300b lovers

Which rectifier is better: full wave or bridge?

Full wave is preferred if you are planning a bipolar (+ and -) power supply. If only a single pole, a bridge does have the advantage of the power transformer not needing a center tap. That advantage is because a center tap is never truly centered, so the output from the transformer applied to a full wave is slightly different with each half of the AC waveform. This means the diodes are making slightly different current spikes as they commutate (turn on and off). 

That might not make much of a difference, since the best way to snub the circuit (to kill the swept resonance that occurs when the diodes shut off) is to snub the rectifier with a resistor and capacitor in series across the input to the diodes. In case its not clear, the source of 'diode noise' is really the power transformer inductance, interacting with the capacitance in the junctions of the diodes. For this reason, if you use a semiconductor rectifier setup, you really want to keep the leads from the power transformer as short as possible.

If you do that right, you can make the nasty silicon rectifiers perfectly silent; obviating any need for a tube rectifier (who's main advantage is low 'rectifier noise'). I put that last bit in quotes since the transformer is what is causing the noise, reacting to the rectifier.

In the old days it was common practice to put a 0.01uf cap in parallel with semiconductor rectifiers. This does nothing and might actually make the problem worse. Any diode junction, if you plan to snub the rectifier itself, must use a resistor in series with a small capacitance to properly snub the rectifier. But usually you can have great success with just a resistor and small capacitance at the input to the rectifiers instead, since snubbing the transformer is putting out the match rather than the forest fire.

I changed the load resistor from 57K to 39K Ohm. The difference in square wave overshoot is very small, but frequency response peak is just +0.5dB instead of +1.3dB.

You can also see that bandwidth has opened up a bit. Something to keep in mind: when you load down that interstage transformer, it changes the load presented to the tube driving it. You might be able to further optimize the circuit by making changes in the driver tube circuit. Also, if I  were you I'd be try to sort out why the FR drops off so dramatically at the bottom end (is it the transformer or something else). 

The bottom end is predictable. ~8K 6sn7 (260v, 9mA) internal impedance and 70-80H transformer primary inductance gives -3db around 20Hz. Actually I measured -2.37dB at 20Hz and -3.62 at 15Hz.

@alexberger Here's something to keep in mind. If you are not using feedback, then its likely at either end of the bandwidth of the amp that the FR will fall off on a 6dB slope. Filter theory tells us that a 6dB slope introduces phase shift to 10x or 1/10th the cutoff frequency. So a 20Hz cutoff will manifest phase shift to 200Hz since there is no feedback correction.

The ear perceives phase shift of a single frequency very poorly, but over a band of frequencies it interprets it as tonality. At high frequencies the ear also uses phase to sense the sound stage. A high frequency rolloff above the audio band can cause darkness in the presentation. At low frequencies a rolloff will be perceived as a loss of impact.

Transformers have bandwidth limits. The smaller you make the transformer, the wider its limits can be. So you have a chance using an interstage transformer to have one that has good LF response- 5Hz is a good place to be. Because you are designing an SET you're dealing with a low frequency bandwidth issue in the output already. All I'm saying here is the less phase shift you present to the output section, the less phase shift will be present in the output.

If you're planning to use a sub with this system this might not be much of a concern. But of you really want to use a sub properly, it should be cut off no higher than about 80Hz otherwise it will tend to draw attention to itself, requiring that it be in the same place as the main speakers. If a sub is used, probably best to drive it with a preamp signal rather than that of the amp due to this bandwidth issue.

And then we get into the deeper waters of the sonics of the regulators themselves. Some are slow and noisy, and intermodulate with the music. Others are fast and silent. Regulators do not all sound the same, and passive CLC supplies can have a signature too, depending on the capacitors chosen. There is no one-size-fits-all solution.

The tricky bit using regulators is bandwidth.

The output of any regulator has a certain impedance. What you are looking for is a linear impedance curve across the entire audio band. The tube regulators often had a problem with this; the output impedance rises when the regulator meets its bandwidth limit. At this frequency the regulator has to be bypassed with a capacitor that keeps the output impedance as linear as possible. It won’t be perfect- usually you wind up with a step in the output impedance at the crossover point. Too much capacitance can cause the regulator to run hotter and with tubes, you can have reliability problems that might occur when the capacitance is charged during warmup.

Solid state regulators often have much wider bandwidth so don’t usually need so much bypassing, but they can have stability problems so are often bypassed with a small capacitance at the output to prevent it oscillating.

Adding additional capacitance to the output of the regulator will do little to improve sound quality and if enough capacitance is added, will increase the heat of the regulator and could threaten its reliability.

If the regulator is properly bypassed and operated well within its limits then they will tend to be neutral.

To help the regulator along, its a good idea to do as much as you can to minimize noise at the input of the regulator. For example a PI network is helpful to reduce the amplitude of the sawtooth waveform at the regulator’s input; this will reduce the work the regulator has to do, which can reduce its operating temperature.

what are your thoughts on using a balanced input stage to drive a single ended 300B? My front end hardware is all balanced and I like using balanced connections. I’m thinking an LTP with a CCS on the cathodes and a push/pull interstage on the anodes. Output of the IT driving a single 300B with fixed bias connected to the other side of the IT secondary. I haven’t decided which tubes to use for the input stage, but was thinking about triode strapped D3As.

@jaytor You’ll want as much gain out of the voltage amplifier as you can get, since the more gain also yields better Common Mode Rejection Ratio. But you’ll need a good Constant Current Source to optimize the gain stage. The D3A, triode strapped, will give you a good mu value to work with, although you could get that with a single 12AT7 and be in the same ballpark. A 12AT7 will allow for plenty of bandwidth.

If you use a driver transformer as a plate load, the issue you’ll be up against is imbalance of plate current between the tube sections. The better your CCS the less you’ll have this problem, and matching is a very good idea. The more current that isn’t cancelled in the magnetic core you can also expect greater distortion. So you can see getting the plates to have equal plate voltage is important. To this end, using a balance pot in the cathode circuit to balance the plate currents, while effective, has the effect of also reducing the differential effect, increasing distortion and reducing gain and bandwidth. So matched tube sections and a really effective CCS are paramount.

One thing to consider is its really impossible to get a perfect center tap in a transformer. Its always going to be a little off.

Since its often difficult to get really well matched tube sections (they should be matched on a curve tracer for best results), I prefer to use RC coupling to a cathode follower, which is in turn direct coupled to the grid of the power tube. The power tube would thus obtain bias from the driver tube. If the driver tube gets weak, the power tube will conduct less also, preventing damage. I explained this topology earlier in this thread.

If you are considering using a differential amplifier as the input voltage amplifier, you really should consider a bi-polar power supply of equal B+ and B-. This will improve the differential effect and the effectiveness of the CCS. If you have such a power supply then you have a good way of setting up that cathode follower I mentioned. If you go this route, care must be taken to make sure the plate of the driver tube is well bypassed so even at full output there is no noise, no artifact caused by the signal on that plate. This will really help the amp assume greater authority.

Are there any tubes that, in a PP IT loaded design, would be reasonably linear and provide enough drive for a single 300B without a cathode follower stage?

@jaytor Sure! We are assuming that you plan to use a high efficiency speaker so you’re not also looking for a lot of gain. But I have to get some clarification- do you want the ability to have enough drive for the power tube, along with enough voltage gain at the same time? They are not always the same.

Tubes that can swing the voltage needed for the 300b include the 12AT7, 12AU7, 6CG7, 6SN7 and so on. I would stay away from the 6DJ8 and similar frame grid tubes; while they are very linear and can handle the current you need, they aren’t very good with microphonics as they are not meant for audio.

A bipolar supply supply is helpful because there is always a bit of differential performance left on the table by any CCS. The more negative the supply, the higher the impedance of the CCS can be (which is a good thing). You can, with modern semiconductors, easily built a good CCS that will work on -100V, but If you are already building a B+ supply, its really not all that hard to build a B- supply from the same transformer. I like to use a separate power transformer for the driver and voltage amplifier, so as to minimize noise appearing in the power supply of the output section from messing with the rest of the amplifier. This helps reduce IMD.

2. I use interstage transformers and separate PS for driver and input tubes. As a result, I have two separate GNDs: 1st - for output circuit and PS 2nd - for input/driver circuit and PS. Should I connect both these GND together to the amplifier chassis at one point? Or can I leave one of these GND floating?

@alexberger I can confirm everything @lynn_olson mentioned regarding your post in his above. Pay attention to all those issues!

Regarding this question which has to do with grounding.

You’ll want to ground your chassis if its metal (and if you’re running single-ended, ferrous materials will provide audio-frequency shielding). So the ground connection of the power cord should be tied to the chassis.

The audio circuit, if tied directly to chassis, leaves you open to ground loops. To avoid this, lay out your amplifier circuit so all the points that go to ground do so to the power supply (star grounding is nice) without touching the chassis.

Grid and cathode connections for each tube section (the ground side of the grid resistor and ground side of the cathode resistor) should be common and use a single wire going to the star ground. This forces noise imposed in that wire to be common mode and so will reduce noise.

The spot in the power supply to which you tie your grounds can be a buss tying all the filter cap grounds together.

Once you have all the grounds starred together, at that point you need to reference the chassis ground. That can be done several ways: you can use a simple resistor from the audio ground to chassis ground. You can use a pair of rectifiers in parallel, each opposite of the other, with a resistor in parallel with that. This higher impedance is presented to ground currents between chassis and the audio grounds which otherwise set up the ground loop.

Be sure to have the input connectors also isolated from ground. Most RCA connectors are supplied with shoulder washers for this purpose.

In this way, the chassis can be a relatively quiet shield for the audio circuit. You’ll find the audio circuit to be quieter with this practice.

Does aluminum provides any shielding? I will use an aluminum chassis.

@alexberger Yes, but only at radio frequencies. No shielding at audio frequencies, and for that reason steel is likely a better choice if you're running single-ended.

We use aluminum chassis in most of our stuff, but its balanced and so does not need shielding to be quiet.

@alexberger The resistor value isn't critical. 50 to 100 Ohms works nicely. As long as both chassis (power supply and amplifier) are tied together and grounded by the power cord, its really not going to matter where that resistor actually is, since it will carry no current under normal circumstances.

The prototype preamp is essentially a circuit that was a thought experiment of Lynn’s for a fully balanced design that has been around for years and no one ever really tried to build and perfect it. 

@donsachs I'm curious what is meant by this. We've been working on a fully balanced design since the mid 1980s, which started selling in 1989. Is it the use of transformers in the circuit path?

Most people think a 300b is a fuzzy rolled off tube.  That is because they heard it in a single ended amp with a mediocre power supply, and most likely an inadequate driver section, and possibly mediocre output transformers.

@donsachs I made this point on another thread just a day or two ago. People get so hung up on the tube, whether a plus or minus, while the circuit design, which is far more important, gets ignored.

Ralph has written in the past that the 6SN7 is a sufficient driver tube “dependent “ on its implementation in a given circuit. I have been told that the 6EM7 provides much more current, power and drive capability compared with the 6SN7.

I will say that I’ve heard an excellent sounding PSET 845 amplifier that utilized the 6SN7. I could certainly be wrong, but isn’t the 845 a more difficult tube to drive than the 300b?

@charles1dad The issue is always how much voltage swing will drive the tube to full output and how much grid capacitance is there. The more grid capacitance, the harder the tube is to drive.

You can get a 6SN7 to drive a single 300b no sweat. But doing it the way is traditionally done in SETs will likely but the tube at a disadvantage.

The 6EM7 does have one section that is considerably more gutsy than a 6SN7 has. But you will always be looking on the collector's market for replacements.

The way to get a 6SN7 to really do the job is to have it wired as a cathode follower direct-coupled to the grid of the power tube. In this way the coupling cap used in the amp stays rather small since its driving the 6SN7 rather than the output tube. Smaller caps are more transparent...

This is the technique we've used in our OTLs for several decades now. Its very reliable. One 6SN7 section is thus able to drive a number of highly capacitive triode grids with the voltage swing needed. I do not understand why this approach hasn't been used in SETs since the design consideration is similar. It might be because a negative power supply would be needed. But from what I've seen, SET users are not particularly worried about cost if the amp gets the job done for them.

RM's minced no words in responding that 6SN7 tubes were a very poor choice of tubes -- "better employed in old TV's", because of their inherent high noise levels.  

@whitestix I encountered that with him as well- although by that time he was beefing about the 6SN7 linearity family of curves. Turned out his source was an early Tung Sol tube date manual that didn't reflect 99% of most production tubes. RCA and Sylvania of course figured things out with the 6SN7 early on and in another year or so, so did Tung Sol. RM had the bad luck to have a Tung Sol manual that was outdated. 

There is a group who prefers this approach and there is a group who prefers the more simple way - using half or the 6SN7 in parallel as one triode and interstage coupled with the next stage, or alternatively using a more capable driver, including a DHT such as 300B itself, 4P1L, 10y, 801, EML’s 20 or 30, etc. 

@ffzz I've found that implementation plays an enormous role in zero feedback amplifiers: grounding, component quality, power supply design and of course getting the operating points right in the circuit design. I challenge the idea that using an interstage transformer is actually a simpler approach- it is once you have a suitable transformer, but a good transformer design is the tricky bit; on that account direct-coupling is IMO easier. It has the same advantage of being able to support class A2 operation as well as instantaneous overload recovery, plus wider bandwidth and lower distortion.

The lower distortion may well be why it sounds less 'dynamic'! The use of that word when associated with SETs is always tricky, since most of the 'dynamic' nature of SETs has to do with how they make distortion. IOW its normal for a lower distortion circuit to sound less 'dynamic'. 

These amps do PP a different way and limit the phase split issue and avoid some SE distortion.   Again, there are many fine examples of all sorts of amp topologies, but these don't sound at all like other PP amps you have heard.  They sound like a great SE amp with the drive of a PP amp.  That is the design goal....

@donsachs +1

I've been harping on this very issue for years. Our amps do not have a dedicated 'phase splitter', relying instead on good CMRR in the differential voltage amplifier.

 Conventional wisdom is that caps etc take a while to run in so do you have to run in the whole design for a a while to get the true sense of it and then tinker with the component parts thereafter.... and then wait for them to run in again before making an evaluation?

@whitestix Usually a thing like a coupling cap will reveal its character fairly quickly. We've been doing this since the 1970s and in that time have yet to see a coupling cap change so dramatically during break-in that it exceeds the character of another, better sounding cap with the same time on it. So you can audition them easily right out of the box. So far the Teflon dielectrics have proven themselves over and over again. A regard paper and oil as very nice sounding parts as well, but they can develop a voltage drop across them which can throw off operating points in the design; IMO not worth the return shipping and frustration!

@lynn_olson I've been harping about the audibility of the higher orders for a very long time. Nice to see some agreement in this regard.

@donsachs @lynn_olson @thom_at_galibier_design 

I confirmed a theory I had about OPTs over the weekend. You might give this one a try; its inexpensive. Replace the bolts that hold the OPT together. Typically these are made of steel and are insulated from the transformer by fiber or plastic shoulder washers.

You can get non-magnetic stainless bolts to replace the steel bolts. Its not a big change, but in the case of SETs or lower powered tube amps in general, every drop counts. I measured about a 7% increase in amplifier power. I suspect this will vary depending on the transformer design as well as the specific alloy of stainless bolt used.

The shoulder washers are supposed to take care of the problem of a magnetic short of course and for the most part they do. But they don't do the job perfectly and I suspect that since this technique is 70-80 years old, tradition has set in and caused no-one to look into it further.

I found out decades ago that the mounting bolt in a toroid transformer, commonly made of common steel, would heat up more than the transformer because the actual toroidal mag field was sloppier than theory. So the bolt was a short to the field. By replacing it with non-magnetic stainless the transformer ran at a lower temperature.

I've been working on a low power PP tube amp recently so decided to give this a try.

The 300B ... all of them ... quite happily accept at least 20 volts of positive grid drive. This is not secondhand info gleaned off the Internet, I’ve seen it for myself on a Tektronix scope screen back in the Nineties. I was frankly surprised, because there wasn’t even a trace of a kink or a glitch as it went from negative to positive grid drive. I was expected more drama from the Big Bad Positive Grid Drive, but nothing, no drama, and no signs of grid or plate overheating, either.

@lynn_olson Back in the 1990s we built an experimental OTL that used four 6300bs (a graphite plate variant of the 300b) per channel. The plate voltage was only 150V since higher than that is impractical in an OTL. To get the tubes to conduct properly we biased the tubes at +15V as their operating point. We played that amp at CES that year. The only reason we didn't produce that amp was it was impractical- that's a lot of money to spend on power tubes for a 15 Watt amplifier! We could get slightly less than double the power using four 6AS7Gs which could be had for less than the cost of one of those 6300bs.

Do you believe that the output transformers in these amplifiers is the overwhelming factor that informs your opinion?

@charles1dad I've said it many times in the past. The greatest limitation SETs have is getting bandwidth as the design is scaled for more power. The OPT is the defining issue.

A thread with this caliber of participants stimulates further inquiries from posters. Informed commentary is valued. 😀

@charles1dad Such is the nature of the internet I suppose, where fact and opinion freely co-mingle. The physical nature of SET transformers are governed by physical law FWIW and that law isn't interested or caring about opinion. It simply is.

For example, floating paraphase phase inverters instead of split-load inverters or Mullard long-tail pairs. The phase division isn’t as precise, which is why they dropped out of favor, but the drive capability is much stronger than the other two types.

@lynn_olson A simple way to improve this circuit is to use the bias supply for the KT88s as a B- voltage and then use a 2 stage constant current source to set the operating point of the differential driver. This improves the differential effect quite a lot and gets rid of the need for slightly different plate load resistors- they can be matched instead.

To this end we use bipolar supplies in our amps; B+ and B- have the same absolute value. No balancing is required in the differential amplifier and often the plate voltages are within 3-4 volts of each other. This improves the CMRR quite a lot which in turn reduces distortion (there is the additional benefit of more gain and wider bandwidth as well). If the input stage is built in a similar manner, even orders will cancel throughout the circuit, resulting in a dominant 3rd harmonic which can mask higher orders.

The harmonics will be found to fall off at a faster rate (than seen in SET circuits) as the order of the harmonic is increased; they will follow an exponential decay based on a cubic function. This works really well for the human ear (smoother sound and greater detail, both on account of reduced open loop distortion)!

To eliminate a frequency pole caused by a coupling capacitor, we’ve been using a differential cascode circuit as the sole source of gain in our OTLs. Because the gain is increased in that single stage of gain, so is the CMRR and overall differential effect. This allows one to use a cathode follower driver direct coupled to the output section. In our OTLs the power tubes are in turn direct coupled to the loudspeaker. If you were to use an output transformer, a pair of DHTs are easily driven- linearity is such that no feedback need be used. Bias is obtained from the driver circuit, so if bias controls are used, they are in the grid of the driver tube rather than the power tubes.

If feedback is desired, it can be wrapped around the circuit and applied in a manner identical to how its done in opamps- using resistor divider networks that mix the feedback with the incoming signal at the grids of the input stage. This technique vastly reduces distortion that the feedback signal would otherwise encounter, which in turn means the feedback is more effective at its job, generating less higher ordered harmonics (caused by non-linearities in the feedback nodes traditionally used in both tube and solid state amplifiers). By doing this a wider range of speakers can be used.

1. Does it matter what volume a power tube is played at? Does that effect tube life?

@markusthenaimnut 

If the tube is operating class A1, the power its making won't matter. If operating class A2 or A3 the higher power levels will probably affect tube life. If the tubes are running class AB then they will run cooler, which could translate to longer life, but higher power will shorten that. You didn't mention the load but that affects things too- its rare that the speaker actually loads the output transformer correctly for a given tap; transformers transform impedance in both directions so a load too low on the output transformer will be a load too low on the output tube(s) as well. That will reduce tube life as more of the power made by the output section will be dissipated in the tube(s) rather than the load!

2. Does it matter if a tube is cooled, say, by a small fan nearby?

It helps! During WW2, 6L6s were used due to shortages to get amazing power levels by being water cooled.

3. If a tube is powered up but not making music does that "cost" tube life just as if you were playing music through it?

Always- how much depends on the class of operation and other variables such as dust on the envelope and so on.

4. What is harder on a power tube? Turning it on and off, let's say twice during a day (two listening sessions totaling three hours) or letting it stay on, let's say for an eight hour period?

Tubes wear out no matter what you do. They also draw power... The longer the off time the easier it is to answer a question like this. If we're talking about an indirectly heated power tube, the turn on should include time where no B+ is applied until the power tube cathode is properly warmed up. We supply a Standby switch for this purpose on our amps. If the tube is directly heated it won't matter. So if it were me I'd shut the amp down when not in use, even if for only an hour.

True, push-pull substantially reduces distortion, but in reality it only reduces even-order distortion ... 2nd harmonic, 4th harmonic, 6th harmonic. etc.

@lynn_olson This statement is not correct. If the circuit is fully differential and balanced (in effect, PP from input to output) distortion is compounded less from stage to stage. The result is less 3rd harmonic than you would otherwise see if only the output section is PP.

In addition I should point out that the 3rd harmonic is innocuous in the same way as the 2nd as long as it is at the same level or less as seen in an SET, which as you pointed out in most cases (except for the example I just gave) it will be.

A properly functioning tape recorder makes a dominant 3rd harmonic that is actually higher than seen in most PP tube amps at full output. No-one really seems to complain about the ’sound’ of tape; rather people seem to like it quite a lot!

How about using SE as opposed to PP in the first stage and a SE to PP interstage transformer between the first and second stage?

@ffzz When you combine single ended and PP in one amplifier you get two distinct non-linearities (since neither circuit is perfectly linear). PP circuits generate what is known mathematically as a ’cubic non-linearity’, while single ended circuits generate a ’quadratic non-linearity’. The cubic or quadratic function describes a lot about how the higher ordered harmonics behave in the circuit. This is important because higher ordered harmonics can contribute to harshness and brightness, as that is the tonality assigned to them by the human ear.

When you combine both circuits in the same amp you get both non-linearities. Norman Crowhurst (one of the technical gurus that wrote about tubes and other technical/engineering topics related to audio) wrote about this 65 years ago, stating that the result is a more prominent 5th harmonic. With this understanding I’ve not found it surprising that SET owners prefer their amps over PP simply because the PP amp combined both types of circuits with the result described.

I hate the sound of solid state devices.

@donsachs There are now class D amplifiers that have a distortion spectra that if you had one on the bench you would be completely convinced you were measuring a really well-behaved triode tube amplifier.

There are three reasons for this:

1) its really easy to get really high Gain Bandwidth Product values, such as 20MHz with class D, so feedback can be supported across the entire audio band (no rising distortion with frequency- like a zero feedback amplifier in this regard)

2) the feedback node can be very linear so the feedback signal does not get distorted prior to doing its job.

3) the things that cause non-linearities in the circuit tend to generate lower ordered harmonics.

Despite inexpensive power being available, I still prefer higher efficiency loudspeakers for the reasons Lynn outlined above. In addition, the harder you make any amplifier work the higher the distortion. For this reason I feel that higher efficiency and higher load impedance is an advantage to all audiophiles.

Maybe.  Except that there are things we cannot measure.

@donsachs That was true back in the 1980s. It really isn't now. Measurement technology has really advanced in the last 40 years!

Its the distortion of any amplifier that is its 'sonic signature'.

There are many class D amps I've heard over the years that I had to struggle to take seriously. But I've heard some now that sound every bit like a very good tube amp; just like in a tube amp where arcane subtleties can make or break a design, the same is true in class D (or any design for that matter). I pointed to what is needed to make a successful amplifier (if you're going to use feedback) in my prior post. Most amps using feedback fall short of the GBP needed so distortion is much higher at higher frequencies than the THD measurement suggests!

THD by itself, if that is the only harmonic distortion metric used, allows a lot of problems to be swept under the carpet. When it is the sole metric, it leads to the myth that 'there are things we cannot measure.' The reality is if the harmonic spectra is measured at several frequencies (including above 1KHz) and if distortion is graphed vs frequency, then we start to be able to predict the 'sound' of the amplifier.

Both the 'measurement only' guys and the subjectivist guys hate this! But Daniel von Recklinghausen was right- if it measures well but sounds bad, you measured the wrong thing.

I’m convinced that no feedback plays a role as every amp I’ve heard with feedback doesn’t have that super inky back blackground or perfectly rendered transients.

@cloudsessions1 That is because in most amps employing feedback, its poorly applied- so my surmise is you've yet to hear one where the feedback was really right. Norman Crowhurst pointed this problem out 65 years ago, describing how the feedback node (the point in the amplifier where the feedback signal is mixed with the incoming signal) isn't linear; therefore the feedback signal gets distorted before it can do its job, thus creating higher ordered harmonics that have given feedback a bad rap. Its not the fault of feedback so much as poor application. Amazingly, little has been done in the last 6 decades to fix this; IMO mostly out of ignorance and a lack of will to do so.

I've heard amps with very high feedback that sound utterly relaxed in the mids and highs and portray depth with ease.

What are some amps you’ve heard with high feedback that you’ve enjoyed? I’ll have to check them out. I’ve seen the measurements for the Purifi and its distortion is incredibly low, to the point it’s at the limits of what the AP analyzer can measure. It does have raising distortion in the treble but if I had to surmise, it is most likely low order as I find the purifi a touch sweet in the treble. I do run it without an input buffer as my preamp is up to the task of driving the Purifi module directly.
I have yet to hear the new Gan stuff that is specifically designed for audio. I know yourself and AGD specifically designed yours for audio applications unlike many of the other brands. Are you guys going to be at the pacific audio fest? I’d love to come hear your Class D amps and hear how it compares to the 300b statement monos.

@cloudsessions1

Our amp of course, which you have to imagine we’ve compared extensively to our class A triode zero feedback OTLs.

The AGD Audion (although I don’t know how much feedback is employed in this design)

Orchard Audio

Digital Amplifier Company (unfortunately the designer passed away last year)

We won’t be at the Pacific Audio Show- I have a prior booking.

@lynn_olson

Interestingly, Class D amps are free of Class AB transition artifacts, so there’s an entire class of coloration that just isn’t there. The big issue for Class D is nanosecond precision of timing for the pulse-width modulation (Class D is pulse-width-modulation, akin to FM, and not PCM), and insensitivity to reactive loads affecting the PWM modulator.

Yes, crossover distortion artifacts are impossible for any class D using an output filter. Timing issues (which cause an unpleasant hiss) are solved by a self-oscillating topology. This is done by exceeding the phase margin of the amp and the oscillation is used as the switching frequency. This makes for a very high stability amplifier and very large amounts of feedback. Since the feedback is not distorted prior to doing its job, higher ordered harmonic generation is avoided. Essentially the feedback simply reduces the existing distortion (in our circuit this tends to be lower ordered harmonics). Distortion vs frequency winds up being a flat line across the audio band. So it sounds very much like a zero feedback triode amplifier, but more ’focused’ owing to lower distortion, which otherwise obscures detail. You should try it- you make excellent amplifiers and know how they are supposed to sound, so you are in a good position to see how this technology advances the art.

There is a simple and inexpensive means of dealing with heat due to amplification. Ducts in the ceiling above the amplifier(s) can be added, along with dryer style hose the connects via a squirrel fan to the outside. This can be installed for a few hundred dollars (mostly labor), is quieter and far less power draw than air conditioning. 

Do you use load a resistor after the interstage transformer? What kind resistors do you use?

@alexberger If any audio transformer is not loaded it can 'ring' (which is to say it will generate harmonic distortion, which can be quite profound). The correct loading will cause a state of 'critical damping' in the transformer, where when a square wave is put thru the device, it will have little or no 'overshoot'. Since the grid of a tube tends to be very high impedance (other than its capacitance) some form of resistive loading is a good idea to explore!

Since you sound like you are up to something with your own design, I recommend an empirical approach, which might be to drive the transformer with a square wave and have it drive in turn a power tube which is operating normally. A potentiometer and oscilloscope's probe across the output of the transformer would then allow you to vary the potentiometer and observe the result on the square wave. In this manner you can exactly dial in the correct loading value.

The difference is Class D switches at 100 kHz or higher, with pulse width translating to signal level, while Class A is non-switching and like a preamp, fully analog from start to finish.

@lynn_olson most quality class D amps switch at 500KHz or more. Also to be absolutely clear, class D is 'like a preamp, fully analog from start to finish.' A very different type of circuit, but analog nonetheless. 

I do not have the resources to confirm this via test bench measurements as you do. I can only rely upon listening experiences with my own audio system.

@charles1dad Lynn is correct- if you are not running feedback then everything in the amplifier design right down to the component quality, wire and solder makes a difference. When you run lots of feedback you get rejection of things like that- often including sensitivity to line voltage.

Zero NFB seems to uncover a masking effect.

@charles1dad That's not quite correct. It depends on how well the feedback is executed and as I've been pointing out a lot recently, with most amps made in the last 70 years or more that's not happened.

But if its done correctly, amps using it can be amazing- no hint of dryness, nice soundstage; everything you want in an amp.

I've described what's needed often enough, no need to repeat it here.

I understand that you are referring to the approach you use with judicious application of negative feedback. I was referring to Lynn’s comparison to the “golden age “ PP amplifier which typically used 20db of NFB. Your current class D amplifiers do not fit this description.

@charles1dad 

They don't! However, Futterman claimed 60dB of feedback in his OTLs. That was during the 'golden age'... Also Kron-Hite made a transformer coupled tube amplifier (UF-101) that claimed even more feedback! It was built for laboratory use but works great for audio as well. Its specs are astonishingly impressive and having had a set (they were single channel) I can say they sounded quite decent. That was a long time ago but we compared them to an ARC amplifier which got its doors blown off.

If feedback is applied properly it is really beneficial. If its not then it will mess things up with the amounts normally found in tube amps (12-20dB...).

In the last couple of years, I had the VTV Audio EVO 1200 Class D amp with the Purifi module, with the aftermarket ministrations of Ric Schultz, and while the sound was as Ralph describes it,

@whitestix I'm going to contest this; I was not describing those amps at all and haven't heard them. You can tell something is up that isn't right since when you go on the web, you see really variable reports about their 'sound'. I notice that doesn't happen with tubes- everyone agrees that tubes sound smoother and often have more detail and so on. To what degree and what emphasis is the differences between the tube amps. I'm saying I've been playing a class D that you would think is in that category if you heard it. Going back to tubes you don't find that tubes are bringing anything more to the party. I get that seems like a tall statement. Keep in mind that OTLs have ruled the roost in the transparency department of the tube amplifier world and I've been building them since 1977 or so. During that time I've heard many tube amps; I repaired audio for a living as I put myself thru college and afterwards until Atma-Sphere was able to keep me busy full time. So I know what tubes bring and I'm telling you there's at least one class D out there that does the same thing. I suspect there are others.

Sorry I won't be able to participate in this thread for a while- I'll be out of town for 4-5 weeks. Y'all have fun and be well! 

 A Cornell Dublier engineer once explained to me that the difference between a 150V rated cap and one rated at 175V was the way the cap was formed up and nothing else. Forming is a process where voltage is applied to the capacitor with a target of the rated voltage; something that is done only once in the factory.

But the cap is almost never used at the forming voltage. Usually for best reliability, they are to be run at about 80% of the rated voltage.

So the cap is not as efficient at the new voltage when new. It takes time for the cap to 'form' to the new voltage. Its important to understand that electrolytic caps have some properties in common with batteries and so are fundamentally different from film caps in this regard. Charging them and polarity are two examples of this similarity. Forming is one way they are unique.

Anyway, when the cap forms up to the new voltage used inside the amp or preamp it will be a more efficient bypass. Its my theory this is what people hear during break-in. I've found it measurable too- the voltage once the caps are formed is every so slightly higher and the supply is less noisy.

Are you at the point where you feel your Class D amps are equivalent to your best tube amps?

@whitestix Yes.  After extended listen sessions as you put it I don't find that our tube amps bring anything to the party that the class D doesn't. I really doubt that we're the only ones that can do this so this has led me to thinking that tube power amps are on borrowed time at this point- that is if sound quality is the only arbiter. People do like the glow of tubes- so do I. But I've found also that I don't miss them in the slightest, despite liking tubes so much as to make a business of them.

Ralph, it seems curious to me that knowledgable audiophiles would rely on anything but sound to be the arbiter of their purchase decisions.  Tube gear is a hassle and costly, but to my ears, it is the price to pay for such magnificent sound. I think buyers buy tube gear because it sounds better, pure and simple.   

Yes. That is what keeps any manufacturer in high end audio in business :)

The Raven is particularly well-suited to driving transistor amps because:

1) The transformers prohibit the transmission of DC pulses to RCA and XLR outputs, under all conditions, including total failure.

2) Unlike solid-state preamps, there are no DC servos to fail. There are no coupling caps on the output to store turn-on pulses, unlike cathode-follower designs. No muting relays are needed, or used, so the signal path is direct from transformer secondary to output.

@lynn_olson FWIW dept.: We've been using a direct-coupled output for our fully balanced tube preamps since their inception in 1989. To that end we've used a servo to sense and correct DC at the output. Here's the FWIW bit: a servo is (or should be) very reliable. We've not had a failure since their inception.

The magazine conventional wisdom would tell you that clarity and beauty is "euphonic coloration". That’s complete horse****. Euphonic colorations can’t add detail, resolution, more depth, and more in-the-room presence ... colorations can twiddle with subjective tonal balance, and usually adds mush, murk, or grain.

@lynn_olson You might want to play around with this applet:

https://www.falstad.com/fourier/

Select 'sine' and the little dots below the waveform are movable and represent harmonics.

It shows why euphonic colorations (which are only the 2nd and 3rd harmonics) can indeed add to (or subtract from) detail and 'dynamics' and alter your perception of depth and soundstage.

If you only play with the 2nd and 3rd harmonics, and also work with their phase, you see some interesting results. For example if the phase of the 3rd is out of phase with the fundamental, the waveform actually gets taller.

Harmonics define the sound of musical instruments. You can see from this little applet that distortion can bring out details of musical instruments or obscure them.

if the input of the next component is balanced and not referenced to ground (e.g. transformer coupled), I don't understand why it is necessary to decouple the output in the source component from any ground reference to achieve the full benefits of balanced connections. Can you please help me understand. Thanks.

@jaytor Part of the issue driving interconnect cables is how the signal travels in the cable. When the shield is part of that connection, its more likely to pick up noise and the actual construction of the cable (what sort of insulation it uses and so on) becomes more critical. That shield is connected to chassis ground at the input of whatever is being driven- so now you also have the possibility of a ground loop too.

So when the source is referencing ground, such as a pair of single-ended outputs, one of which is out of phase with the other, you have a problem where the ground circuit return is active in the shield of the cable. Suddenly the dielectric in the cable is playing a role that it did not when the shield was only used for shielding with no signal on it.

It is precisely this problem which is why there are 'high end audio' balanced line cables now that might cost up to $1000/foot or more (put another way, most 'balanced outputs' on 'high end audio' equipment actually references ground as if the designers were not aware of the balanced line standard)! If the connection is done properly, you won't be hearing the sort of differences between cables that might convince someone (who might have a touch of audiophile nervousa) to spend that $1000/foot.

I'm saying that an inexpensive cable can sound just as good in every way.

The proof of this is the vast number of recordings that were made in exactly this way- proper balanced outputs and inputs. Its part of why you could have 150 feet or more of interconnect cable between a microphone and the input of the tape recorder in 1958, nearly 20 years before Robert Fulton showed off his first 'high end audio interconnect' cable, yet the resulting recording just gets better and better as you improve your system's ability to winnow more information out of that recording. That can only happen if the cables used to make that recording are absolutely transparent!

Put simply, you have to dot your 'i's and cross your 't's if you want this system to work properly.

But let's look a bit closer at that balanced source that references ground. It may well be rack mounted in a relay rack and through that rack its chassis is grounded to every other bit of equipment in the rack or maybe even in the studio. Some of that equipment might be on the input side or the output side. So a ground loop could easily be introduced! 

You might think that because you're not using a 7' tall steel relay rack at home that you won't have that problem, but keep in mind that the equipment is also grounded into the wall. That's where you get in trouble: you must be sure that ground is ignored with both inputs and outputs; that ground is only used for shielding in cables and never for any kind of signal ground! If you don't do this, the benefit of balanced operation is eroded. It was designed so that exotic cables aren't needed and grounding issues are eliminated.

Think about the advantage of having cables that sound as good as the best out there price no object, but not having to pay that price- for all the interconnects in your system, you might have only a few hundred dollars invested at the most, rather than $1000s or $10,000s. And they don't go out of date or any such thing...

The idea of introducing a tube-driven class A stage to achieve "better cable drive, partly signal conditioning, scraping off RFI and noise induced in the cables" is appealing. How would you recommend I learn about this?

@lewinskih01 You might want to study how balanced lines work. Properly done, balanced lines are the best cable drive available to audio. RFI and noise are rejected due to the low impedance aspect of balanced lines (in the old days the studio line inputs were 600 Ohms; these days its more like 1-2KOhms); weak signals induced in the cable are swamped by the low impedance. In addition the input that is being driven has a high Common Mode Rejection Ratio, which is to say that signals common to both the inverted and non-inverted inputs (such as noise and RFI) get rejection.

In a true balanced line system ground is ignored to eliminate ground loops. If using tubes this is usually done using an an output transformer which can float with respect to ground. Its also possible to direct couple using a Circlotron output, for which Atma-Sphere has several patents.

If you are supporting the balanced line standards (AES48 is one of the standards; the other is the low impedance aspect) these two methods are the only ways to do it.

Why buy an high-end audiophile component, with audiophile pricing, made from off-the-shelf $5 parts.

@lynn_olson

It might be because those parts work...

Tube gear costs a lot because transformers and vacuum tubes are inherently labor-intensive, and the parts are not inserted on circuit boards with pick-and-place machines. I’m one of those madmen who think zero-feedback circuits are interesting, and I like tubes. Nelson Pass is your man if you like zero-feedback JFET/bipolar transistor circuits.

If you are more sensible, read ASR reviews, ignore the comments section, ignore the single-dimension SINAD number, and look at the noise floor of the multitone IM distortion graphs. That is the true wideband IM distortion, and multitone is the most severe test of the entire circuit.

FWIW, we use surface mount parts in the module we designed for our class D amp. We assemble them to the board by hand (no machines). You use different tools for that- a different soldering station, and special reader’s glasses so you can see what you’re doing.

You missed one of the more vital measurements: distortion vs frequency. Why this is important is that it can show you if the amp is going to make more distortion (and audible, annoying distortion) than the specs would otherwise show.

Zero feedback amplifiers have a ruler flat line across the audio band in this regard. Beyond that the distortion spectra must allow the distortion to be innocuous. That’s why SETs sound they way they do.

When the amp has feedback, that’s when you can have troubles with distortion rising with frequency. This happens because the design, whether tube or solid state, has insufficient Gain Bandwidth Product (and also points to poor engineering; feedback is control theory, which is a field that is well understood elsewhere in the electronics industry). For those that do not know this term, GBP is the frequency where the gain of the circuit has fallen to a value of 1 (unity gain) and so is the highest frequency where a sine wave can be relatively undistorted. Obviously an amp with a gain of one is not useful- 25 to 30dB is more useful so a preamp can drive the amp in a conventional manner (SETs don’t need quite so much gain, but since they don’t usually use feedback they aren’t part of this discussion).

For example if the amp has a GBP of 1 MHz and we are looking for 30dB of gain (a gain of 1000) out of the design, you divide 1MHz by 1000 and you get 1KHz. That is the frequency where the feedback will fall off on a slope (starting at 6dB/octave, but as frequency is increased, falling off faster)- and the distortion will rise on a converse slope.

This is why a simple THD value can hide dirt under the carpet; the fact that distortion will be much higher at 7KHz than it is at 100Hz. Its why most solid state amps can play bass just fine, but sound bright and harsh- you’re getting more of the audible annoying kinds of distortion at higher frequencies than the specs otherwise show! This has been one of the bigger disconnects between the spec sheets and what we hear over the years and has given rise to the myth that there are things we can hear that we can’t measure and explains why amps that ’measure poorly’ can sound so good.

It is recently become possible to build solid state amps that have so much GBP (we have 20MHz in our class D) that the distortion vs frequency is a ruler flat line, just like in an SET (but of course, overall much lower distortion, so greater detail is audible since distortion can obscure detail); IOW the feedback employed in such amps is supported across the entire audio band. That is why its now possible to build solid state amps that sound for all the world like the best tube amps.

FWIW ASR does on occasion graph distortion vs frequency on their site, but its apparent to me that they don’t understand its significance: the line that exists between that graph and what the amp actually sounds like. If you have all the measurements you can know that!

As for rising THD versus frequency, I haven’t experienced a refinement of treble with linear THD across the spectrum. Properly designed SS has been overall terrific in the upper registries over the last 2 decades. Pass labs, benchmark, and purifi all have terrific top end and all of them have rising THD versus frequency.

@cloudsessions1 Just so you know, this statement is false. Most self-oscillating class D amps, such as the Purifi, do not have rising distortion with frequency. Where ever you got that your source is wrong.

Regarding this comment:

Humans are inherently bad at hearing harmonic distortion don’t take my word for it there’s many blind tests you can do online to see how much distortion it takes before you notice.

This test is probably not done with attention paid to distortion rising with frequency- and in that context your statement is correct. Most of the online stuff I've seen does not have that built-in to the software. So its not the same thing. When distortion rises with frequency, it puts emphasis on higher ordered harmonics. This is at the root of why solid state has had a reputation for being harsh and bright, and also why feedback has gotten a bad rap in high end audio (because it can mess with a tube amp in a similar fashion).

I've already described how Gain Bandwidth Product causes the rise in distortion with frequency. What I've not mentioned in this thread so far is how feedback is usually applied in amplifiers so that the feedback signal itself gets distorted before it can do its job mixing with the incoming audio signal. As a result higher ordered harmonics and intermodulations are created because the feedback node is not linear. Norman Crowhurst (a well known audio guru of the late 1950s and 1960s) wrote about this over 60 years ago, but almost nobody really did anything about it.

You can apply feedback without distorting it. That is done the way opamps do it, by mixing the feedback with the audio signal using a resistor network at the input of the amplifier, rather than inside the amplifier. Resistors are far more linear than any tube or transistor! We've employed that technique in our smaller OTLs for decades now.

@cloudsessions1 Thanks for the links!

Given how much Bruno Putzeys has written about this topic it was my assumption that all of his designs conformed to his ideals. Assumptions can get you into trouble...

I'm Ok with the agree to disagree. I've met very few solid state amps that I could actually live with; hence 45+ years in business making tube amps. I do agree its less of a problem now as opposed to +20 years ago. The semiconductors needed to do the job really didn't exist in the 80s and early 90s.

A differential circuit has a current source or high-value resistor in the common cathode (or emitter) circuit, which is why they are called "long-tailed pair" in the literature. This forces differential operation, but has a limitation because the two tubes (or transistors) are effectively in series. If one device cuts off (impedance goes to infinity), then the other device is hard-limited to 2X the quiescent current. It can never go further, because the long-tail or current source hard-limits total current to both devices.

This statement is false. The devices are not in series, else Kirchhoff's Law would prevent the second device from conducting if the first were in cutoff... At any rate if one device is in cutoff, the other will be in saturation which is the limit of any device's ability to conduct!

@lynn_olson

You might want to read this article:

The Power Paradigm

Most zero feedback tube amplifiers are Power Paradigm devices.

Regarding some of your comments in your post above:

How an output section behaves was not the topic when you brought up this bit of conversation (balanced vs differential). I never said anything about an output section. FWIW its possible to build a differential circuit so a tube can saturate when the other half is in cutoff. Its all about operating points as you rightly pointed out.

FWIW the first differential amplifiers were single pentode circuits; the grid being one input and the cathode being the other. IOW all tubes behave differentially- they amplify what is different between the cathode and grid. On this account, you can see that setting the operating point is the crucial bit which may or may not allow the tube to swing from saturation to cutoff. Drive has a lot do do with it of course. My surmise is Allen simply didn’t set his operating point correctly in your anecdote.

We were building tube differential voltage amplifiers before Allen came on the scene- we were the first worldwide to offer them to the public in a audio product meant for home use. My recommendation is to spend more time working with them and see if you might arrive at a different conclusion.

The above, in a nutshell, is why tube amps behave very differently at clipping than SS amps. It is why, with the right speaker of course, that a 60 watt tube amp can sound like it has the drive of a 200 watt SS amp,

In case there’s any question about why this might seem so, its how the tube output section makes distortion at clipping. A zero feedback tube output section has a very gentle clipping character; at early onset you don’t hear the amp breaking up at all, but the distortion has skyrocketed and the higher ordered harmonics cause the amp to sound louder than it really is, despite no obvious breakup. Its an illusion.

It has led to the myth that tube power is more robust than transistor power. But in simple terms a Watt is a Watt; but how distortion interacts with our ears is a different thing altogether. A sound pressure meter will reveal the truth of the situation easily enough.

The previous point about Class A operation in a differential stage still holds: what happens when more than 100% of the current programmed in a current source is exceeded?

It won't if the circuit is properly designed!

The real question is what happens when the drive to the differential gain stage exceeds the range of that gain stage. The answer is one of the devices saturates while the other goes into cutoff. Picking the right amount of current in the constant current source (if there is one, differential amplifiers do not need a CCS to work... the first circuits we built employed a bipolar power supply; the cathode resistor had the entire B- Voltage dropped across it; this limited current to the same extent that any cathode resistor might in any single-ended circuit) is the key to making sure that the design isn't limited by the CCS. Instead you want it limited by other parameters- the tubes themselves, the plate load, etc. The addition of a CCS increases differential effect- thereby increasing gain and decreasing distortion, as well as improving bandwidth, assuming that the CCS does not impose a bandwidth limit.

@donsachs I get it. I was trying to point out the difference between what sounds 'louder' and actual sound pressure; as you know from playing tube amps the two are not always the same. IMO this is one of the bigger failings of SETs with zero feedback since, more than any other kind of amplifier made, they tend to sound louder than they really are due to how they make distortion.

60 Watts isn't a whole lot less to our ears than 250 Watts is due to the logarithmic nature of our ears. So as long as the 60 Watts can adequately drive the load it can do quite well. This is the same reason we didn't try to build a super high powered class D amp. It was more important to get it right than it was to make a lot of power- as it is, it makes 200 Watts into 4 Ohms (250 at clipping). If your speaker really needs more than that kind of power to really fly, its borderline criminally inefficient, since to merely double the sense of volume to the ear, you need ten times the power. To my understanding there are no 2500 Watt amplifiers that sound like music.

@donsachs Yes- those Joseph speakers were nice.

I can see 88-89 for smaller speakers. I have a little 5 Watt tube amp  I designed for desktop or a bedroom system and I use a pair of Fritz Carbon 6s with it, which are 88dB and I never run the amp out of gas (but I never play it that loud either).

But if they are going to be large there's no reason they should be hard to drive. I keep telling people that if you want to get the most out of your amplifier dollar investment, its best served by a speaker that is higher impedance and easier to drive, on account of the simple fact that the harder the speaker is to drive and the lower the impedance it is, the more distortion the amp is going to make. IOW a simple way to make any amp sound smoother and more detailed is to have it drive a higher load impedance (if all other things were somehow equal, which they never are...).

Since he is no longer with us to defend his design, what do you think are the positive attributes of a differential output stage in a tube power amp?

@jaytor Since Lynn isn’t going there I’ll take this one. The advantage is the differential effect reduces distortion in the output section and makes the output section easier to drive since it will have a bit more gain.

There are a lot of differential output sections in well known tube amps- such as the Dynaco ST-70. What makes it differential is the use of a common cathode resistor. A Constant Current Source (CCS) can help performance but isn’t needed for the gain stage (whether an output section or not) to be considered differential.

Despite Lynn’s remonstrations, if designed properly a CCS in the output section of an amplifier will not limit current right up to the full power of the amplifier; in fact if the output section isn’t differential, using a pair of cathode resistors rather than a common one, the output power is unchanged or even reduced. I’ve seen applications where the use of the CCS actually increased the output power by a few Watts since the distortion was held in check to a higher power level.

What might not be obvious WRT an output section is you can set up the cathode circuit regulator to sense B+ variation and adjust the cathode voltage in response, which reduces distortion and increases tube life. This eliminates the benefit of a regulated B+ which would otherwise be a hefty lift in terms of execution and cost. IMO Lynn is missing a bet on this one and leaving performance on the table.

@alexberger The 6SN7 can support fairly high current which is why it can make a good driver tube if used properly. It can also make an excellent voltage amplifier (1st stage of gain) since it is quite linear if used correctly.

SETs usually do not need much gain since the speaker used with the amplifier should be high sensitivity (if a 300b power tube, +100dB is a good value). The 6SN7 will allow plenty of gain for this. One section can be used as the voltage amplifier and the other section the driver.

You should be able to get plenty of bandwidth using a 6SN7 and an interstage transformer! The issue will come down to the quality of the transformer itself.

Ralph, I sincerely invite you to build your own 300B amplifier.

@lynn_olson I did just exactly that some years ago and played the amplifier at a CES in the late 1990s. The 300bs were driven by a cathode follower direct coupled to the grids of the power tubes so the bias was obtained from the driver tube which in turn ran fixed bias. In this manner the capacitance of the grids was a non-issue.

How would you compare the sonics of your 300B amp to the the Class D amp you make now? You’re the creator of both, so you’re in the best position to evaluate and compare. I only spent a half-hour of casual listening to the Purifi at the show (and Audio Group of Denmark), so I’m hardly an expert on the subject.

Its been a long time so my comparisons have a set of our M-60 amplifiers in between if that makes sense. The M-60s were overall less colored by distortion with wider bandwidth and a greater sense of palpability. Peter Moncreif had us do a direct comparison at the show.

So the M-60 compares to the class d in that they have a similar tonal balance- with no grain or edginess in the mids and highs. The big tell that we hear and that customers report is that the class D is more focused in that its easier to hear what’s going on in the rear of the sound stage, pick out details and that sort of thing.

And if you really want to get hardcore, make sure all the cathode circuitry, of each section, comes down to a single star ground on the main ground bus-bar.

If I can add to this, make sure that the grid circuit and cathode circuit of each tube employ a single wire that goes to ground for both of those circuits. If you are using terminal strips, to do this the grid resistor and cathode resistor would tie to the same point and then a single wire to ground is used.

Tubes amplify differentially, which is to say the grid and cathode are out of phase with each other. So if a single wire is used for ground and noise is injected into that wire, it will be rejected by the tube. If you use separate wires the tube will be more noise susceptible.

McIntosh MC30 has the similar cathode follower driver 12ax7 to drive 6L6 directly. The output tube works in fixed bias in this schematics. Isn't it? Can you explain how the output tube bias is self adjusted?  

@alexberger 

Yes. R21 (120K; refer to schematic) sets the bias point of the 12AX7 (which to me seems a terrible choice for this application- a 12AU7 or 12AT7 would have suited better), which in turn sets the bias point of the power tubes.

The operating point it based on the idea that as the tubes weaken, at some point you just replace them, rather than readjusting the bias as the tubes age. The output power is not a function of the bias- its affected by the condition of the tubes! That is why the operating point was chosen to be class AB2 so the actual operational point is not critical and the power tubes will run cool with long life, high power and low distortion (due to the various feedback means). A good driver tube will present the output tubes with a very consistent bias voltage over time- that circuit is quite stable and has some ability to handle some grid current in the output section.

The bias rectifier and its power supply play a role in this. The original rectifier was a selenium device which has a larger voltage drop across it so if you renovate the amp and replace the rectifier with a new device (recommended- those old seleniums were terrible) this is a minor thing to pay attention to if you want to set the same operating point, although as I said its not critical.

From this distance, I wonder about circuit stability.

@lynn_olson They were quite stable and reliable. But the circuit is carefully designed around some serious issues with phase margin, which is why you really don’t want to try any serious mods other than updated coupling caps unless you do the math and think it thru.

The Dynaco ST 70 had good transformers, but the driver section was barely adequate and if you replace it with one of the more modern boards, especially the all octal tube ones, the amp is greatly improved. The small chassis limited the quality of the power supply you could fit in there though.

@donsachs The real weakness of the ST-70 is the rectifier. Its not got enough current and so is the most likely tube to fail in the amp. If you replace it with solid state the B+ will be too high. It is possible to use dual rectifiers but you have to replace the power transformer to do so. Triode Electronics of Chicago makes a power transformer that drops into place. You can use the filter can location for the other rectifier; the power supply caps can be installed beneath the chassis.

There are adapter sockets on ebay that allow you to run a 6GH8 in lieu of the 7199; that gets you a far more common tube that also runs the amp with a bit lower distortion.

@donsachs Unless the power transformer is replaced, I think its a bad idea to drop in a board using 6SN7s as much as I like that tube. 6SN7s have a 600mA filament; running 4 of them is 2.4 Amps! The transformer would burn up in due time.

Yes, you can install a supplemental transformer...As much as I like the ST70 (it has decent output transformers) IMO when you get that far into it, I feel like starting from scratch with a chassis that is large enough to really do what you want is a better move- and leave the ST70s for what they are good at: inexpensive and competent tube power.

Why is so much harder to drive 300B? 

@alexberger Mostly because of the voltage swing. RC coupled circuits are a bit inefficient if you also want current to deal with that input capacitance (which admittedly isn't all that much). That is why transformer coupling or better yet, direct coupling, does the job better.

Hi.. you are only running 3 6sn7 tubes.   Input tube and then a long tailed pair driver on each side using a single 6sn7....

@donsachs So 'only' 0.9 amps more... I think I've seen too many power transformer failures in my life, although that's mostly been due to bad filter caps. I treat old power transformers carefully- to avoid Bad Things happening.

The tube Don, myself, and many other manufacturers would like to see go back in production would be a 45.

@lynn_olson As best I can make out,  Sophia and EML both make a type 45.