Tube VCA PSpice voodoo!!???

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Learner

Well-known member
Joined
Jun 4, 2004
Messages
341
Location
Area 621.xxx
Hi guys,
I've been playing with the idea of building a tube VCA by controlling the cathode current using transistors to control the gain, here is what I've come up with and it seem to simulate ok..........
I am very skeptic about the result of the simulation due to previous voodoo results that I have got from it, it has a bit of distortion at the output but by lowering the plate resistor value that seems to help a bit. Does anyone know if this thing will actually work at all!???

I am looking to use the transistor array chips such as the THAT 340 or the SSM2220/2210, donno if that will make much difference than using general NPN and PNP transistors.....

tube1_copy.jpg
 
all i can say is that if you think it's going to work and the simulation says it might, then break out the protoboard and try it.. :green: that's the only way to know for sure!

let us know how it works!
 
Acutally I am but the problem is that I don't have 6l6 tubes and I am not really interested to buy some 6l6 tubes, because I have a small stockof 6SK7 and 6BA6 which is what I am looking to try it on. The 6BA6 and the 6SK7 are pentodes and I am planning to wire the supressor grid straight to the cathode, I just want to get some opinion before I wire it up.

Thoughts people?
 
> a tube VCA by controlling the cathode current using transistors

So where is your control voltage?

U1 Q1 force 1mA. U2 Q2 force 15V, but only have 1mA of grunt. What is the voltage and current at the 6L6 cathode?

You can make a perfectly respectable VCA with a 6L6, grounded cathode, triode-strapped, and vary the grid bias negative. You could do the same by DC-grounding the grid, forcing the cathode positive. For maximum gain-range, start from a high current like 50mA. That means plate resistor of 2K or less. For 1V peak output, you can force the cathode current down to maybe 1mA, gain change of about 7 or 17dB. There will be 70V of DC-shift from max-gain to -6dB gain, quite a bit compared to 1V peak signals.

When designing a VCA, you have to start with a basic question: will it be constant output with varying input, or constant input with varying output? The former is what we want for an AM radio, a limiter, or for bringing various inputs to mix-level. The later is the case when playing CDs loud or soft, or controlling the output of an analog synth (is my age showing?). Unfortunately, this plan (like any similar plan, including the beloved "vari-Mu") is really best suited to constant input with output varied from max down to zero.

I never trust SPICE with pentodes. I've had very bad luck with screen grid modeling. Pentodes have their places, but SPICE models are dumb.

When working a tube at currents outside "normal", don't trust the model. When I took a 12AU7 down below 1uA, my SPICE model was showing Gm declining with mathematical precision, which I know aint true for a real 12AU7. If you are running that 6L6 at 1mA, the model is giving you numbers that are plausible, but wrong.
 
[quote author="PRR"]> a tube VCA by controlling the cathode current using transistors

So where is your control voltage? [/quote]

Hi PRR,
I found the voltage to current converter circuit from The Mordern Converter and Filter Circuit Encyclopedia by Rudolf Graf, according to the book the current at collector Q1 = +ve terminal input voltage of U1 divided by the shunting resistor(15K in this case). I was hoping to use the rectified feedback DC into a log amp then a diff DC amp for threshold, and feed that into U1 terminal to control the cathode current source.



U1 Q1 force 1mA. U2 Q2 force 15V, but only have 1mA of grunt. What is the voltage and current at the 6L6 cathode?

According to the simulation the voltage at cathode, Q1 collector and Q2 emitter are all 15V and the current is -9.3mA at Q1 collector and -12.9mA at Q2 emitter. The current at the plate is -3.6mA and 158.6V, does this seem to be right?

What can I do estimate and recognise what the realistic value should be?

I am not quite sure how to apply circuit analysis into this situation, still haven't done enough practice to apply it fluently...... :mad: :?
 
I've taken off the signal source and simulated and got the same result Q1 collector and Q2 emitter are all 15V, the current is -9.3mA at Q1 collector and -12.9mA at Q2 emitter. The current at the plate is -3.6mA and 158.6V.

I don't understand how the current at the plate 3.6mA which would be the source, after passing the grid can get to 9.3mA + 12.9mA = 22.2mA shunting to the ground at the cathode!??

or in real electron flow (which I prefer to think) how can 22.2mA at the cathode becomes 3.6mA at the plate, since 250V divided by (22.5K plate resistance plus 25K plate resistor) = 5.26mA. :?

Also, how does Q1 sink 9.3mA when it is set for 1 mA??? :shock: I am not quite sure how Q2 works either...... :mad:
 
[quote author="PRR"]> a tube VCA by controlling the cathode current using transistors

If you are running that 6L6 at 1mA, the model is giving you numbers that are plausible, but wrong.[/quote]

I changed the value of the shunting resistor 0f Q1 and Q2 to 3K Ohm which allows 5mA at the Q1 collector, the value of Q1 collector became -27.47mA and -31.15ma for Q2 emitter the voltage stays the same at 15V. The plate current stays the same at -3.6mA.......

here is the netlist......

*Spice netlist for Circuit: C:\WINDOWS\DESKTOP\TEST.CKT
Q1 5 7 6 Q2N4265
EU1 7 0 8 6 1E5
V4 8 0 DC 15V
C3 5 10 4.7uF
V5 12 0 DC 15V
EU2 9 0 12 5 1E5
Q2 4 9 5 Q2N4258
C2 5 0 470uF
XV1 11 10 3 5 X6L6GC
V2 13 0 DC 250V
R3 0 4 3k
R2 0 6 3k
R5 0 3 100k
R4 10 11 50k
R1 11 13 25k
.SAVE V(3) V(4) V(5) V(6) V(7) V(8) V(9) V(10) V(11) V(12) V(13) @q1[p] @q1[ic]
.SAVE @q1[ib] @q1[ie] @v4[p] v4#branch @c3[p] @c3 @v5[p] v5#branch @q2[p]
.SAVE @q2[ic] @q2[ib] @q2[ie] @c2[p] @c2 @v2[p] v2#branch @r3[p] @r3
.SAVE @r2[p] @r2 @r5[p] @r5 @r4[p] @r4 @r1[p] @r1
*BKGND=RGB 0 0 0
*BINARY RAW FILE

* Selected Circuit Analyses :
.OP

* Models/Subcircuits Used:

*2N4265 Si 350mW 12V 200mA 300MHz GenPurp pkg:TO-92B 1,2,3
.MODEL Q2N4265 NPN (IS=20.3F NF=1 BF=325 VAF=62.4 IKF=0.2 ISE=8.06P NE=2
+ BR=4 NR=1 VAR=24 IKR=0.3 RE=0.258 RB=1.03 RC=0.103 XTB=1.5
+ CJE=6.91P VJE=1.1 MJE=0.5 CJC=6.39P VJC=0.3 MJC=0.3 TF=530P TR=368N)

*2N4258 Si 350mW 12V 200mA 700MHz Switching pkg:TO-92B 1,2,3
.MODEL Q2N4258 PNP (IS=20.3F NF=1 BF=130 VAF=62.4 IKF=0.12 ISE=20.2P NE=2
+ BR=4 NR=1 VAR=18 IKR=0.18 RE=0.258 RB=1.03 RC=0.103 XTB=1.5
+ CJE=4.22P VJE=1.1 MJE=0.5 CJC=7.1P VJC=0.3 MJC=0.3 TF=227P TR=158N)

*6L6GC Vacuum Tube Tetrode (Audio freq.)
.SUBCKT X6L6GC 1 6 3 4
B1 2 4 I=(((URAMP((V(7,4)/8)+V(3,4)))^1.5)/1456)*ATAN(V(1,4)/10)
B2 7 4 I=((URAMP((V(7,4)/8)+V(3,4)))^1.5)/9270
C1 3 4 10P
C2 3 1 0.6P
C3 1 4 6.5P
R1 3 5 1.5K
R2 2 4 100K
D1 1 2 DX
D2 4 2 DX2
D3 5 4 DX
D4 6 7 DX
D5 4 7 DX2
.MODEL DX D(IS=1.0P RS=1.0)
.MODEL DX2 D(IS=1.0N RS=1.0)
.ENDS X6L6GC
.END

:oops: :oops: :oops:

Any idea what went wrong up there!??? :sad:
 
> how can 22.2mA at the cathode becomes 3.6mA at the plate

I'm confused by the sign convention. But it makes perfect sense if Q1 is NOT acting as a conventional transistor, but is broken-down. Is if base voltage 14.4V, or +22V? Or maybe Q2 is broken-down.

Therefore one transistor is forcing current into the other, not to the tube.

I do not see the point of forcing the node with both a current and a voltage. If you force a current, the tube will find its own voltage. If you force a voltage, the tube will find its own current. If you force BOTH, then anything the tube doesn't eat has to come out somewhere else.

Skip U1 Q1. Skip Q2 too, because U2 is an ideal opamp that can source-sink infinite current (and apparently infinite voltage). Or just use a Vdc voltage source for simulation, because you can run a PARAMetric analysis with a Vdc at several voltages (try 0, 3, 6, 9, 12, 15V).
 
[quote author="PRR"]
I'm confused by the sign convention. [/quote]

Sorry about the confusion, I'll stick to conventional current flow.....

But it makes perfect sense if Q1 is NOT acting as a conventional transistor, but is broken-down. Is if base voltage 14.4V, or +22V? Or maybe Q2 is broken-down.

The base voltage of Q1 is 15.8V but the current is whopping 32.47mA!!! :shock: :? The Q2 base is 14.2V with 26.14mA! :?

What would be the cause of this? Does this reflect what will happen in real life?


Therefore one transistor is forcing current into the other,

Would it be correct for me to assume that the current is going in a circular motion drawing current from the ground then from Q2 into Q1, would this work as a constant current source at the cathode?

not to the tube.

Wouldn't the tube draw current because of the higher voltage at the plate?


I do not see the point of forcing the node with both a current and a voltage. If you force a current, the tube will find its own voltage. If you force a voltage, the tube will find its own current. If you force BOTH, then anything the tube doesn't eat has to come out somewhere else.

hmmm.............if Q2 is generating voltage at the cathode, wouldn't that effect the current at the Q1 collector? Since Q1 need 15V at the collector to shunt 1mA from the cathode, so is it possible to use Q2 as a voltage to control the Q1 current feeding into the cathode to accomplish gain control??
 
> Wouldn't the tube draw current because of the higher voltage at the plate?

Sure it draws current. 3mA if we believe SPICE. That's wrong but in the right ballpark: vacuum conducts very much less than silicon.

Curse me for a novice. Q1 collector is reverse-biased. The tube current will fall near zero if U2 Q2 force +15V on the cathode. But U1 Q1 live to force 1mA into the tube; or rather, through R2. Since the tube won't take 1mA, U1's output rises. That won't make the tube conduct any better(!) but at +15.6V the collector juction of Q1 reverse-biases. Current flows backward through the collector junction and into Q2, until Q1 emitter is at +15V. (You have to see Q1 as two forward-biased diodes, not a transistor.) U2 wants to see +15V at its - input and will ask Q2 to take current. But Q2 can only take about 15mA due to the 15K resistor in its collector, so the rest has to run straing through Q2's emitter-base diode and into U2 output.

I don't know why it says 32mA, 26mA, etc. There is some kind of conflict that I am missing.

But what is this scheme supposed to do? I don't have a clue and I suspect you don't either. DO you want to control the tube voltage or the tube current?

And do us a favor. While a 6L6 is a good choice in practice, I don't like SPICE pentode models. Use 2A3, 300B, or 12AU7 triode and get the basic operation understood. For 12AU7 at +250V supply, try cathode currents of 10mA to 0.1mA, or grid-cathode voltages of 0V to -15V.
 
[quote author="PRR"]

But what is this scheme supposed to do? [/quote]

This is actually what I wanted to do, I built this circuit during the holidays and couldn't get it to work due to some stupid grounding mistake.......... Now it seems to work OK, here is the schem.......

tubeamp.jpg


I still need to finalise the biasing value for the 6BA6 screen resistance and the 6SK7 biasing resistors, since I have yet to measure the pot resistance....
 
here is a shot of the cro at the output, channel 1 measures 800mV P-P and channel 2 measures cloese to 20V P-P.

The 3 tube preamp puts out very modest,

20log 20/0.8 = 27.95 dB

can get higher gain with higher feedback resistance at the last tube but increase distortion and would oscillate.....

upload_344338.jpg


here is a shot of the noise figure, less than 20mV.......

upload_344345.jpg
 
> I built this circuit during the holidays and couldn't get it to work due to some stupid grounding mistake.......... Now it seems to work OK, here is the schem.......

I've been quiet, hoping that someone else would po> I built this circuit during the holidays and couldn't get it to work due to some stupid grounding mistake.......... Now it seems to work OK, here is the schem.......

I've been quiet, hoping that someone else would point out the obvious:

You are not clear in what you are doing!

What do U1 Q1, U2 Q2, U3 Q3 do for you? Nothing! Take them out, use just a 3K resistor and the big cap, it will work fine. In fact I don't see how it can work as shown, unless reality is not the same as the drawing (would not be a suprise).

Also using G2 as the output of V2 is unusual. It has been done when you need a low-power triode and all you have is a glass pentode. G2 is a fine "Plate", if a little easy to cook. Then the original plate may be grounded, which acts to shield the rest of the tube.

And the unbypassed G2 on V3 is odd. Normally you must bypass G2 it get good gain.

The 10K load on V3's 1Meg variable plate resistor does not look right. Plate resistors are usually smaller than the load. And 10K is a fairly low load for these tubes.

> The 3 tube preamp puts out very modest, 20log 20/0.8 = 27.95 dB

That's stunningly modest. 6SN7 gain-stage, 6BA6 gain-stage, and 6SK7 cathode follower would give voltage gain over 1,000 or 60dB and THD under 1% at 1V out without any feedback or big stability issues. You seem to have over-complicated, got "modest" performance, and at this point nobody understands what that circuit really does.

Start with the 90 year old tube amplifier designs. Don't go sticking Silicon up their butts without knowing what it really does. int out the obvious:

You are not clear in what you are doing!

What do U1 Q1, U2 Q2, U3 Q3 do for you? Nothing! Take them out, use just a 3K resistor and the big cap, it will work fine. In fact I don't see how it can work as shown, unless reality is not the same as the drawing (whould not be a suprise).

Also using G2 as the output of V2 is unusual. It has been done when you need a low-power triode and all you have is a glass pentode. G2 is a fine "Plate", if a little easy to cook. Then the original plate may be grounded, which acts to shield the rest of the tube.

And the unbypassed G2 on V3 is odd. Normally you must bypass G2 it get good gain.

The 10K load on V3's 1Meg variable plate resistor does not look right. Plate resistors are usually smaller than the load. And 10K is a fairly low load for these tubes.

> The 3 tube preamp puts out very modest, 20log 20/0.8 = 27.95 dB

That's stunningly modest. 6SN7 gain-stage, 6BA6 gain-stage, and 6SK7 cathode follower would give voltage gain over 1,000 or 60dB and THD under 1% at 1V out without any feedback or big stability issues. You seem to have over-complicated, got "modest" performance, and at this point nobody understands what that circuit really does.

Start with the 90 year old tube amplifier designs. Don't go sticking Silicon up their butts without knowing what it really does.
 
Back
Top