diytube.com

Hey guys,

I don't like silicon in my tube amps. However, I'm about to go off to college, and I want to make my ST-70 as roommate-proof as possible. Not to mention the fact that this would just be really cool and would eliminate any need to rebias an amp.

My initial idea was to make a self-biasing amplifier. This isn't exactly fixed bias, per se. I was thinking about making a voltage source using a zener diode and a voltage divider to get between 400mV and 450mV. Then using a differential amplifier, a PNP transistor, and a resistor, one could effectively make a simple feedback circuit to make sure there is always the correct DC voltage on the grid. One input of the differential amp is taken off of the cathode resistor (10Ω in my case) and the other is the voltage source (400mV or so). The output is connected to a PNP transistor, and the emitter is grounded. The collector is tied to a resistor (which goes to the negative DC bias voltage) and the grid.

Could this work? I think it'd be a super simple op-amp based circuit. Has anyone done anything like this? Is there a better way?

Kyle

John Broskie of tubecad.com fame worked on an autobias circuit.

I have an interpretation that I used on a trioded EL36 that has near-veritcal curves in triode mode.

http://geek.scorpiorising.ca/GeeK_ZonE/ ... pic=1728.0

Cheers!

Lovely! I'm glad someone else has done this. It's definitely a little more complex than I had hoped, but maybe I can work on a simpler one. I might go try and put something together now. I'm just nervous about trying to use it with my ST-70 since I definitely don't want to blow anything...

Do I need to worry about this circuit at startup?

Also, why would this circuit really affect the tube curves? I would think if you capacitively smooth the voltage at the cathode resistor, it couldn't really change the bias voltage depending on what you have flowing through the tube, right? Am I making sense?

Kyle

Hi,

antiquekid3 wrote:It's definitely a little more complex than I had hoped, but maybe I can work on a simpler one.

If you can do it safely, be my guest!

This is the "Muntzed" version already. There's a commercial digital version out there from the 70's that makes Broskie's full version look like a 1-hour-effort.

I'm just nervous about trying to use it with my ST-70 since I definitely don't want to blow anything...

My rule in power amp building - add cathode fuses.

Do I need to worry about this circuit at startup?

No.

Also, why would this circuit really affect the tube curves?

Were did I say that? It's been 5 years since I played with it...

I would think if you capacitively smooth the voltage at the cathode resistor.....

.... and destroy the response time ;-)


Cheers!

You could also use a TL783 as a adjustable constant current sink on each EL34 cathode with a bypass cap. It is almost foolproof and once adjusted you can swap in different tubes and the cathode currents stay spot-on.

mesherm wrote:You could also use a TL783 as a adjustable constant current sink on each EL34 cathode with a bypass cap. It is almost foolproof and once adjusted you can swap in different tubes and the cathode currents stay spot-on.

But if the cathode currents are actively made to remain constant by a ss device in each cathode lead, how is any power output developed then? Does it hold the static (i.e quiescent setting) while the bypass cap allows for peak currents to be drawn? If so, how different is the performance of this approach to that of traditional cathode bias?

Dave

My bad -- read right over the word "sink" and assumed source. Need to slow down and read s---l----o----w----e-----r!

Dave

Geek, I think I misread what you typed; forget what I said about affecting the tube's curves.

This autobias would have to be done for every tube (all four), correct?

Okay, so the bypass cap allows for the AC fluctuations to pass, while the constant current sink takes care of the quiescent current, right? This seems as though it would change the sound much more than the autobias circuit.

Also, what would you do about the grid bias if you were to use a CC sink?

I would love to try the op-amp approach, but I'm not sure if I could do it by myself. Could someone help me figure out an autobias circuit that would work? I have lots of random op-amps laying around.

Kyle

http://home.earthlink.net/~mikesherman/images/PP-CC.jpg

This schematic demonstrates how i've used TL783's in the past on a 6V6 PP amp. The CC sink works like cathode bias using a cathode resistor bypassed with a capacitor. The advantage is that once the current is set it doesn't drift even if you change tubes. Same principle as using a regular 3 terminal regulator like a LM317 as a CCS except an LM317 is only good for 37 volts max whereas the TL783 is good to 125 volts.
I have used the same configuration in an EL34 SE amp and the current stayed rock solid once set. You don't need the fixed bias circuit at all. Just a grid resistor to ground just like a cathode bias scheme. As I noted in my 6V6 description, small clip on heat sinks on the TL783's were fine but for EL34s a larger heat sink is needed. On my EL34 SE amp I used TO-220 insulator kits to mount the TL783s to the chassis and that worked fine.
The chief advantage is that once set you don't need to worry about the bias anymore. I verified this in the amps I have built. I swapped the tubes around from socket to socket and even used different brands of tubes but the cathode current always stayed the same.
People never complained about the ST35 sound and ALL FOUR output tubes flowed through the same 95 ohm resistor bypassed with a 100mf cap.

Okay, that makes sense. Could I (since I have quite a few LM317s) have an extra series resistor to take off some of the load on the regulator?

If you were going to use an LM317 for an EL34 (at 40mA), how would you do it? Also, how big and what kind of capacitors should I use?

Kyle

antiquekid3 wrote:This autobias would have to be done for every tube (all four), correct?

Yup!

Okay, so the bypass cap allows for the AC fluctuations to pass, while the constant current sink takes care of the quiescent current, right? This seems as though it would change the sound much more than the autobias circuit.

The time constant is set so that any audible changes are in the infrasonic region.... naturally filtered out by the iron.

Also, what would you do about the grid bias if you were to use a CC sink?

As in mesherm's idea?

Just return the grids to ground.... it's cathode bias.

Mesherm's idea is wayyyy simpler to implement, but also limits power to what you's get with cathode bias.


Cheers!

Does it only limit the power because it is shaving off almost 40V or so from the plate? I doubt I would notice too much of a difference, as I really don't push my amp that hard.

Any ideas about how to implement my LM317s in my ST-70? I calculated approximately a 31 ohm resistor from sense to output to get 40mA. However, I would be pushing the regulator's max voltage, so can I just stick a 400 ohm resistor in series to cut the voltage and power across the regulator?

Kyle

I would not recommend using LM317s or any standard bipolar type voltage regulator even with series resistors (or zeners) taking some of the voltage. The TL783 is an MOS device with additional safeguards in place to prevent overload and accidental current set resistor failure.
The bypass cap voltage should be double the expected bias voltage. An electrolytic with a voltage rating of 100-150 volts should suffice. As to the mf value...one could do a lot of calulations based on capacitive reactance and the lowest expected frequency but I would swag 200-400mf would work. Electrolytics with 100-150 wvdc and 200-400mf range are not that big physically and not that costly either. Might be worth using Nichicon Muses even if you have some.
I think I used 200mf in my 6V6 amp and the bass was good down to 30hz.
The advantage with this is its maintanence free and tube newb proof. Once the bias currents have been set they will not drift and your PP amp will stay in balance even if the tubes are replaced or swapped.

http://cgi.ebay.com/TL783-High-Voltage-125V-Variable-Linear-Regulator-x4-/180482808923

Since the regulators are in fact acting as a constant current source, I would be very interested in what their use does to sustained power output. Music is basically an infinite number a transients, so the large bypass caps you specify could allow for most of that type of signal to pass on to the load. But high levels of softer more repeating signals like sine waves will act to increase the average current draw in the output stage of traditional fixed or cathode biased class AB1 amplifiers. In this case however, it would seem that the average current draw will not be allowed to increase. Have you measured the effect that the regulators have on the THD performance or sustained full power sine wave power output when the regulators are substituted for more traditional bias arrangements? It would be interesting to know.

Dave