I have never seen a single zener diode used to self (cathode) bias an amplifier. I was wondering if I may be missing something.
I want to bias my push-pull circuit at 6V. The 'Q' current of each pentode is 6 mA (12 mA combined), and when either side is at peak signal the current will be 65 mA. In my mind, a 6V 1 Watt zener should work as shown.
I want to bias my push-pull circuit at 6V. The 'Q' current of each pentode is 6 mA (12 mA combined), and when either side is at peak signal the current will be 65 mA. In my mind, a 6V 1 Watt zener should work as shown.
Attachments
I don't think this will work as desired.
The zener diode will behave like a voltage source. The DC current through the tube is not limited in this case. Therefore you shall see that the tubes have a thermal run away, easy to be seen in the red coloration of the anodes of the tubes. The bias circuits normally used for tubes are current sources for DC and short circuit for AC.
The zener diode will behave like a voltage source. The DC current through the tube is not limited in this case. Therefore you shall see that the tubes have a thermal run away, easy to be seen in the red coloration of the anodes of the tubes. The bias circuits normally used for tubes are current sources for DC and short circuit for AC.
Even with simple capacitor bypass, Zener is noisy.
Zener and Avalanche breakdowns blended for flat
(but noisy) voltage over a range of temperature.
Best to quiet that sort of voltage reference with a
more thoughtful filter. Aim source or emitter of so
quietly referenced P-sand at your cathodes.
Your circuit is balanced. If it stays always in class A,
might cancel some common mode noise. Not likely
to blow up if you want to try as drawn.
Zener and Avalanche breakdowns blended for flat
(but noisy) voltage over a range of temperature.
Best to quiet that sort of voltage reference with a
more thoughtful filter. Aim source or emitter of so
quietly referenced P-sand at your cathodes.
Your circuit is balanced. If it stays always in class A,
might cancel some common mode noise. Not likely
to blow up if you want to try as drawn.
Last edited:
It is fully possible that with fixed 6 V bias voltage the current of one pentode is 5 mA and 7 mA of the other, or the current of both is essentially higher or lower than 6 mA.
Therefore I would ask why don't you plan to use normal fixed bias with adjustable negative voltage and preferably combined with adjustable bias balance ?
Yes it is heavy class B. It's actually 7mA Q point now that I look at the load line again. I may move the Q point down to 4 or 5 volts and see how it sounds.
Attachments
Post regarding direct mains connection removed. Please read the forum rules. It's insanely dangerous, not just to the experimenter but to innocent bystanders.
Ofcourse I was talking about quiescent currents.
But thanks for reminding me how the pp- amplifier works.😀
There are 2 reasons.
1) I'm using an isolation transformer, so I don't have a low voltage secondary to create a bias supply.
2) It's a minimal parts count amplifier, using what I have available. I have 5 and 6 volt zeners, so I was wondering why it isn't commonplace to just put a zener from cathode to ground. You don't even need a cap, the zener will regulate the voltage as long as the current stays within the operating region of the zener diode.
A zener alone may not give satisfactory results, however, when paired with a transistor it works quite well. Current variation through the zener is reduced by a factor of beta, reducing the avalanche noise considerably. Since the impedance looking into the emitter is quite low there is generally no need for a capacitor bypass. A capacitor across the diode will reduce any hash created by the zener and swamp any CB capacitance of the bjt.
Attachments
I have used a zener diode string, a bunch of LED's, and an "amplified zener" as it was called in a Popular Electronics (I think) magazine article somewhere around 1970 (same circuit as shown in post by jerluwoo.
The zener works, the LED's work better, and the amplified zener works best. I prefer to use a separate zener (or alternative) in the cathode of each output tube.
The zener can be noisy, but a few millivolts of noise will not usually be noticed in an output stage. Zeners can be temperature sensitive, and you need to derate them by at least 1/2 their rated power. They do short out when they blow.
LED's are extremely variable. Some have a very poor knee, with a lot of slope in the VI curve, while others are quite sharp and steep. The only way to be sure is to test them. I found some camera phone flash LED's that worked great and lit up the whole lab. I used 3 or 4 in series for each EL84 in an SPP board.
The circuit can be made to resemble a neat perfect zener with a good high beta transistor. I have seen an adjustable version somewhere.
It works exactly like fixed bias, and thus the grid resistor must adhere to fixed bias rules.
The zener works, the LED's work better, and the amplified zener works best. I prefer to use a separate zener (or alternative) in the cathode of each output tube.
The zener can be noisy, but a few millivolts of noise will not usually be noticed in an output stage. Zeners can be temperature sensitive, and you need to derate them by at least 1/2 their rated power. They do short out when they blow.
LED's are extremely variable. Some have a very poor knee, with a lot of slope in the VI curve, while others are quite sharp and steep. The only way to be sure is to test them. I found some camera phone flash LED's that worked great and lit up the whole lab. I used 3 or 4 in series for each EL84 in an SPP board.
The circuit can be made to resemble a neat perfect zener with a good high beta transistor. I have seen an adjustable version somewhere.
It works exactly like fixed bias, and thus the grid resistor must adhere to fixed bias rules.
Attachments
Most of the world's fixed-biased amps generate their bias voltage from the same winding as the HT. It only takes a diode or two, a couple of caps and resistors. OK not quite as minimalist as a single Zener, but not far off.
Thank you for the responses
Thank you for all the responses, lots to think about.
The tubes are 60FX5, and I may get about 4 watts / channel.
This amp is more or less like the old kitchen table radios, it just simply amplifies my mp3 player or computer audio so I can hear it, it's not "hi-fi".
I'm building one channel on a breadboard so I can play around with different bias circuits.
I had considered doing a resistor voltage divider and a cap. Then the idea of a zener popped into my head to get and maintain the exact voltage. So as I was calculating currents I noticed the cathode current will always be in the active range of the zener, so why not just put the zener in the cathode path (which is why I posted this question).
I will probably use separate zeners, I have many small volt, 5 Watt zeners, so I will try different bias voltages to see how (if) the sound changes as I move the class A->B point.
I have read that in other forums, and since this is such a "low-quality" amp I wasn't to concerned about a little noise. If it does get noisy I will try the transistor circuit jerluwoo posted or just go back to a resistor in the cathode path.
Thank you for all the responses, lots to think about.
The tubes are 60FX5, and I may get about 4 watts / channel.
This amp is more or less like the old kitchen table radios, it just simply amplifies my mp3 player or computer audio so I can hear it, it's not "hi-fi".
I'm building one channel on a breadboard so I can play around with different bias circuits.
I had considered doing a resistor voltage divider and a cap. Then the idea of a zener popped into my head to get and maintain the exact voltage. So as I was calculating currents I noticed the cathode current will always be in the active range of the zener, so why not just put the zener in the cathode path (which is why I posted this question).
I will probably use separate zeners, I have many small volt, 5 Watt zeners, so I will try different bias voltages to see how (if) the sound changes as I move the class A->B point.
I have read that in other forums, and since this is such a "low-quality" amp I wasn't to concerned about a little noise. If it does get noisy I will try the transistor circuit jerluwoo posted or just go back to a resistor in the cathode path.
The trouble with exact voltage is that it does not adjust itself with changes in the HT or valve aging. Do you plan to fix the HT voltage too? If not then you're probably better off with good old cathode resistors.
I'm using an isolation transformer / bridge / caps, so my HT is always 166V unless my mains change.
If the zener is in the cathode path, then as long as the current through the tube is > 1mA and < 800 mA, the cathode should always be 6V higher than the grid.
If the zener is in the cathode path, then as long as the current through the tube is > 1mA and < 800 mA, the cathode should always be 6V higher than the grid.
Since your HT is not fixed, but you want a 'stiff' bias voltage since you're near class B, you might consider tracking bias. Just a thought.
http://www.diyaudio.com/forums/tubes-valves/274253-auto-bias-vs-manual-bias-vs-efb.html
- Status
- Not open for further replies.