Tube Preamp Paired with Tube Phono Stage?

05-20-12: Chervokas
The load impedance of a device should be at least 10X the source impedance of the device driving it to minimize insertion loss. Even at 10:1 you have a 1/10th loss of voltage IIRC.

That's close. It's actually 1/11th, assuming that both impedances are essentially resistive at the frequencies being considered.

Now if you plugged a mid output Benz with a 24 ohm output impedance and 0.8 mV output into the 100 ohm phono stage I think you'd have around 0.5 mV insertion loss (if I'm remembering the math correctly)

Assuming both impedances are essentially resistive at the frequencies being considered, you would lose:

0.8 x 24/(24 + 100) = 0.15 mv

Resulting in (0.8 - 0.15) = 0.65 mv appearing at the input of the phono stage. Which can be calculated a little more directly as:

0.8 x 100/(24 + 100) = 0.65 mv

All of which brings to mind a pet peeve of mine. When the specs for a cartridge indicate "impedance," what are they talking about? Impedance at zero Hz (i.e., DC resistance), impedance at 1 kHz? 20 kHz? Other? For many and perhaps most LOMC's it won't make much difference, because of their low inductance. But it would certainly be nice if DC resistance and inductance were separately and explicitly specified (which they sometimes are, but all too infrequently), so we don't have to make assumptions and guesses.

Regards,
-- Al

Al, thanks for the clarifications. I just installed an AT33EV at home and noted that the specs supplied in the manual were pretty thorough on those scores -- 10 ohms impedance at 1 khz; 22 uH inductance at 1khz -- but I gather that's not always the case.

I'm trying to learn the tech of cartridge loading so it's less trial and error. I get impedance matching and insertion loss. I get how you have to manage the resonant frequency of the RLC circuit with higher inductance pickups by being careful about capacitance loading (that resonant frequency never seems to get close to the audio range with low inductance pickups and any amount of real world capacitance you run into). I get the need to watch out for the cutoff frequency of low pass filters formed by output impedances and cable capacitances, though again that seems to be a rare real world problem. I'm not sure what other tech stuff comes into play. I guess the need for an RC network at the front of the phono stage with a cutoff frequency sufficiently low enough to reject RF noise but not too low to affect audio signal. What I don't quite understand is what, if any, disadvantages there are to lightly loading the cartridge. Like what are the problems that arise when loading a low impedance MC with a high impedance? Why would you want to load it down?

05-21-12: Chervokas
I just installed an AT33EV at home and noted that the specs supplied in the manual were pretty thorough on those scores -- 10 ohms impedance at 1 khz; 22 uH inductance at 1khz -- but I gather that's not always the case.

Yes, those specs are unusually well defined.

What I don't quite understand is what, if any, disadvantages there are to lightly loading the cartridge. Like what are the problems that arise when loading a low impedance MC with a high impedance? Why would you want to load it down?

The one reason I am aware of (although I'm not completely certain that there are no other reasons) is the one Ralph (Atmasphere) described. Heavier loading will reduce the magnitude of the ultrasonic resonant peak you referred to, which as you realize is caused by the interaction of the inductance of the cartridge and the capacitance of the cabling plus the input capacitance of the phono stage.

The greater the magnitude of that resonant peak, the greater the amplitude boost that will occur, relative to other frequencies, of frequency components that may enter the phono stage that are near the frequency of that resonant peak. Depending on the design of the phono stage, that may increase the likelihood that intermodulation or other effects caused by those ultrasonic frequencies will affect audible frequencies. Extraneous ultrasonic or RF energy may be present as a result of tics and pops, ultrasonic musical information that is on the record, RFI pickup, etc.

Some references that may be of interest:

http://www.hagtech.com/loading.html

And the post dated 8-14-10 by JCarr (Lyra cartridge designer) in this thread

Best regards,
-- Al

Chervokas, to be clear: if you load the cartridge with a lower resistance than your 10:1 or 11:1 ratio, all that happens at audio frequencies is you may reduce the output voltage of the cartridge. Its bandwidth will be unchanged.

So in theory you would want to load it as lightly as possible such that the RF issues are managed. The RF issues *can and do* act like changes in tonality. I admit that its confusing, but if you study this stuff on the test bench the fact of it comes out in spades.

Thanks, yeah, that's what I'm trying to get a grip on, if we're talking about a resonant peak way up above the audio band....maybe 600 khz a the low end or even pushing 2 Mhz at the high end; where presumably it's far enough above the audio band also to have limited if any impact on the phase performance of the audio signal, and even if it's basically undamped or at least certainly underdamped by the cartridge impedance loading (in many instances it seems like even 100 ohm loading would leave the ultrasonic peak underdamped), what precisely is the expected impact in the audio band?

I also realize that these aren't just electrical devices but also mechanical ones and the actual in-use performance isn't always perfectly predicted from trying to model the performance based on a circuit model of the device.

Chervokas, As far as an undamped LOMC goes, there is no effect whatsoever in the audio band, other than the voltage output- with no load its output will be a little higher, not much.

The issue is that the coils in the cartridge have so little inductance that the ringing issues that often affect inductive devices do not come into play until you are at very high frequencies, well into the MHz region. It is this ringing that supplies energy for the tuned circuit that results from the cartridge and the interconnect cable. Note that the capacitance of the cable always plays a role, the lower it is the higher the frequency of the resulting tuned circuit. So its rare that this can be solved by reducing the capacitance of the cable!

What is paramount, IOW, is that the phono section be stable with RF at its input.

Right, I get that, it's an RLC circuit and the capacitance is tuning the resonant frequency, and the lower the capacitance the higher the frequency (and the R is damping it).

So what does it take in terms of good design for the phono stage to be stage be be stable at RFs or to RFs -- I assume there's some kind of RC filter at the input of the phono stage that's a low pass filter typically, to reject RF generally, no? Just trying to understand. Thanks for your help.

Chervokas, one of the problems with RF is it can be rectified by elements of the preamp circuit itself. Tubes often tend to be a little less susceptible on account of transistor junctions can rectify the RF energy fairly efficiently if not properly set up. Tubes don't do that nearly so easily. Mechanical connections can also be a problem.

If the RF can get in at sufficient level, non-linearities caused by overloading the circuit can drive any preamp into conniptions, so the trick to do things that limit that; shielding at RF frequencies (aluminum is often better than steel for this), correct use of 'RF beads' that limit how much RF can enter the input connectors, proper 'grid stop' resistors in tube circuits ('gate stop' resistors in solid state gear), layout and grounding all play a role.

IOW its all in the design and the designer's attention/awareness of this as an issue.