KT88 power amp

[sith]

Hi, I came over this one, and I wander is it possible to change 6AU6 with noval base lamp (EF86, ECC8x) or similar. Is it possible to connect triode mode to ultralinear OPT.

[SY]

The 6AU6 is triode connected, but in a funny way. The biasing of that tube is also... odd. Unless I'm reading the value of the feedback resistor wrong, which may be possible because the drawing is pretty fuzzy. You could sub a real triode, if you wish. A 12AX7 works well in that spot, but the plate resistor would need to be bumped to 100K and you'll have to redo the cathode circuit to bias it properly and introduce the feedback. If you search for a schematic for the old Eico HF87, how to do this will become clear. If you want to use an octal, I'd go with a 6SL7 or a 5691.

You can indeed go UL on the KT88s, but don't skimp on the feedback.

[rcavictim]

Originally posted by SY
The 6AU6 is triode connected, but in a funny way.

Sy,

Since the majority of pentodes, moreso the power ones, have G3 internally tied to the cathode there just isn't any option for the circuit designer when desiring triode mode to accept that G3 will operate at anything other than cathode potential giving one a triode with a suppressor grid. That ain't natural. When someone, in exercising a somewhat rare opportunity of freedom ties G3 to G2 you find this funny?

Just for fun, what would have been the effect on you if the designer had tied all three 6AU6 grids together? Would you have been in a state of LOLARAOTF ?

I mean no disrespect, I'm just good naturedly pulling you leg. I just thought your describing the hookup 'funny' was itself funny. OK, I don't get out enough.

I find it funny that the designer didn't just use a triode. 6AU6's generally make better microphones anyhow than gain stages IME.

If this is an existing chassis and the input tube socket already exists with 7 pins may I recommend a 6C4 (and rewiring as appropriate) if the loss of gain can be accomodated?

[fdegrove]

Hi,

The biasing of that tube is also... odd. Unless I'm reading the value of the feedback resistor wrong, which may be possible because the drawing is pretty fuzzy.

Am I correct in reading_whilst torturing my already sore eyes_the FB resistor as 330R, the frequency correction cap as 3.9nF and the cathode bias resistor as 6.8R?

What's the purpose of the cap the bias resistor is standing on? Never seen it done like this before....

The rest of the diagram looks straightforward enough..it's just input I'm having trouble with.

Cheers,

[rcavictim]

Originally posted by fdegrove
Hi,



Am I correct in reading_whilst torturing my already sore eyes_the FB resistor as 330R, the frequency correction cap as 3.9nF and the cathode bias resistor as 6.8R?

What's the purpose of the cap the bias resistor is standing on? Never seen it done like this before....

The rest of the diagram looks straightforward enough..it's just input I'm having trouble with.

Cheers,

Frank,

You are right, that is really odd. With my blurry schematic decoder glasses turned up full I see the cap bypass R is 4700 uF, the 6.8 ohm R in series with it allows some portion of the NFB to act on the tube cathode (developed across this R). Note that the DC cathode current flows in its entirety through the speaker winding of the output xfmer. Very odd circuit! This means that loudspeaker loading and back EMF will effect bias on the input stage. That huge 4700uF cap is there to try to smooth over this back EMF problem I guess. Maybe this amp is prone to motorboating and this was the cure? Play with this cap value and one could have a dynamic processor amplifier all in one!

Now that's also funny!

[SY]

Thje 4700u cap lets the 6R8 set the feedback and overall gain at AC. It has to be that big because of the small size of the feedback resistor. Part of the "funny" may be my eyes- I thought the screen resistor was 330K not 330R. If the latter, this is a pretty normal triode connection. One can still do better with a higher mu "real" triode like the ones I mentioned. Better still is the 12BZ7, but it's a rarish bird.

One disadvantage to having the cathode current flow through the secondary is the effect on magnetization; that could be cured with a CT secondary. Another is that you'll inevitably get some offset. The back-EMF stuff is normal for any feedback amp.

And finally, I have a prejudice against the 12AU7 used in the diff amp- I've never been able to use this tube without experiencing higher distortion. My own choice for that hole would be a 6SN7. But "check on a song oooh," as the French say.

[fdegrove]

Hi,

Better still is the 12BZ7, but it's a rarish bird.

But it's an excellent bird...as is the AZ7, BTW.

Thanks for the explanation...I figured something like that but didn't think it funny, rather errrrrrr...not so smart given the alternatives.

I have a prejudice against the 12AU7 used in the diff amp- I've never been able to use this tube without experiencing higher distortion.

You're not alone on that one...I avoid 12AU7s/ECC82s in a NFB loop like the plague: bright and anemic come immediately to mind, o-scope shows it too.
Either the 6SN7 or if real estate is tight a 6CG7/6FQ7 would cure it for me.

Cheers,

[Idefixes]

This his a WAD schema. On this net URL you can see a recent build version based on with an optimised PSU. Sorry but it's in french.

Marc

[FrancoB]

I happen to be building this amp and couldn't be more happier with two gurus discussing mods on this circuit.

So, fdegrove and SY. Please go on!

Thanks,

Franco

[EC8010]

You chaps are saying that the feedback is funny, but if it were a transistor amplifier, you'd say it was completely conventional! It's a better method because it degenerates the gain to unity at low frequencies instead of producing a sub-sonic bump.

There really isn't any problem with passing that little bit of DC through the output transformer. Suppose that first stage passed 3mA of cathode current, with a typical output transformer turns ratio of 30:1, that would be equivalent to an imbalance of 100uA in the output valves. Would you be worried about that?

That feedback scheme is far better than the traditional method...