Crossover distortion

[lowell]

Hey all,
In my custom build I have 2 6v6 and 2 6l6. I am getting some nasty crossover distortion on the clean channel. (haven't scoped OD channel yet) On the output the wave (1khz sine) is showing lots of crossover dist and is also vibrating/oscillating when viewed on the scope.
Would someone please see the schem below and see if I have something off in the power amp. I can't find anything wrong w/ it. I did however rebias the 6v6 w/ a 400ohm Rk and each 6v6 is around 36ma. Each 6L6 is around 63ma. Va on power tubes is 374vdc. The 6v6 rebias helped a bit. The power supply (not on schem yet) is full wave rectified, through a 200ohm sag resistor, into 16uf, 5h choke, 16uf, etc.... Maybe the initial 16uf cap needs to be more around 40uf? There is a swirling high pitched distortion thang going on. It's there whether the "turn of the 6l6" switch is on or off. The scope looks good until the plates of the power tubes. I can't see how my tubes are over biased (too little current) but maybe I'm missing something. I don't know, are the grid leaks good/ok values? Oh and the 6V6 screen resistors are actually 100ohms not 470. I'm thinking I should increase those.

[zbyszekskibinski]

Crossover distortion problem

I think what you could do is to plug in the generator to the amp imput (frequency 1000hz sine wave-its also available on youtube) and use your oscilloscope to find the problem by connecting it via low value capacitor in series with the scope. To isolate the fault you could start searching for the symmetrical sine wave from the input stage following to the next stage and going up to the power stage. if at the output of one of the stages the sine looks distorted and at the input it's clean, probably the problem is one of the preamp tubes or the components around it. If you have a generator you could pass the signal from it at different voltage values, if the volatge at the input of the amp is abit too high, the sine wave at the scope display should come up as cut off symmetrically on both sides (that's theory I remember from the books)Also check if distortion increase with turning up the volume or not. If it does, the problem is before volume control somewhere in your preamp sections.(that's what I think but I'm not sure about it) It's always better to hook up a large resistor at speaker output with resistance and power corresponding to speaker power and impedance to avoid annoying sine wave squeals. If you isolate the problem at one of the stages you could test resistors which sometimes drift off their values, shorted decoupling caps blocking amp stages and tubes themselves, wish you good luck with it but also wait for some other posts as I'm not a very experienced technician and some guys may give you better advice with it

[defaced]

Are you getting crossover distortion on just one set of tubes or both? Meaning, have you pulled tubes and scoped.

[lowell]

In the schem you can see I have a switch that cancels signal at the grids of the 6l6s. It actually works very well and w/ the scope on the grids of the 6l6 there is NO signal. So whether the switch is on or not, I'm getting this crossover-looking distortion.

[defaced]

But that says nothing about whether the 6L6s are experiencing cross over or not, just that the 6v6s definitely are. Pull the 6v6s and see if it goes away, then you can figure out what battle(s) you've got to fight.

[lowell]

Ok so w/ the cathodes of the 6v6 disconnected the plates of the 6L6 are only showing a very small inkling of crossover distortion. However the plates are clipping fairly hard at +/- 260v p-p. The 6V6 are showing insane amounts of crossover distortion when the volume of clean channel is cranked and no signal going into the 6L6. The 6V6 do not show any crossover dist on the scope until the volume of clean channel reaches 75%.

Ok and fyi at the 6v6 grids, the signal starts to clamp around 120v on the top and bottom. Don't think that'd be related to crossover dist.

[lowell]

Alright so I added 470 screen resistors to the 6v6 cause I don't think the choke creates sufficient v-drop. I rechecked the 6v6 bias and they are right at 90% of max Wa (67ma for both tubes at 374v).

I then pulled the 6L6 and no change in the crossover distortion on the plates. I did notice however that the crossover distortion ALMOST goes away and is quite minimal when the 6L6 and 6V6 are both fully functional.

[lowell]

Other than cold bias, what could cause crossover distortion?

sorry for so many consecutive posts...

I'm baffled. I've rid of the 6L6 tubes and rebiased the 6V6 to 90% and still crossover-looking distortion on plates. Is there something other than cold biasing that would cause crossover distortion?

[lowell]

ok I started the process of elimination. Turns out the 100uf bypass cap on the 6V6 is causing it. I removed it and voila. I did try a 33uf to see if that changed it but it did not. I'm sure there's a great explanation of why this is happening. If someone has it please divulge. Does the Paul Ruby mod have something to do w/ this, or putting a zener on the cathode somehow?

[Chuck H]

It's worth noting that your using a cathodyne type PI. This type of PI will crumble the moment you try to draw any grid current from the power tubes. The more you increase bias on the 6V6's the more the 6V6's will try to draw grid current. The crossover distortion could be due to 'grid loading'. The coupling capacitors from the 6V6's are charging as the 6V6 grids try to draw current but the PI cant supply it. You also have the 6L6's trying to draw a little current. The power tubes always 'try' to draw at least a little current at the grids. So you have a double whammy with two seperate bias points. The Paul Ruby mod will only fix this problem if set the zener voltage right at or barely above the cathode voltage for the 6V6's when using a cathodyne PI for this circuit.

You might actually try biasing the 6V6's cooler and use the Paul Ruby Zener mod with the zener voltage set 1V above the cathode voltage. Provided that doesn't set the idle for the 6V6's too cold. You can also place another zener across the cathode resistor itself to act as voltage regulation at the point of crosover distortion. Effectively "fixing" the bias above that quiescent.

Chuck

[Dave H]

I did notice however that the crossover distortion ALMOST goes away and is quite minimal when the 6L6 and 6V6 are both fully functional.

That will be because when both sets of tubes are driving the OT each pair of tubes will effectively be seeing a higher impedance load than when a single pair of tubes is driving the OT.

To have the optimum impedance for both the 2 and 4 tube settings you really need to also switch the OT impedance.

Dave H.

[lowell]

Ok I will try that out. Can you please explain the mod to me? I cannot find a real schem or detailed explanation on google or anywhere on ampage.

[Dave H]

Ok I will try that out. Can you please explain the mod to me? I cannot find a real schem or detailed explanation on google or anywhere on ampage.

If you are asking me for an explanation I was thinking that it would work better if you were to change the OT secondary tap to make the primary look like 4k for 6V6 + 6L6 operation and 8k for 6V6 only operation. eg. If you have an OT with a 4k primary and 4ohm and 8ohm secondaries connect an 8ohm speaker to the 8ohm tap for 6V6 + 6L6 and connect the 8ohm speaker to the 4ohm tap for 6V6 only. That would be great if you were using 4 x 6V6 but it’s a compromise when using 6V6 + 6L6.

I’ve just noticed that the 6V6s are cathode biased at 36mA and the 6L6s fixed biased at 63mA. The difference in bias current is another reason why the 6V6s have more crossover distortion and they are cathode biased with 400R which will cause bias shift which increases crossover distortion. To reduce bias shift I wouldn’t want to use much more than 300R for the bias resistor.

Dave H.

[lowell]

Sorry should've been more specific. I understand the OT primary impedance issue. I could not however seem to find the Paul Ruby mod but I just found it. It's here if anyone needs it: 18-Watter Buzz Info

You say to use the 300ohm Rk. At 400ohm Rk and 374 Va I'm already at 90% Wa max. I guess I could try 100% Wa max but am a bit worried the tubes will be short lived.

[Dave H]

You say to use the 300ohm Rk. At 400ohm Rk and 374 Va I'm already at 90% Wa max. I guess I could try 100% Wa max but am a bit worried the tubes will be short lived.

I see, my 6V6s only run at 90% with a 300ohm Rk. If your OT is for 6L6s its primary impedance will be too low for cathode biased 6V6s. There would be less crossover distortion if they were fixed bias. I'd replace the 400ohm Rk and cap with a zener (24V ?) to make it fixed bias at the same bias current you have now.

Dave H.