Hi,
I have a tube amp which uses c3m as the driver tube. The c3m is fixed biased but I found that the bias current vary significantly across different c3m tubes, even within the same brand/type. It's very difficult to find a c3m that can drop in and achieve the same bias current.
So what is the best way to adjust bias in such cases? The problem is that if two c3m tubes are used, one for each channel, and if they are not matched, then even if I somehow make them bias the same (current) their operating point will likely be different, which may lead to different gains, etc.
Am I understanding it correctly or is there any good way to this biasing problem?
Thanks.
I have a tube amp which uses c3m as the driver tube. The c3m is fixed biased but I found that the bias current vary significantly across different c3m tubes, even within the same brand/type. It's very difficult to find a c3m that can drop in and achieve the same bias current.
So what is the best way to adjust bias in such cases? The problem is that if two c3m tubes are used, one for each channel, and if they are not matched, then even if I somehow make them bias the same (current) their operating point will likely be different, which may lead to different gains, etc.
Am I understanding it correctly or is there any good way to this biasing problem?
Thanks.
Thanks for the suggestion. I think my amp doesn't use feedback - the c3m's cathode is grounded via diodes.
My thinking is that to adjust bias I can change the anode voltage but that will change the tube's operating point so not sure if doing that differently on the two channels will result in unbalances, gain or otherwise ...
My thinking is that to adjust bias I can change the anode voltage but that will change the tube's operating point so not sure if doing that differently on the two channels will result in unbalances, gain or otherwise ...
Change bias voltage moves the load line up and down. Change load resistor (anode voltage) rotates the load line. Either way you are changing the tubes' operating points, just differently. Isn't that what you want to do? Change it to balance and otherwise un-balanced tube.
John Snell Electronic, and DF96, I agree.
Start with individual cathode bias (with bypass capacitors), then check the gain of the two tubes (since there is no negative feedback, you can see how good or bad the gain match is). Write the gain of each tube down, and if necessary try more tubes, write down the gains, and find a match.
The problem with diode or LED bias method is that the bias voltage is roughly fixed to be the same voltage for many different tubes, but the current is not the same for different tubes unless they are matched.
The practical way to get a fixed diode/LED bias voltage & a fixed quiescent plate current is to use a current source in the plate circuit. This works real well for triodes. But it does not work well for pentodes, since pentodes are a current sources themselves. I believe the tube you are talking about is a pentode, right? Again, try individual self bias resistors, with individual bypass caps (your LED had low impedance, as a proper bypass cap would).
Start with individual cathode bias (with bypass capacitors), then check the gain of the two tubes (since there is no negative feedback, you can see how good or bad the gain match is). Write the gain of each tube down, and if necessary try more tubes, write down the gains, and find a match.
The problem with diode or LED bias method is that the bias voltage is roughly fixed to be the same voltage for many different tubes, but the current is not the same for different tubes unless they are matched.
The practical way to get a fixed diode/LED bias voltage & a fixed quiescent plate current is to use a current source in the plate circuit. This works real well for triodes. But it does not work well for pentodes, since pentodes are a current sources themselves. I believe the tube you are talking about is a pentode, right? Again, try individual self bias resistors, with individual bypass caps (your LED had low impedance, as a proper bypass cap would).
Last edited:
If the C3m is directly connected with the next stage you must check the bias point carefully so the selection is more complicated because the working point of the next stage is directly related to the point of C3m. In addition the selection of the tubes with the standard tube tester or with 2-3 different bias setting is not complete. It is much better (in addition) to select the tubes in a dinamic way sending different input signals so you can look at the tubes with the closest performances.
Walter
Walter
DF96, Alllensoncanon, 6A3sUMMER, waltube: Thanks for the suggestions
The c3m is capacitor-coupled to the next stage. The B+ is supplied by a shunt regulator using zeners; cathode and g2 are both diode biased.
My main concern is whether the c3m for the two channels will operate in the same or similar operating point for best inter-channel balance (gain or otherwise). I have various c3ms (Siemens, TFK) and I am quite surprised by the wide variations across different tubes. Perhaps these tubes are meant to be cathode-biased?
However, come to think of it even cathode bias will result in different operating points if two tubes are not matched so in the end it seems the only way is to first find matched tubes, and then adjust the operating point accordingly.
To 6A3sUMMER: You mentioned CCS as the plate load - in fact I thought about that too. Could you explain a bit more why CCS won't work with pentodes?
Thanks.
The c3m is capacitor-coupled to the next stage. The B+ is supplied by a shunt regulator using zeners; cathode and g2 are both diode biased.
My main concern is whether the c3m for the two channels will operate in the same or similar operating point for best inter-channel balance (gain or otherwise). I have various c3ms (Siemens, TFK) and I am quite surprised by the wide variations across different tubes. Perhaps these tubes are meant to be cathode-biased?
However, come to think of it even cathode bias will result in different operating points if two tubes are not matched so in the end it seems the only way is to first find matched tubes, and then adjust the operating point accordingly.
To 6A3sUMMER: You mentioned CCS as the plate load - in fact I thought about that too. Could you explain a bit more why CCS won't work with pentodes?
Thanks.
I am not familiar with bias servo. Do you have pointers where I can read up? Thanks.
There are a couple articles on the Tubecad site about bias servo's. However running these fixed bias is not really needed in IMHO. Cathode bias has the benefit of stabilizing the gain if i remember correctly. A 10 percent mismach in gain between channels might not be so worrisome anyway, the hearing is logarithmic.
Last edited:
CCS as load for a stage presents a “squirrelly load” to the situation: one that is very sensitive to operating point of each individual valve. It acts (for analytic purposes) as a huge resistor attached to a huge B+ supply, but with real 'clipping' limits at the top of its voltage dropping region.
Consider the case with a 10,000 volt B+ and a 12AX7 or other small tube. You definitely can set its operating point with cathode bias (or any bias method) to say 1 mA. And you can in turn calculate the precise anode resistor to drop all but 100 volts, 9,900 volts at that mA operating point!. Well, obviously at 1 mA that'd be R = E/I = 9,900 ÷ 0.001 = 9.9 MΩ.
(Using lab-bench higher voltage supplies, I actually tried this once. “Just to see”. With a 12AX7 too! It worked!)
So while you can find an operating point with the CCS-as-LOAD where your tube(s) will have an anode voltage “where you want it”, its also mighty likely to change over time as the particular valve in service ages.
Just saying,
GoatGuy
PS: interestingly, the “10,000 volt example” is also similar to using a very high value inductor (choke) as an anode load element for a stage's valve. It works by being “close to the ideal inductor” — very low DC resistance compared to very, very high AC impedance. Values of 250 Hy or higher are common in this model. Using
Z = 2πFL … F in Hz, L in Hy, Z in Ω then
Z = 2π (20) 250
Z = 31 kΩ … at 20 Hz
Z = 2π (1000) 250
Z = 1.6 MΩ … at 1,000 Hz
Z = 2π (20,000) 250
Z = 31 MΩ … at 20,000 Hz
You can see that the choke load acts as a “nearly fixed DC voltage CCS” for all intents and purposes.
Thought all y'all would like to see.
GoatGuy
Consider the case with a 10,000 volt B+ and a 12AX7 or other small tube. You definitely can set its operating point with cathode bias (or any bias method) to say 1 mA. And you can in turn calculate the precise anode resistor to drop all but 100 volts, 9,900 volts at that mA operating point!. Well, obviously at 1 mA that'd be R = E/I = 9,900 ÷ 0.001 = 9.9 MΩ.
(Using lab-bench higher voltage supplies, I actually tried this once. “Just to see”. With a 12AX7 too! It worked!)
So while you can find an operating point with the CCS-as-LOAD where your tube(s) will have an anode voltage “where you want it”, its also mighty likely to change over time as the particular valve in service ages.
Just saying,
GoatGuy
PS: interestingly, the “10,000 volt example” is also similar to using a very high value inductor (choke) as an anode load element for a stage's valve. It works by being “close to the ideal inductor” — very low DC resistance compared to very, very high AC impedance. Values of 250 Hy or higher are common in this model. Using
Z = 2πFL … F in Hz, L in Hy, Z in Ω then
Z = 2π (20) 250
Z = 31 kΩ … at 20 Hz
Z = 2π (1000) 250
Z = 1.6 MΩ … at 1,000 Hz
Z = 2π (20,000) 250
Z = 31 MΩ … at 20,000 Hz
You can see that the choke load acts as a “nearly fixed DC voltage CCS” for all intents and purposes.
Thought all y'all would like to see.
GoatGuy
Last edited:
Jack,
A small difference between the pentodes is common, even in the tube top manufacturer.
If you use C3m as a pentode driver, the simplest way to set Ua and Ik to be same value on both tubes is a slight change of voltage on G2.
You have zener regulated Ug2; insert small values trimmer pot into a series with zener diode and change Ug2 ...... or so
A small difference between the pentodes is common, even in the tube top manufacturer.
If you use C3m as a pentode driver, the simplest way to set Ua and Ik to be same value on both tubes is a slight change of voltage on G2.
You have zener regulated Ug2; insert small values trimmer pot into a series with zener diode and change Ug2 ...... or so
v4lve lover: Thanks for the pointers.
GoatGuy: Thanks for the detailed explanation. So the problem is that CCS may run out of anode voltage and cause hard clipping for certain tube characteristics.
Walter: What is minimum test set and how one select tubes in a dynamic way?
RajkoM: The difference I'm seeing is quite large. In my amp the correct voltage at the anode of c3m should be around 120V but with different c3m it range all the way from 30V to 180V as the bias current varies. That's why I have a hard time finding c3m that works in the circuit.
Moreover, the bias also seems to change as tube ages. I put in one with anode running at 130V at the beginning and recently measured only 30V. Hence I think eventually I may need to find a solution to this problem.
Your idea of adding a trimpot to g2 sounds like a good solution. Will probably try that in the future. Thanks.
GoatGuy: Thanks for the detailed explanation. So the problem is that CCS may run out of anode voltage and cause hard clipping for certain tube characteristics.
Walter: What is minimum test set and how one select tubes in a dynamic way?
RajkoM: The difference I'm seeing is quite large. In my amp the correct voltage at the anode of c3m should be around 120V but with different c3m it range all the way from 30V to 180V as the bias current varies. That's why I have a hard time finding c3m that works in the circuit.
Moreover, the bias also seems to change as tube ages. I put in one with anode running at 130V at the beginning and recently measured only 30V. Hence I think eventually I may need to find a solution to this problem.
Your idea of adding a trimpot to g2 sounds like a good solution. Will probably try that in the future. Thanks.
30 to 180V plate voltage variation for a simple resistor load is abnormal. What is connected to the Anode?
It's a CCS using IXYS current regulator follow by zener//cap, then anode resistor, and finally anode of c3m. Signal out is taped at the c3m's anode as usual and cap-coupled to the grid of the next stage (a triode).
Under the fixed bias voltages (cathode and g2 all diode-biased) different c3m conducts different currents so the voltage drop at the anode resistor (~ 70KR) caused wide range of voltages at the c3m's anode.
In extreme cases of low voltages the c3m would run out of voltage (to go down) and exhibit distortions.
Under the fixed bias voltages (cathode and g2 all diode-biased) different c3m conducts different currents so the voltage drop at the anode resistor (~ 70KR) caused wide range of voltages at the c3m's anode.
In extreme cases of low voltages the c3m would run out of voltage (to go down) and exhibit distortions.
What can I say that I already haven't?
You've got a current regulator in the anode supply. Whether its “feeding a zenner + cap” or not hardly matters. It'll introduce voltage-drop as it sees fit to regulate the current flow. Indeed, following it with a shunt zener regulator is a “band-aid on an unresolved problem”. Redesign the B+ power supply.
Moreover, when you say you have "G2 is diode-biased", what does this mean? G2 is nominally the “stable accelerator ("screen") grid”; it usually runs at a voltage varying from 40%-to–90%(common)-to–130%(very rare) of plate voltage. What voltage are you running yours at? It very well could be substantially higher (especially at VA = 30 V) than anode.
If you've not already done so, how about a schematic — with values and voltages — of the actual amplifier section(s) you are having us discuss. “A picture is worth a thousand words”
Respectfully just asking...
GoatGuy
Please post the schematic.
I have never seen a C3m. Sounds like it is more common to Europe.
Are you in Pentode mode? Or Tetrode mode (G3 tied to plate)?
There are different curves for each mode.
Lets look at one operating mode of the C3m:
6500 uMhos
Ip 16 mA
Is 3 mA
Rk 250 Ohms self bias resistor
19mA x 250 Ohms = 4.75V self bias
ri 250k
Screen 155V 3mA
Anode 225V 16mA 3.6W plate
ug2g1 19
26 second heater to get to 4mA anode current (very long warm up time)
An example stage using the specifications above:
Lets use a B+ of 325V
325V - 225Vp = 100V
100V/16mA = 6250 Ohms plate load resistor
There is a 6.5mA change of Ip for 1V change of Vg1.
6.5mA x 6250 Ohms = 40.625 The stage gain is 40.25 (lower than that because of the parallel load of the Rg of the next stage).
If you use a current source as the plate load; and LED, diode, or fixed bias, the plate voltage will go all over the map when you swap out tubes. This will not work well.
Even using self bias, and a plate load of a current source will have the plate voltage all over the map since the plate resistance, Ri, is 250k Ohms.
Note A. Operating the C3m with 325V B+, 6250 Ohm plate load, 155V screen, self bias cathode resistor 250 Ohms should nail down the plate voltage for good c3m tubes (in the ball park of 225V).
If the plate voltage is still all over the map, then I personally would not ever attempt using C3m tubes, too much variation for me.
Find a few that work good and are similar as operated in note A. Then if your circuit is different, but is another valid application of the C3m, it should work too.
I have never seen a C3m. Sounds like it is more common to Europe.
Are you in Pentode mode? Or Tetrode mode (G3 tied to plate)?
There are different curves for each mode.
Lets look at one operating mode of the C3m:
6500 uMhos
Ip 16 mA
Is 3 mA
Rk 250 Ohms self bias resistor
19mA x 250 Ohms = 4.75V self bias
ri 250k
Screen 155V 3mA
Anode 225V 16mA 3.6W plate
ug2g1 19
26 second heater to get to 4mA anode current (very long warm up time)
An example stage using the specifications above:
Lets use a B+ of 325V
325V - 225Vp = 100V
100V/16mA = 6250 Ohms plate load resistor
There is a 6.5mA change of Ip for 1V change of Vg1.
6.5mA x 6250 Ohms = 40.625 The stage gain is 40.25 (lower than that because of the parallel load of the Rg of the next stage).
If you use a current source as the plate load; and LED, diode, or fixed bias, the plate voltage will go all over the map when you swap out tubes. This will not work well.
Even using self bias, and a plate load of a current source will have the plate voltage all over the map since the plate resistance, Ri, is 250k Ohms.
Note A. Operating the C3m with 325V B+, 6250 Ohm plate load, 155V screen, self bias cathode resistor 250 Ohms should nail down the plate voltage for good c3m tubes (in the ball park of 225V).
If the plate voltage is still all over the map, then I personally would not ever attempt using C3m tubes, too much variation for me.
Find a few that work good and are similar as operated in note A. Then if your circuit is different, but is another valid application of the C3m, it should work too.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.