or Connect
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Low Voltage 6N6p-OTL by H. Ahammer
New Posts  All Forums:Forum Nav:

Low Voltage 6N6p-OTL by H. Ahammer - Page 5

post #61 of 89
Thread Starter 

Voltages added in RED. Is there anything I missed?

 

 

post #62 of 89
The voltages all seem to work out more or less -- there is about 1.8V bias on the first tube and ~1.2 on the second. More might be better, but that is high enough that you should not be easily over driving things.

Do you have the same issue in both channels? The E88CC is a pretty high Gm tube that is happy to oscillate if it is not well cared for. So that's one place to look. Make sure all the resistors are right on the pins, and that there are no long leads anywhere.

The next step, assuming that everything is hooked up as it should be, would be to test each section independently. If you have a cmoy (or any other amp), run the input stage into it and see what you get. Similarly, try listening to just the output stage on its own, and see if you can pin down where the issue arises.

Going forward, if you plan on keeping the amp, there are a few changes I'd make. In particular, unless you are planning to add a feedback loop, I'd ditch the 6922 for something with a little less gain. -- 5687, 12AU7, 6SN7, another 6n6, etc. Second, I'd build a negative rail by adding 2 diodes to the rectifier and use that with a CCS in the tail of the driver tube. Then I'd scrap the RCRC section of the PS and add a small regulator, and probably increase the voltage and the plate resistors on the first stage. Just my $0.02.
post #63 of 89
Quote:
Originally Posted by dsavitsk View Post

Do you have the same issue in both channels? The E88CC is a pretty high Gm tube that is happy to oscillate if it is not well cared for. So that's one place to look. Make sure all the resistors are right on the pins, and that there are no long leads anywhere.

 

+1. And the same for the 6n6p. I only see one grid stopper ?

 

Quote:

Going forward, if you plan on keeping the amp, there are a few changes I'd make. In particular, unless you are planning to add a feedback loop, I'd ditch the 6922 for something with a little less gain. -- 5687, 12AU7, 6SN7, another 6n6, etc. Second, I'd build a negative rail by adding 2 diodes to the rectifier and use that with a CCS in the tail of the driver tube. Then I'd scrap the RCRC section of the PS and add a small regulator, and probably increase the voltage and the plate resistors on the first stage. Just my $0.02.

 

 

Agreed on all points.

 

As is, the input stage is indeed extremely poor. With only 180r in the tail, it doesn't behave as a proper ltp and is extremely unbalanced. A quick sim gives a gain of 4 right now :blink:. A quick look at the curves doesn't place the current bias point in a very linear place either. So, if you don't have any other tubes, I'd at least put a ccs in the tail as suggested (about 4ma per tube, so 8ma), raise the input voltage at least up to 130V and change the load resistors to 10k. Not exactly how I'd design an amp but the reduction in distortion is dramatic in sims. Gain would be about 10x.

post #64 of 89
Thread Starter 

Thanks for the input guys.

 

I don't have other input stage tubes, except for some more 6N6P, but I do have a pair of 5963 (12AU7) in the mail right now.

 

Sound-wise, it's in both channels.

 

In terms of long resistor leads on the E88CC - the only long ones are the 180R tail resistors to ground and the 7K5 resistors are connected by a 2" wire to the caps, everything else is under an inch.

 

I'll try to test the stages independently if I get some time this weekend, kind of busy lately. :(

 

In terms of "install this part here", how would I create a negative rail and install a CSS on the tail?

 

As for the input voltage - I'm assuming you're saying replace the 7K5 resistors with 10K. Correct?

 

Also, in terms of whether I'm keeping the amp - I'm completely impartial to keeping or rebuilding it into something else, but as I already have the casework for 4x 9-pin tubes and the 4-pin rectifier, it would be nice to keep that configuration. :) Aside from that, anything goes. Of course keeping the transformer and PSU is a huge bonus, but I'm happy to modify the PSU, as long as I get to reuse some of the parts. :)


Edited by FallenAngel - 1/25/14 at 1:43pm
post #65 of 89

Since you asked for open-ended opinions...  if you can get about 180V from the supply (by increasing the cap in between the rectifier and the choke to 33uf... what's the dcr of your transformer secondaries ?), you'd be better off with a white-cathode follower output. I wouldn't reinvent the wheel and just use the aikido, as suggested by Broskie. Check for the schematic here: http://www.diyaudio.com/forums/tubes-valves/133201-6n6p-aikido.html#post1659341 , you want the one with 12au7-ecc99 (right column). Just replace r11 by a pair of diodes (Bijou-like, or SY-like, or Broskie-like, whoever you want to credit for that).

 

You'll be able to keep pretty much all your casework intact.

post #66 of 89
Have you tried shields on the tubes yet? I have seen the 6n6 make some pretty horrible noise without them.

Example of a negative rail here. http://tubelab.com/designs/tubelab-se/schematic/
post #67 of 89

To add something constructive. If you replace the ecc88 by a 12au7, you don't really need a negative supply for a CCS. With 4ma per section and the anode at about 100v, you have 3V of bias, just enough to fit a jfet ccs.

 

So you could try to use the attached schematic.

 


Edited by 00940 - 1/26/14 at 6:30am
post #68 of 89
Quote:
Originally Posted by 00940 View Post

To add something constructive. If you replace the ecc88 by a 12au7, you don't really need a negative supply for a CCS. With 4ma per section and the anode at about 100v, you have 3V of bias, just enough to fit a jfet ccs.

LM334 will work in place.

Actually, why are the two stages cap coupled? The easiest thing to do might be to put a CCS under the output tube (10M45 or DN2540 will work fine) and eliminate the cap and allow the thing to find it's own bias. You might need to tinker around with operating points a bit, but I'd think it would be worth the effort.

None of that helps with the noise problem, which I still say might be solved with some tube shields.

SK-9PINF-2.jpg?1381959151
post #69 of 89
Quote:
Originally Posted by dsavitsk View Post

Actually, why are the two stages cap coupled? The easiest thing to do might be to put a CCS under the output tube (10M45 or DN2540 will work fine) and eliminate the cap and allow the thing to find it's own bias. You might need to tinker around with operating points a bit, but I'd think it would be worth the effort.

 

So that the both stage can find their own bias ? :wink_face:    I don't see much hurt in a small coupling cap with a significant dc potential across it, especially when there's a big coupling cap at the end of the day. I'd think you'd gain more in maximizing the voltages at the anode of both tubes than trying to directly couple them and you need the cap for that. But that could very well be a matter of tastes.

 

I agree on the solid state ccs though. A dn2540 set at 40 to 50ma could replace the 1k resistor and the cathode resistor adjusted for about 40v at the grid of the 6n6p (or top of the ccs, same thing).

 

 

 

Wrt the noise... Would it be possible to have pictures of the current state of the amp ? Another thing: what did you do with pin 9 of the 6n6p ?

post #70 of 89
Thread Starter 

Current stae of the amp:

 

Pin 9 of the 6N6P is grounded.


Edited by FallenAngel - 1/27/14 at 4:06pm
post #71 of 89
Thread Starter 

In terms of transforming this into an Aikido-like amp running off about 160V, what resistor values should I be using? I know this is asking for quite a bit, but the link (http://www.diyaudio.com/forums/tubes-valves/133201-6n6p-aikido.html#post1659341) shows it running off 250V, not 160V.

post #72 of 89

Here is schematic that should do the trick.

post #73 of 89

Here's how I'd do the aikido on 160V. For a 12au7-6n6p combo (the 6922 has too much gain), keep all the values given by Broskie but make:

 

r2-r4: 330R

r11: replace it by two 1n4007 diodes in serie. No need for capacitor bypass anymore.

 

The problem is R11. With only 80V across each tube, you're very close to bring the grid to 0V to get any decent current flowing. From the published curves, a 1.4V drop across r11 should give you about 15ma flowing.

 

I must confess it's a mystery to me why Artsi requires 5W resistors. 0.6W resistors are sufficient.

 

At the risk of repeating myself, if going the aikido way, my point of view is that your PS is loosing way too much voltage. 3uF as first cap is just a waste. With your transformer, 5u4gb-33R (at each diode plate)-22uF-10H/160m-220uF followed by 220R-100uF by channel should give you a decent 210V to 220V supply without overtaxing your rectifier tube. Especially since the aikido stage cuts your standing current by two (about 20mA per channel). If going that road, make r2-r4 470R and use 3 diodes in serie.

post #74 of 89
Quote:
Originally Posted by 00940 View Post

 

I must confess it's a mystery to me why Artsi requires 5W resistors. 0.6W resistors are sufficient.

 

I have just calculated simply 160v and around 0.017A through 6N6P. 160x0.017=2.72w or have i calculated totally wrong? It could work with burning hot 0.6w resistor too.

 

With my DIY amp's 6AS7 150v and 0.07A =10.5W, and 10W resistor goes very hot, or have i understood these wattages totally wrong and 0.6W would work better with my DIY amp?

post #75 of 89

Under normal working conditions*, there's never 17ma/160V across those two resistors. The cathode one for example will dissipate 43mW (R*I²,0.017² * 150). That's the good thing with WCF outputs, all the heat is dissipated by the tubes. That's why Broskie specifies only 0.5W or higher.

 

And neither will you have 17ma with a 150r cathode resistor with a 6n6p at 80V (that'd be -2.5V at the grid, you need 110-120V across the tube for that)

 

* The only conditions where I could see those resistors dissipating a lot of heat would be a failure of a triode as short. At this point, whatever their ratings, they'll go poof with the tubes.


Edited by 00940 - 2/2/14 at 8:36am
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Low Voltage 6N6p-OTL by H. Ahammer