• 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.

Origin of the so-called "Blumlein Garter"

Just be sure to use Rg values that are quite a bit less than the tubes specified values.
As DF96 mentions, this circuit trades off stability.
Be careful, or this configuration will have thermal runaway to the max.
Also, if one tube goes gassy, it may very well cause the other tube to go bad too.

Do not forget that even though the currents are DC balanced even if the tubes transconductance is not well matched, the AC currents will not be balanced. Hmm, this asks for balanced tubes, which is what the circuit was supposed to do (only does DC balance, not AC balance).

"You should make things as simple as possible, but no simpler" A. Einstein.
In my experience, if instead you use reasonably well matched tubes and individual self bias, the simpler circuit will have very well balanced DC currents, very well balanced AC currents, and no thermal runaway.

I have a vendor that re-tests tubes very carefully, and at a good price too.
Go the route of well balanced tubes, and save parts count.
 
Last edited:
I've had nothing but successful results from using garter bias on my builds, and have lately gone to lower value grid resistance just as a precaution. Even with obviously gassy tubes I've yet to encounter any sort of thermal runaway.

Either way two extra resistors per channel is almost not even worth mentioning as a downside, but for most builds standard individual cathode resistor setups work fine for most setups anyway. It just depends on preference really.

Now that I think about it, since the grid resistance is tied to sort of a "virtual ground" rather than a fixed reference, would it be considered bootstrapped, the same way that a cathode followers or concertina splitter is when set up for self bias? If so, one could safely use a much smaller grid resistance without hurting input impedance much. I could be wrong though, of course :)
 
In the Garter circuit, the series resistors in the cathodes are bypassed with capacitors.
Therefore, there is no appreciable motion of the cathodes with signal applied.
That is not a bootstrap for the grid resistor, Rg.
The driver will 'see' the grid resistor without the benefit of bootstrapping.
 
None of you are newbies, but for their sake: Be sure to put in the grid leak resistors too. And if the one tube is gassy, or if the grid leak resistors are too high of a resistance, then we can remember the old lyric "run run run run runaway".

Incomplete schematics like the one at post # 3 get the newbies in trouble.

I am sure all of what I said now has already been said on this thread, but I started reading it today from the beginning. Could a newbie do that too?

As far as having balanced currents in the output transformer, you can use a couple of current sources for the cathodes with a couple of bypass capacitors. Oh, wait, that may not be true for tetrodes, pentodes, and beam power tubes. That is because the screens of two tubes do not necessarily draw the same current. Oh, are we back to either matched tetrodes, pentodes, beam power tubes. Or you can triode wire them, and the cathode current will equal the screen plus plate current. Or more simply a pair of unmatched triodes (they will have the same quiescent current, even though with signal the unmatched nature of the two will show up). Hmm, I guess there is a reason to use matched tubes.

Sorry for resurrecting an old thread but I don't understand the grid leak caveat. Are not the resistors from the grid to the junction of the cathode resistors providing the necessary ground path?
 
Sorry for resurrecting an old thread but I don't understand the grid leak caveat. Are not the resistors from the grid to the junction of the cathode resistors providing the necessary ground path?

Yes, the 100K grid resistors ARE the grid leak.
So the diagram in post #3 is not incomplete in that sense.

And yes, thermal run away of one tube - once started - will also affect the other one.
On the other hand, it may be less likely to start run away in the first place because there is twice the cathode degradation in place. This might explain why it is rarely observed.
Also, if one tube goes bad one would most likely replace the other tube anyway ...
 
Sorento,

Referring to Post # 3 Schematic:

Cathode Self Bias Degeneration is according to the 200 Ohm resistor from the cathode to where the 100k grid resistor connects.
So, the bottom 200 Ohm resistors do not cause any Cathode Self Bias Degeneration.
Only the top 200 Ohm resistors (that connect to the grid resistors) cause Cathode Self Bias Degeneration.

Strictly speaking, the Garter Bias circuit is a Poor Man's attempt to take tubes that are not well matched, and through the "magic" of the Garter bias circuit
'sort of' make the tubes match.
 
Last edited:
Strictly speaking, the Garter Bias circuit is a Poor Man's attempt to take tubes that are not well matched, and through the "magic" of the Garter bias circuit
'sort of' make the tubes match.

This is not a poor mans way of emulating a matched pair. This is an extremely clever piece of Engineering design to make a circuit more stable and robust.
A pair of matched valves is an intrinsically unstable situation because those valves will age differently and the match become poor over time with less chance of a degraded valve been "restrained" by its linked partner.

Its a shame that experienced builders cannot see the intrinsic brilliance of Garter Bias. What is not to like about a circuit which is intrinsically stable over long periods of operation and with just about any valve's you throw at it. I am totally against the need to adjust bias on a near weekly basis to get a valve design to give of its best.

Shoog
 
Last edited:
Sorento,

Referring to Post # 3 Schematic:

So, the bottom 200 Ohm resistors do not cause any Cathode Self Bias Degeneration.
Only the top 200 Ohm resistors (that connect to the grid resistors) cause Cathode Self Bias Degeneration.

Yes, but ... nevertheless there seems to be a small effect which mitigates the increase of plate current.
Here's a little spice "experiment".
To emulate a tube going into runaway, the left one is being replaced by a much stronger tube.
(Ayumi models for 6CB5 / EL84 and 6CA7 / EL34 used)

First: 2x EL84 in triode, 470ohm single cathode R, 300V B+
2nd: EL34 / EL84 in triode, 470ohm single cathode R, 300V B+
3rd: 2x EL84 in triode, 2x 470ohm garter, 312V B+ (to compensate for 12V drop in 2nd R)
4th: EL34 / EL84 in triode, 470ohm garter, 312V B+ (to compensate for 12V drop in 2nd R)

Result:
plate current rises from 25.9 mA to 44 mA without garter bias
plate current rises from 25.9 mA to 38 mA with garter bias

The effect is not big but in the right direction, mitigating the rise in current.

It also shows that the other tube suffers too when one goes bad.
 

Attachments

  • Draft1.png
    Draft1.png
    13.7 KB · Views: 138
  • Draft2.png
    Draft2.png
    14.3 KB · Views: 147
  • Draft3.png
    Draft3.png
    15.3 KB · Views: 139
  • Draft4.png
    Draft4.png
    16.1 KB · Views: 154
Last edited:
For years, I have used individual self bias, and individual bypass caps for my push pull amplifiers output tubes.

I went to the expense and trouble to use well matched tubes (I have a vendor that is Very Good at that).

After 100s, and even 1,000s of hours operation, the tube quiescent currents are still very well matched.
No Garter Bias required (but I did start with very well matched tubes).

That is just my mileage.

Your mileage may vary.