18 watt plexi with 6v6's?

[chuckb]

Hey guys, I just finished an 18watt plexi with 6v8's using Mark Huss' schematic and I'm having some issues trying to adjust the bias. I believe that the bias range for 6v6's is 23mv to 27mv? I have been getting a reading of 30mv with the bias pot all the way down, so I increased the 15k bias resistor to 22k and the bias went higher? I changed out the 22k for a 10k and it lowered but still not enough. so I went all the way down to a 6.5k and the bias was steady at 27.5mv? i thought that by increasing the bias resistor that it would help reduce it but it had the oppisite effect. I''ve attached the schematic of the build and hope someone can shed some light on why this is happening?

Chuck.

[jbltwin1]

Increasing the restrictor resistor(that's how I keep them straight) REDUCES the amount of control grid voltage and control grid voltage is what RESTRICTS the current. More negative means less current. Increase the resistance to increase the current. Counter to what you would normal think.

[chuckb]

Ya, it was oppisite of what I was thinking. so I have it running at 27.5mv now, is this a little high for a jj 6v6s?

[g1]

I believe that the bias range for 6v6's is 23mv to 27mv?

A number like this is kind of meaningless without knowing the plate voltage the tubes are running at. In the schematic there are 1 ohm resistors between each power tube cathode and ground. By measuring the voltage across those resistors the current is easy to calculate. I=V/R, so a 27mV reading across a 1 ohm resistor means there is 27mA of current flowing. Now multiply the current by the plate voltage to calculate the idle dissipation in watts. The schematic shows approx. 440V at the plates. So .027 times 440 gives approx. 12 watts.
The schematic shows 14mV across the 1 ohm resistors, so he has designed it to run the 6V6's at approx. 6 watts. That is half of what you have them biased at.
As a rough guideline, 6V6's are considered a 14 watt tube and in push-pull would commonly be biased between 50% and 70% of that, so 7 to 10 watts.
Where did that 27mA figure come from, and what is the plate voltage? I'm wondering if the 27mA would be for 2 tubes in a cathode biased arrangement? Or running at much lower plate voltage?

[chuckb]

Hey g-one, i measured the voltage across the 1ohm resistors and got a reading of 28mv,the vlotage on the plates is 380v so if I have my calculations right 10.5 watts.So I have the tubes running a little higher than 70%?

[g1]

Correct. You are about 76%. In the schematic you provided, Huss has them running at a cool 45%. Somewhere between those 2 figures will be your own happy medium between tone and longevity .

Edit: With regard to your original problem with the bias circuit. I believe the problem may be due to using different power transformer. Suggest that rather than changing the 15K, decrease the 220K before the diode. Try to get the -54V shown at the diode/15K junction.

[chuckb]

Thanks for your help much appreciated.

chuck.

[chuckb]

Hey g-one, I have my amp running at aprox 8watts. thanks for the help on that one. I'm now trying to bias another build(BAD Cat Hot Cat with cathode bias, class A). The schematic gives me a plate voltage of 376v and a cathode voltage of 24v. problem is the transformers I'm using are giving me 470v on the plates and with a 270ohm/10w resistor on each tube(El34's) I'm getting 20v on the cathodes. l=v/r - 0.027 x 470= 9watts. I believe that with the el34 at idle dissipation is max 25v, I'm to low @ 9watts.I increased the cathode resistor to 450ohm and now have a voltage of 24v which gives me a idle curret of 11 watts,so it still has to come up. Rather than increasing the cathode resistor, is there a way to possibly lower the plate voltage to a more managable level?

Chuck.

[g1]

I'm getting 20v on the cathodes. l=v/r - 0.027 x 470= 9watts.

Nope, 20/270 gives .074. Then we use plate to cathode voltage as it is cathode biased. So .074 x 450V gives approx. 33W, way hot.

[chuckb]

Sorry about that,I got a reading 20v(not 27v on the cathode and a reading of 470v on the plates.
I'm not understanding where you got 20/270 and is there a different way to calculate the bias on a cathode biased amp?

Chuck.

[g1]

It is calculated in the same way as the other amp. The first amp had a 1 ohm cathode resistor, this one has a 270 ohm. For the first amp, I=.027/1 = .027A or 27mA. For this amp, I=20/270=.074 or 74mA. In both cases, to calculate the wattage we multiply by the voltage across the tube itself (plate voltage minus cathode voltage). We just used the plate voltage on the first amp because there is not enough cathode voltage to be significant. In the second case the 20V on the cathode is big enough that we take it into account, so the voltage across the tube is actually 450V. 74mA x 450V gives about 33 watts idle dissipation.

[chuckb]

Well i understand it now. I have since added a 180ohm resistor in series to the 270ohm to give me a total of 450ohm. so I have a cathode voltage reading of 24v/450ohm=.053ma x 470 plate voltage =25ma which is the max idle dissipation for el34's correct?

[g1]

Close enough . You meant 25W, not 25mA, and .053A not .053mA.
And you need to account for the voltage at the cathode. So the actual voltage across the tube is 470V minus 24V = 446V. Idle current x voltage across tube = idle dissipation in watts. So .053 x 446 = 23.6Watts. And 25W is max. for EL34.

[chuckb]

Close enough . You meant 25W, not 25mA, and .053A not .053mA.
And you need to account for the voltage at the cathode. So the actual voltage across the tube is 470V minus 24V = 446V. Idle current x voltage across tube = idle dissipation in watts. So .053 x 446 = 23.6Watts. And 25W is max. for EL34.

yes, I mean 25W & .053A . I wasn't accounting for the voltage @ the cathode. So I have to subtract the cathode voltage to get the actual voltage across the tube. 23.6W still running the amp a little to hot, i'll have to try it out and see if I like it although i may try to get it lower. 70% of max would give me 17.5W which be ok. thanks for your help, I've learned alot here, although i may be back with more ?'s before the days over

chuck.

[g1]

Well, if you want to cool it down that's up to you. But class A is generally considered biased at 100%.
I'll leave it to you to read up further on class A if you want, this is a good article: Class A