Tube Amp CCS?

[JojoD818]

Hi Guys,

I would like to use a CCS for the attached circuit. SY's evolved Morgan Jones CCS with leds is what I had in mind, I had great success with it in preamps.

The first thing I want to do is to remove the 470 ohms and two 220 ohms in the driver tubes and replace that with a CCS. This should work right?

Now for the input triode, which should I replace? the 47K on the anode and use a current source or the 470 ohms on the cathode and use a current sink?

Any comments or suggestion would be a great help.

Thanks!

An externally hosted image should be here but it was not working when we last tested it.

 

[SY]

Nope. If you change the cathode resistor to a CCS, you'll kill the gain of that stage. This particular circuit is not terribly amenable to CCS use.

Now, if you wanted to go open loop, you could replace the cathode resistor with an LED or LEDs, then put a CCS in the plate circuit...

 

[JojoD818]

Hi SY,

Which are you talking about, the input tube or the driver tube? It's 5AM here and coffee is nowhere in sight...

 

[anatech]

Hi SY,
Your feedback would be "shorted" by the LED. Differential front end then?

-Chris

 

[SY]

Chris, that's why I said "open loop."

I was talking about the input stage. The driver stage could be set up as a CCS-long tail, but since it receives a balanced signal from the split-load phase splitter, that's not really necessary.

 

[JojoD818]

Hi Chris, SY

so installing a CCS in the driver tubes (cathode) won't make much of a difference then?

what if I leave the 470 ohms cathode resistor of the input tube and then use a CCS for it's plate circuit?

 

[SY]

CCS plate loads hate cathode resistors.

Nope, this is a design (Williamson) which hangs together as a piece. Incorporating CCS will take some significant redesign. And if it were me doing the redesigning, <broken record>I'd use something like a 6SL7 diff amp input direct coupled to a 6SN7 differential driver</broken record>.

 

[JojoD818]

just as I thought, so it's a hopeless case then until I break it apart...

 

[anatech]

Hi SY,

Chris, that's why I said "open loop."

Gee, I didn't know where were being completely literal! I thought you were just indicating the higher gain from a CCS load. Can't get anything past you these days!

<broken record>I'd use something like a 6SL7 diff amp input direct coupled to a 6SN7 differential driver</broken record>.

That broken record is playing a nice tune. It would be interesting to see a complete design.

-Chris

 

[SY]

I'm tempted to do one in my HF87 re-rebuild. On va voire.

 

[anatech]

Now you are teasing.

You have my full attention sir. If I had a spare HF-87, I'd try it.

-Chris

 

[Johan Potgieter]

Also to add a little,

I would not change the original in this specific case either. This is not entirely a Williamson, but there was a reason why that amp was one of the cleanest around. A CCS as an anode load (1st stage) will increase the gain by perhaps 20% - the 6SN7 mu is only 20. Linearity will also not improve drastically as the signal amplitude there is quite low.

Regards!

 

[rdf]

CCS plate loads hate cathode resistors.

Bumped my head against a knowledge limit. Why so?

 

[SY]

I'm just guessing here- because of the magnification of the cathode resistor by the mu to form a nice, big nonlinear late resistance. I don't know if this is entirely the reason, but it's something I've observed consistently, poor distortion from degenerated stages with CCS plate loads.

 

[Tubelab_com]

CCS plate loads hate cathode resistors.

I don't get that one either, since all of my recent amps use a triode with a CCS (10M45) load and a (bypassed) cathode resistor for the gain stage. Gain is almost equal to Mu (depending on load) and distortion is very low.

After many suggestions from SY, I tried replacing the bypassed cathode resistor with an LED. It works pretty good, but I still haven't found the right LED. My plate voltage is too low. I only have a few LEDs in my collection. The green ones work best, so far. The red ones are usless, not enough voltage drop. I have yellow and orange, but I haven't tried them yet. They are at least 20 years old.

 

[SY]

Take a look at what that stage is doing open-loop without the bypass. I've tried this with 12AT7, 6SL7, and E88CC with consistently bad results. Unexpectedly, I should add.

 

[Tubelab_com]

OK, I did try it without bypassing the cathode, with poor results. The voltage gain was low, something like 10 with a 5842. Gain went to 40 with a bypassed resistor. This is with a CCS plate load and a mosfet follower (PowerDrive). The gain stage distortion is uder 1/2 % with over 50 volts out.

My latest amp uses 1/2 of a 12AT7 as the only gain stage. I get a voltage gain of 50 with a CCS load on the plate cap coupled to the output stage grid with a 120 K grid resistor to ground. I have not tried this amp without the cathode bypass because I need the gain.

The 12AT7 seems to work well with an unbypassed LED though. I just need the right LED.

 

[SY]

I was surprised to see the distortion of a 12AT7 with an unbypassed normal red LED be as low as it is. With a JJ, you can get 0.03% distortion at 15Vpeak out. The other alternative is to bite the bullet and supply the LED with 8 or 9 extra milliamps with a CCS.

 

[rdf]

I don't disagree, I was fishing for an explanation of the pure R case. Interestingly enough, the low cathode resistance (or maybe the non-linear impedance?) of series LEDs can be beneficial. The third harmonic of a CCS-loaded 6C45 biased by two series green LEDs (5ma, ~4 Vdc) rises almost 10 dB when the cathode is bypassed by 1000uF. 2nd doesn't budge.

 

[SY]

With the simple 12AT7 circuit, bypassing the LED dropped the degeneration, but I couldn't measure any change at all in HD at the plate. Not totally unexpected. I saw the same thing with some 6SN7 tests with two seriesed red LEDs- no change in distortion spectrum with bypass. Holding current constant does help...