Question About Capacitor Upgrade in Tube Amp

I can see how adding too much capacitance to a regulator could cause the regulator to oscillate, but oscillation in tube amplifiers is rarely due to the power supply, and when it is, it's because the power supply has inadequate regulation (too high an output impedance), not the other way around. 

Yes, obviously changing capacitor values in the amp itself, and particularly in the feedback circuit, can cause an amplifier to oscillate. I spend a fair amount of time reading DIY audio sites (and building my own amps and preamps), and I don't think I have seen a case where adding a small amount of additional capacitance to a coupling cap or power supply cap ever caused an amp to oscillate. 

Tubes are capable of several MHz operation.

Oscillators are made of LCR and a device to drive them. The combination of the PSU, wiring - printed or P2P, transformer LCR and a tube to drive them can create a self-resonant circuit.

If one searches the DIY sites there are innumerable posts of capacitor changes turning tube amplifiers into oscillators.

@ieales - I'd be real interested to understand how adding some film decoupling caps cross the B+ supply could cause a tube circuit to oscillate. 

Prove it.

HF bypass caps are common in IC circuits as opamps can have multi-MHz bandwidths coupled with 100dB gain.

Tubes can also be multi MHz oscillators, but willy-nilly hanging a film cap across the B+ filter maybe just the ticket to convert an amplifier into an oscillator.

A simple answer. Yes, the small cap across the large el4ctrolytic cap reduces high frequency noise.

@ieales   Absolutely. Total agreement here. LC - LC forever!

But my reading of the OP is that the question concerns coupling and bypass caps. Are you seriously suggesting bypassing with electrolytics?

BEFORE worrying about bypass or blocking caps, ALWAYS ensure the electrolytic PSU filter caps are up to snuff. Way too many tube amps have inadequate primary filter capacitance. Dropping the 120Hz by 10dB is far preferable to 100kHz by a couple.

Too often they are under voltage, too high ESR, too small, etc.

"On SET amps the key here is not just Caps, resistors and etc but the trannies are the keys elements which will make the amps sound best. "

Not on SET amps - on all music reproducing or generating equipment. Thats like saying "The key on binoculars is having a good lens!", as if cameras, microscopes, etc don’t need that either.

The second most important thing to keep in mind is don’t let anyone say to you, "I doubt this change will matter because of X Y Z position or capacitance etc". They dont know what you have to begin with, and the part may be so cheap it causes a lot of damage to the audio because of its physical nature. Replacing it with audio-grade may make a world of difference, no matter where in the circuit it is. A Leica lens may have 20 glass elements which serve a different purpose in the schematic, but if glass #14 or #4 is made out of ikea plastic, replacing it will naturally cause a huge improvement.

Recently I had my eyeglass lens replaced with a very high grade material for 2x the cost, and while the Rx was still slightly off, the overall experience of contrast, detail, color, distortion, was far superior. The optometrist thinks only in terms of his narrow view of measurements, getting the focus to land correctly, which is important. But won't usually talk to about the myriad of esoteric enhancements a better material brings, even with the wrong Rx. Audio in electronics is similar, hence why many systems dont measure well or are accurate, but present special qualities of music.

@ieales

Agreed that those properties have not been mentioned. But which of them do you consider more important than DA? You seem to have left that thought unfinished.

You say, "OH, yes they should." But my reading of the OP is that the question concerns coupling and bypass caps. Are you seriously suggesting bypassing with electrolytics?

Of course modern electrolytics are better, but that is not the question as I read it.

I would also suggest that Dielectric Absorption (DA) is the most important measure for modding, because the other measures should have been considered in the design phase.

Nope. DA is not the most important by a long shot.

Of properties mentioned this far, left out are:

• Geometry, rolled or folded
• Lead material, length and attachment methods
• Foil material
• Foil inner / outer connection direction for DC blocking
• Microphonics
• Age

The properties you list are critical when using electrolytics, which are not (should not) be under consideration here.

OH, yes they should. Trying to clean up a marginal supply with a magic bullet is silly and rarely effective. A better supply can transform an amplifier.

Capacitor technology has improved immensely in recent years. More capacitance in smaller packages with reduction in ESR, ESL.

@calieng 

Some people believe that a capacitor with a certain capacitance value must sound the same as another with the same capacitance value regardless of the type of construction. That goes against common sense for anyone with a basic understanding of science or materials.

Goes against common sense.+1.

Charles 

@terry9 

Thanks for the additional info.

I was just making the point that there are physical properties to capacitors that can alter the sonic characteristics - but the cut and paste text was more related to electrolytic yes.

Some people believe that a capacitor with a certain capacitance value must sound the same as another with the same capacitance value regardless of the type of construction. That goes against common sense for anyone with a basic understanding of science or materials.

I would also suggest that Dielectric Absorption (DA) is the most important measure for modding, because the other measures should have been considered in the design phase. The properties you list are critical when using electrolytics, which are not (should not) be under consideration here.

You ask, "Does anyone know how the power supply bypass capacitor effects high end frequency response?"

Maybe. The component may sound smoother, with a less etched treble if the power is cleaned up with such a capacitor. This etched treble is distortion which has been removed. Some who are used to digital may not like the effect.

Capacitors of different construction types will have different sonic properties due to a number of factors (see below).

So as I understand it the properties of a silvergold in oil foil cap will be different than a copper foil in bees wax as they are impacted by voltage, current, and temperature.

As most with basic electrical circuit understanding would confirm, you can change the bass and treble response of a circuit by altering the capacitor value of coupling caps and cathode caps etc. Change from 0.1uF to 0.047uF coupling cap for example will reduce bass response.

So for just one specific example if the TC (temperature coefficient) of SGO Supreme cap of 0.22uF is different than Bees Wax cap of 0.22uF, then under operating temperatures in a hot tube amplifier their actual capacitance may be different and hence their bass and treble response will be different. How much of an impact that actually has or all the factors combined has is up to your ears to say. 

Some key properties of capacitors:

1. Nominal Capacitance (C). Capacitance refers to the amount of electrical energy a capacitor can store within its electromagnetic field. This value is represented in the unit Farad, including pico-Farads (pF), nano-Farads (nF), and micro-Farads (µF or sometimes uF for simplicity). Fixed capacitors have a specific capacitance that cannot be adjusted. Variable capacitors can be modified to achieve a desired capacitance within the available range of that capacitor.
2. Working Voltage (WV). The working voltage is the maximum amount of voltage a capacitor can receive continuously without damage or failure. Voltage can be DC (direct current) or AC (alternating current). The WV printed on the capacitor itself will typically refer to the DC rather than AC. The WV is affected by temperature, as the value listed only applies within a specific temperature range. Extreme heat or cold may affect the working voltage a capacitor can withstand.
3. Tolerance (±%). The capacitance value listed for a capacitor can sometimes vary more or less. The value can only vary by a certain range to be accepted, which is its tolerance. Tolerance can vary anywhere from real low tolerance like 1% all the way up to part with a -20% to +80% tolerance. Common tolerance values are 5%, 10% and 20%, this of course varies based on the type of capacitor it is. Capacitors are ranked in quality based on their tolerance. The lower the tolerance, the closer the actual capacitance is to the value listed by the manufacturer, thus the higher the quality of the capacitor (and oftentimes more expensive as well).
4. Leakage Current. A capacitor contains a non-conductive material known as a dielectric. The dielectric will typically allow a small amount of electricity through, referred to as leaking. Leaking happens because of the strong electromagnetic field that exists between the plates when voltage is applied. Extremely low leakage in a capacitor, such as in a film or foil type, is said to have a high “insulation resistance” (Rp). High leakage, which is more typical in electrolytic capacitors, is referred to as “leakage current.”
5. Working Temperature (T). Temperature affects a capacitor’s ability to store electrical energy. For example, extremely high temperatures can cause a liquid electrolyte in an electrolytic converter to evaporate and change the capacitance. In contrast, extremely cold temperatures could cause liquid or gel electrolyte to freeze and impact its capacitance.
6. Temperature Coefficient (TC). The temperature coefficient measures the change in capacitance that could occur within a particular temperature range. As temperatures rise, some capacitors increase their capacitance values and are considered Class 2 capacitors, such as this one GRM155R71C104KA88D, and others decrease their capacitance value. Capacitors that are able to maintain their capacitance within a temperature range are considered to be Class 1, like this one CC0402JRNPO9BN101. The importance of temperature when it comes to capacitance depends on the job at hand. If you know that temperature may be an issue, the TC is an important characteristic to be aware of.
7. Polarization. This refers to the charge of the plates within a capacitor. In most capacitors there is a positive end and a negative end, similar to a battery. When applying voltage it is necessary to match positive voltage with a positive terminal, and negative voltage with a negative terminal. Incorrect polarization can lead to severe damage within the capacitor and throughout the circuit and device.
8. Equivalent Series Resistance (ESR). This is a term for the total resistance of every part of a capacitor that resists, rather than conducts, electric current. It includes the resistance of the plates, the dielectric, the terminal leads, and the connections to the dielectric. ESR is the sum of all of these, measured within a specific frequency and temperature. This determines energy loss for a capacitor.

Upgrading capacitors is a no-brainer. Ditto resistors, wire, tubes, sockets, transformers, connectors, layout, etc. 

Expecting 'warmer mids' or 'silky highs' when installing Caps-4-Bux or any other component change in myriad divers applications is hubris.

The lack of rigor in accommodating environmental and CBLF changes negate generalized comments.

The expectation that the sonic effect in one device will translate to all others is pure fantasy

Yet experienced listeners who try better regarded capacitors report consistently of improved sound quality. I don’t believe that this is coincidence. Upgrading coupling capacitors will far more often than not improve sonic performance. I believe what they say they hear.

Charles

there was an improvement in frequency response and lower distortion replacing the cheap factory caps with one of better and different material construction.

Did you measure the C or look up the ESR and ESL of the caps or just read the value printed on the schematics / old cap?

Did you measure the frequency response, distortion, phase shift, transient response, etc. or just listen?

@calieng 

Does anyone know how the power supply bypass capacitor effects high end frequency response?

Capacitors have properties Capacitance, Equivalent Series Resistance [ERS] and Equivalent Series Inductance [ESL], all of which are frequency dependent. Dielectrics, foils, leads and geometry affect the frequency response of these properties. At some frequency the combination of these properties will make the system self-resonant. Hence a band limited digital system may exhibit different response when fed by a disc system with much higher bandwidth.

Most think the capacitors just filter the DC from the rectifiers. However, current flows in @ 60/120Hz [50/100Hz] and out of the capacitor bank @ DC to many kHz.

Add in additional effects of construction, circuit design, wiring, printed or otherwise, program, level, related equipment, environmental conditions and universality approaches ZERO.

The expectation that the sonic effect in one device will translate to all others is pure fantasy.

And while the Willsenton at around $900 plus shipping might seem to be a cheap amp compared to other equipment discussed on here I do have to say that it sounds on a completely higher level than my Luxman or Naim soild state amps when paired with the Forte IV speakers. Like sitting in the front row of a jazz club. The music surrounds you and you can pick out the location of all the instruments.

I don’t doubt this observation at all. SET if properly implemented will connect you with the core of the music. Exceptional naturalness and realism  in my listening experience.

Charles

@alexberger 

Yes the Mundorfs seemed to be a best bang for the buck. I found an online dealer selling them around $30-$40 each depending on the value for SGO Supremes.

And while the Willsenton at around $900 plus shipping might seem to be a cheap amp compared to other equipment discussed on here I do have to say that it sounds on a completely higher level than my Luxman or Naim soild state amps when paired with the Forte IV speakers. Like sitting in the front row of a jazz club. The music surrounds you and you can pick out the location of all the instruments. Mind you that is with Lumin U1 and Denafrips Venus II front end.

I firmly believe that we should support products from USA, UK, and others as opposed to from some other locations with questionable governments but if you are looking at buying products like Line Magnetic, PrimaLuna, Cayin, Icon Audio....The Willsenton and Muzishare can also be considered from the same factories and for less than half the price. Maybe just benefiting from a few capacitor upgrades as I have done.

People should also be aware that dealers from this country of origin can be pretty sketchy so for some the extra cost of a USA based dealer with the other brands is well worth it....but that is another discussion entirely.

Hi @calieng ,

You did right thing to chose relatively affordable Mundorf capacitors for your Willsenton R300 amplifier.

Duelunds and V-Cap CuTf are very good but they are overkill for such not expansive amplifier. With all respect coupling and power supply bypass capacitors shouldn't be the most expansive part in any amplifier.

Regards,

Alex.

Based on the good results with the Willsenton R300 I went ahead and upgraded the Muzishare X7 KT88 amp. It required four 0.22uF coupling caps and one 0.47uF filter cap bypass film cap.

I used the Mundorf SilverGold Oil 1000VDC again.

Similar results. Most noticeable change - the depth of the soundstage significantly increased and more holographic image. Slight increase in soundstage width. Better treble crispness and rounder fuller bass with better instrument separation.

Upgrade cost approx $200.

I am using Gold Lion KT88 and 5AR4 tubes with Mullard reissue 12AX7 and 12AU7s in the Muzishare.

Between the two upgraded amps, the Willsenton R300 has the edge in vocals and holographic sound. The Muzishare X7 has the edge in treble response and bass punch. Both are wonderful amps with the upgrade.

And if there was any doubt - after doing this for a second time, good coupling caps make a significant improvement in sound over the factory cheap caps.

I peeled the labels off both the factory branded Willsenton and Muzishare coupling caps and there is no label underneath. So the source of the factory caps is unknown. Although the Muzishare factory caps were much larger than the Willsenton with what looks like a better construction.

And once again the Willsenton R300 comes from the PrimaLuna/Cayin factory (Zhuhai Spark Electronic Equipment Co, LTD) and the Muzishare from the Line Magnetic factory (Zheng brothers Zhuhai factory) so other components in the amps do appear to be decent and not requiring an upgrade (Nichicon electrolytic, Alps potentiometer etc.)

@jaytor - Thanks for the reply!

@calieng - the primary purpose of the bypass cap on the power supply electrolytic is to shunt high frequency noise on the B+ rail. I doubt it will have any affect on the high frequency response of the amp. The value of the cap is really not that significant as long as it is a good quality film cap. 

The biggest influence on the high frequency response will be from the output transformer and possibly from the feedback network. I wouldn't recommend changing these parts without a solid understanding of how the amp works as well as having adequate test equipment to make sure the amp is stable after the changes.

Using a higher quality coupling cap also might give you a small improvement in high frequency response if the one that is in there now is not of very good quality.  

Does anyone know how the power supply bypass capacitor effects high end frequency response?

For example the Willsenton R300 amp uses a .33uF and is a slightly rolled off sounding amplifier while Muzishare amps are using .47uF on the filter cap and are brighter sounding amplifiers. PS audio apparently uses a .1uF

Is this power supply bypass cap impacting high frequency response? Or are the bass and treble response all in the various coupling caps and cathode caps etc in the preamp stage?

Thanks.

NOS Allen Bradley Carbon comp resistors are all the rage for tube guitar amps. For their warmth and organic tone. I wonder how they fair in hifi? They do tend to drift in value a fair bit. So maybe not so good in hifi setting.

Carbon comp resistors are very noisy and add distortion. This may be (is likely) desirable in a guitar amp, but not what I want in my hifi amp. 

And for really low values, less than a couple of ohms, nichrome wire just can't be beat.

The nude Vishays are nice, use them in my 300B amps.

Interesting.

NOS Allen Bradley Carbon comp resistors are all the rage for tube guitar amps. For their warmth and organic tone. I wonder how they fair in hifi? They do tend to drift in value a fair bit. So maybe not so good in hifi setting.

The Vishay Z-foils are very transparent and low-noise. These resistors can be ordered in arbitrary custom values from Texas Components. The latest version is the TX2575. These are only available in sizes from 10 ohm to 100Kohm. The older TX2352 is available from  1 ohm to 250K ohms. 

For higher power requirements or resistors outside these ranges, I like the Audio Note non-magnetic tantalums and silver-tantalums, but these can get pretty pricey.

Resistors make a difference too. I tested half a dozen candidates with a rotary switch in my preamp. Nude Vishay resistors were the best, no-name metal films from Taiwan second, aerospace hyper-expensive last.

Each type of capacitor has a specific frequency response and voltage related distortion.

It should stand to reason that better constructed capacitors should have a better frequency response and less distortion. Whether you can hear the difference in reality or it is simply a fresh capacitor not being run in yet that sounds different I do not know for sure. But it seems in my case there was an improvement in frequency response and lower distortion replacing the cheap factory caps with one of better and different material construction.

In this article the title seems to take the opposite opinion but have a read thru and there is some information that at least confirms capacitor construction does impact sound quality.

I have dorfs in my 300b amp does it sound better because of them? maybe. 

@jaytor And most important, these amps have LETHAL voltages inside. Capacitors can hold a lethal charge for a long time.

So I should have been careful soldering that bypass cap on the 330uF 450V Nichicon filter cap from hot to ground terminal?

Oh dear maybe that is why it was so exhilarating doing the upgrade. A real exciting experience.

😀

But seriously that is good advice. Use the sorting cable with a suitable 10W resistor to discharge electrolytic caps first.

@eworkflow 

Agreed. 
Completely unnecessary obnoxious and immature forum behavior.

Charles

@ieales I don’t know what social setting you acquired your accusatory tone from, but your ridiculous accusations of me being irresponsible have no merit. First, I wouldn’t suggest changing it to anything different nor ever stating that capacitors do not affect sound in the audio path. So, in short, I am "irresponsible" for saying that the primary role of low-capacitance capacitor shunting DC power supply is to filter HF noise. Secondly, please present a literature source to confirm your "power supply impedance and phase" being audibly affected by a low-capacitance shunt capacitor or just shut up. I am well aware of the fact that power supply quality will affect sound, but you are taking your holier-than-thou mission too far.

As for capacitor used to shunt DC power supply, they are necessary to eliminate high-frequency noise present there, but never affect the sound quality as much as a different coupling (serial) capacitor used in the audio path.

@eworkflow Are you an engineer? Do you have a test lab? Verified the bypass capacitor values actually sink just the noise and haven't changed the DUT frequency / phase response?

'Noise' is not a constant and changing capacitor value / type may actually make things worse.

Changing type / value changes the power supply impedance and phase. Whether or not it is an improvement may be dependent on the DUT, the rest of the system AND the CBLF*.

Making willy-nilly recommendations is irresponsible.

*: the listener with all its inconsistencies.

I'm also a strong believer in using high quality film caps. I've had good luck with V-cap ODAMs (very neutral and clear) and Miflex KPCU (a bit warmer with a beautiful midrange tone). 

A couple things to keep in mind. These caps all take a few hundred hours of break-in to sound their best. And some will actually sound worse during the break-in period. 

And most important, these amps have LETHAL voltages inside. Capacitors can hold a lethal charge for a long time. Make sure you are thoroughly familiar with the safety procedures before playing around inside a tube amp. The OP mentioned doing DIY guitar amps so is probably already familiar, but anyone else contemplating doing this kind of upgrade should use extreme caution. 

Capacitors have their own technology, dynamics and very often audible characyeristics. Anybody who disagree most likely never replaced one in the audio path of their high-quality equipment piece.

As for capacitor used to shunt DC power supply, they are necessary to eliminate high-frequency noise present there, but never affect the sound quality as much as a different coupling (serial) capacitor used in the audio path.

Good for you if the Mundorfs work for you long term. Every time I've tried Mundorfs, initial impressions good, over time always heard spot lit high freq, rather artificial sound quality result over longer term. Hopefully these work for you long term, caps exist for all needs.

There was a very very slight smoothness added in the very upper treble after about 20 hours run in versus very sharp and defined treble notes when caps were first installed. 

Just last night I installed NOS RCA rectifiers and preamp tubes. So it will not be possible to report back any more cap break in changes. The resolution of the music increased along with the smoothness and musicality with the NOS tubes. I also tried new production Tung-Sol but they sounded gritty in comparison. Maybe new tubes need to break in as well so I am running them in another amp for now.

I am waiting on a delivery of PSVANE 300B TII matched pair power tubes as the last upgrade.

This is THE amp for my Klipsch Forte IV speakers. A perfect match.

Would I say it is the best amp I have in every case - no. The KT150 amp is better with my Dynaudio speakers for example. So system matching is still more important than these other tweaks.

Lots of great recommendations and suggestions on upgrading.  I for one really believe in it.  I went from Janzens to Duelund copper CAST coupling capacitors in my tube-preamp and at first, I thought not such a good upgrade, but after ~100-hours of break-in, wow...  Increased inner detail and for lack of a better term, body in the music. My good friend has a pair of Welborne 300-B SET's that we're going to do a component upgrade to sometime next year.  They sound surprisingly great for their cost and what they are.  Simple circuit and all...  I believe they will sound notably better with these upgrades.  As many have said, please don't judge the switch until you've given some time for break-in !  

In a break-out box, I have compared V-caps with Solen teflon f&f  and Relcap teflon f&f. The latter are good and quite similar IMO, and the V-caps are bright. To my ears, in my system, in my comparison. YMMV.

i personally would not spend so much on caps upgrade unless i know what trannies that are fitted on the unit.  

The amp comes out of the same factory a Cayin and PrimaLuna. I have seen it reported that there is some amount of sharing of parts. The transformers are indeed decent.

For the Muzishare amps they are made in the Line Magnetic factory. And once again some comments online and photos indicating identical components being used.

Electrolytics are Nichicon so just for the Willsenton they cheaped out on the coupling capacitors.

Alex 

Does anyone compared directly:

1. Vcap CuTf vs Mundorf Supreme Silver Gold Oil?

Vcap Cutf is in the class by itself if untilizing in electronics application. If you are using it for speakers then the Mundorf SGO is a much better cap on speakers. 

On SET amps the key here is not just Caps, resistors and etc but the trannies are the keys elements which will make the amps sound best.   To me it doesn’t make sense to put a $500 caps ( duelund)  on a $1k amp, You are not going take or make that amps sound like a $10k amp. Yes improvement, for sure there would improve compare to the original parts but to what extent, cost and time?  So to summarize it up i personally would not spend so much on caps upgrade unless i know what trannies that are fitted on the unit.  

Yes given the cost of the map I did not want to spend too much on the upgrade but honestly I do not think it could sound any better. Bass, middle, and treble are all clear and in balance with each other. 

I am sure all the other caps recommended would have made a significant improvement as well. Maybe with just slight variations in bass or treble response. But the difference from the original cheap factory caps was much more than I had expected.

Vcap CuTf are great for 300B SET - clean, clear, transparent, fast, soundstage not harsh at all.

Duelund Custom are excellent too - great tone harmonics, soundstage, a little bit slow and full bass.

Bu I think these two options are too expansive for such not expansive amplifier.

Does anyone compared directly:

1. Vcap CuTf vs Mundorf Supreme Silver Gold Oil?

2. Duelund Custom Cu vs Cu-Sn versions?

@calieng 

Congratulations with your positive outcome. It seems that the stock coupling capacitors were a sonic bottleneck. I’d say that was 100 USD well spent.

Charles

sounds like much more than an “ initial impression “….