• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

shared parts between channels

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,

Is it a common practice to share components in cathode bias circuits in integrated amplifiers ? What is the disadvantages of this (crosstalk?) ? Does it have an effect on stereo seperation ? Is it another main difference between monoblock (or dual-mono) and integrated amplifiers ? (the other difference is using seperate supplies for each channel)

MB
 
The main disadvantage is the need for matched quads instead of matched pairs. As the tubes age, there will probably be more imbalance in idle current and operating point. Crosstalk? Not for anything in the audio band, assuming you've bypassed the resistors in a competent way.
 
Ex-Moderator
Joined 2003
SY said:
Crosstalk? Not for anything in the audio band, assuming you've bypassed the resistors in a competent way.

I wouldn't be so sure. Some amplifiers shared cathode resistors and capacitors across channels in their pre-amplifiers. Equivalent series resistance of small electrolytics isn't zero, so it will cause some crosstalk, and once you've done that a few times, it all starts adding up. I didn't make a formal measurement, but I remember thinking that the crosstalk on such an amplifier wasn't all that amazing (and I'd just replaced all the capacitors!).
 
I ran a few quick numbers as a reality check. Back-of-the-envelope worst-case analysis stuff. My assumption (and maybe it's a bad one) was that the poster was inquiring about the bypass caps on the output tube cathodes. So, I'll continue down that garden path.

Let's assume a 470uF bypass cap. At 100 Hz, which is around where low levels of crosstalk could become sonically significant, the cap's impedance is about 3.4 ohm.

I'll assume a push-pull amp here, since cross-talk is the least of one's worries with single-ended. Now, in the operating region below the amp's idle current (the so-called Class A region), there's no AC voltage at the top of the cathode resistor, so no crosstalk. At higher powers, let's throw the number "100 ma" for a pair of small triodes or pentodes (like EL84s) as a reasonable max current. The resulting imposed voltage would then be something like 0.3V. The drive voltage to the other channel will be on the order of maybe 20V. So, the resulting crosstalk is roughly 20 log (0.3/20), or 36 dB down. As you say, not spectacular, but not terrible, either. I'd be delighted to have a phono cartridge that good.

As we move higher in frequency, that number gets even better. As we move lower, it gets worse, but that's also a region where localization by ear becomes less critical, too. And, of course, one can always go up in size on that cap, since it's a low voltage (in the world we travel in!) application.

For single-ended lower level circuits, you lose the push-pull advantage, but the voltages in these cathode circuits are so low that really big caps become quite practical. 4700u at 16V is a pretty reasonable cap. And, of course, the currents are much smaller in these stages, even taking into account the scaling of the voltage swings.
 
Ex-Moderator
Joined 2003
Yeuch.

SY: I accept all of your arguments, but I was thinking of single-ended stuff in pre-amplifiers where one stage might have a 470 Ohm cathode resistor shared between stereo channels. Just because a cartridge doesn't achieve much more than 25dB separation is no excuse for saving pennies by making an amplifier that is only marginally better. I think it's a horrible technique...

Why, a man who would share cathode bypass capacitors between stereo channels would probably drink wine that was so cheap that the bulk of the purchase price is duty, rather than spend a bit more and get something decent.
 
More envelope backs

Well, now that I've polished off this lovely bottle of Royal Magreb (Morocco's finest, produced by winemakers who have had their hands cut off, see my review of this wine on the Cull Page section of my website), let's look at that nasty little bypassed resistor. Low voltage, so you can get some decent amount of microfarads on the board. Let's say... 1000uF, for a nice round number. A 10 volt Panasonic at that value is about 1 x 1.2 cm and costs US$0.50, so this isn't unreasonable.

At the magic 100 Hz, this cap will run about 1.5 ohm in impedance (the ESRs are in the milliohm range so can be neglected). For a typical low level stage, the current will modulate at roughly a milliamp (less in very low level stages, but let's be generous) and the signal level is more or less 1 volt. So, the crosstalk voltage will be about 1.5 mV, which when referenced to the 1 V signal is better than -50 dB.

If that's not good enough, or if you really are concerned about separation at 20 Hz (from an integrated amp, yet!), a 4700u cap of the same working voltage runs US$1.34 in onesey-twosey, and is roughly 1.2 x 2 cm. The Nichicons run even less money and are similar in size.
 
diyAudio Senior Member
Joined 2002
WINE, MORE WINE...

Hi,

Errr....Where else can we cut corners?

Lessee....One PS for the two channels perhaps?

Ok, Ok....this is an INTEGRATED AMP, right?

Makes you wonder why on earth they invented stereo in the first place....:D

Cheers,;)

P.S. I still have both my hands...
 
"And why cannot you sign the papers, old man?"

Well, heck, Frank, I don't use cathode bypasses in my low level circuits, nor do I use cathode bias in my output stages. But... one of my favorite integrated amps of all time did use a common cathode resistor, so how can I complain?

Three guesses as to the amp.
 
diyAudio Senior Member
Joined 2002
Hi,

Three guesses as to the amp.

A Dyna with Moroccan bottles?:D

Makes you wonder what that little integrated could have sounded like with individual bias resistors and matched output bottles...

I think that in those days they couldn't care less about stereo separation given the limitations at the source already....
Besides, it did avoid that hole in the middle effect, didn't it?

Just havin' fun,:)
 
Well, with a Decaa 4RC and Quads, you've got the system that I used happily for quite a few years. The separation limitation was surely the Decca. By an order of magnitude, too, I'd expect.

I don't have to wonder, I converted more than one of those little beauties to fixed bias and 6CZ5 or 6973 outputs.
 
Ex-Moderator
Joined 2003
Re: More envelope backs

At the magic 100 Hz, this cap will run about 1.5 ohm in impedance (the ESRs are in the milliohm range so can be neglected).

The typical, rather nasty, capacitors that are used in an amplifier that cost-cuts in this way have significant ESR. I've just found a 100u 6V3 (which would be typical) having an ESR of 3.45 Ohm. (Rubycon MS5 series.) Assume gm of 2mA/V, giving rk of 500 Ohm, and we have crosstalk across the band at -43dB just for one stage.

By the way, FFSS (applied to Decca cartridges) stands for Fatally Flawed, Sometimes Superb. I had Deccas too... Excellent mono cartridges (if you got a good one), but stereo was a signal too far.
 
diyAudio Senior Member
Joined 2002
Hi,

No need for Quad IIs, the integrated was... integrated. ESLs, my lad, ESLs.

You sure have me puzzled there...:bigeyes:


Lemme guess....Full Frequency Stereo System?
I suppose they didn't have a dictionary on stereo systems in those days either...he,he...

but stereo was a signal too far.

Wouldn't even track a railroad....

Cheers,;)
 
Ex-Moderator
Joined 2003
Not quite; Full Frequency Stereo Sound.

Deccas could track, just not vertical stuff. And they needed a really robust arm. I designed my first "proper" arm with the Decca Gold in mind. It was an oil-damped unipivot and it worked, but end-of-side distortion was still a problem, and the vibration loosened the screws securing the platter bearing on my Garrard 301. The whole thing needed to be tightened every three months!

This is wandering way off topic - it's just as well no moderators are watching.
 
jlsem said:


What is so terrible about crosstalk? It occurs to me that it is much easier to measure than to hear.

John

This is true of most phenomena in audio. For ultimate performance, you want the cross-talk as low as possible. For audibly "blameless" amplification, you can tolerate quite a bit before it becomes audible.

EC: Just for grins, I measured the impedance of a 1000u/35V cap I happened to have laying around here. US$0.59 from Mouser, if memory serves. About 1 x 2 cm. Under 100 milliohms total impedance at 10kHz. Admittedly, I haven't measured it in a warmer environment yet, but...
 

G

Member
Joined 2002
metebalci said:
Hi,

Is it a common practice to share components in cathode bias circuits in integrated amplifiers ? What is the disadvantages of this (crosstalk?) ? Does it have an effect on stereo seperation ? Is it another main difference between monoblock (or dual-mono) and integrated amplifiers ? (the other difference is using seperate supplies for each channel)

MB

Steve Deckert of ZEN Triode Amp fame uses one cathode resistor and bypass capacitor shared between the channels of his SE amps. I tried it when I built my psuedo-Zen and I never changed back. It improved the imaging in my system. YMMV.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.