Problems biasing Gold Lion reissue KT66 in Heathkit W5

[GordonW]

Well, I'm wondering if I'm just dealing with a set of inordinately "hot" KT66s here.

Been trying to re-bias a pair of Heathkit W5s, for use with reissue "Gold Lion" KT66s. The amp was just fine with a set of GE 7581s... but, with the KT66s (BTW- the amp was originally specified for KT66s), I've done the following, and STILL have slight red-plating (at the seams, very faint but there):

- Replaced the 330 ohm 7w cathode resistor with a 390 ohm 10w.

- Made a voltage divider for the voltage feeding the grid pins of the output tubes- I replaced the 100K resistor feeding the grids, with voltage divider consisting of a 180K from the source (bias balance pot) and a 270K to ground. The voltage there was around +8 volts, now it's around +4.5 volts... so I've basically biased the grids 3.5 volts MORE negative than stock, while still keeping right at 100K total parallel resistance (taking the parallel combination of the 180K and the 270K) to the grids... which is what KT66 and/or 7581 wants...

In essence... I've increased effective grid/cathode bias voltage from around -45V (with the 7581s) to around -53... and the output tubes are still not completely happy.

I did a calculation of dissipation... it's right at 27 watts cathode draw... around 109ma total current draw at idle for the pair. Which is also just within spec for the KT66 (25 watts plate plus 3 watts screen, or 27 watts total in UL mode). It just seems that this is a LOT of change necessary, to get it below that (I'm probably going to have to go to a 470 ohm resistor, which will probably make the bias more like -56V).

The amp plays fine... in fact, it's got quite a bit more "oomph" with the KT66 tubes in it. It's just that it seems like this is an AWFUL LOT of mod required to get it to bias up right.

The only other factor of possible interest, is that B+ is around 510V now on the plates... factory spec is 480V. Ostensibly, that's the difference between 110 and 120V, roughly, on the line input.

Is there any data out there, indicating that the reissue Gold Lion KT66s just "run away" when run on plate voltages higher than 500V or so? That would be odd, given that the specs for KT66s indicate tolerance for up to 550V... it would be a case of the tubes significantly deviating from specs...

I hesitate to go much further... because this type of change is making it where OTHER tubes won't work right (6L6GCs and 7581s will be far too "cold")... is it normal for KT66s to be THIS hot??

BTW: For info sake... both amps behave EXACTLY the same. So, it's very unlikely that it's problems elsewhere...

Regards,
Gordon.

 

[GordonW]

BTW again: These W5s have the Peerless 16309 output transformers, instead of the "heavier" 16458s. So, I'm especially paranoid about idle current draw, as you could expect...

Regards,
Gordon.

 

[FalconEddy]

Outside of the mods you made, how does the amp compare to the original schematic? Any changes or omissions? How about other feed voltages?

That's a fair amount of mods to make and still have those KT-66's about at the point of runaway. Especially when your pair is running those less robust 16309's.

You could pull down that 510V feeding the plates, but then you're getting into a territory where you've limited tube choices down to a KT66.

I was searching but haven't seen data to support an 'over 500V' issue for the KT88 GL reissue.

I'm speculating that you don't have a vintage quad of KT66's just laying around doing nothing, right?

. . Falcon

 

[GordonW]

The only other non-stock item on the amps (they've been recapped, BTW- to original spec'd values) is a 12BH7 tube in the diff amp of each (5814 in the first tube position). And they DO have the correct 5R4 rectifier tubes (so there's no other viable choice for rectifier, to reduce B+... the 5R4 is about the highest voltage drop there is in a tube rectifier).

Other feed voltages in the amp are proportional (the same percentage higher than stock) compared to the main B+ to the plates. This indicates that it's almost undoubtedly just the higher line voltage (120v vs 110v) causing the slight voltage increase...

Regards,
Gordon.

 

[Kegger]

Gordon, regardless of how you end up biasing them I would get that plate voltage
lower then 500, even 6L6GC is tough at 500 an over, not all like that voltage either.

I'd be trying for 475v if I were lowering it, and probably drop the primary side of the
power tranny, as the heaters are probably a little high anyway.

edit: (cathode bias, voltage drop)

 

[JonL]

Maybe use a bucking transformer to bring all the voltages back down to the design values?

 

[Kegger]

Find a good way to drop the voltage down and your biasing problems will probably be fine as well.

 

[Jim McShane]

Gordon, regardless of how you end up biasing them I would get that plate voltage lower then 500, even 6L6GC is tough at 500 an over, not all like that voltage either.

I'd be trying for 475v if I were lowering it, and probably drop the primary side of the power tranny, as the heaters are probably a little high anyway.

With an amp that has the tube cathodes above ground the following has to be considered:

The plate voltage to ground may be 500+, but with 50 volts or so on the cathode the tube is only seeing 510 minus 50, or 460 volts. 460 volts is the real plate voltage (remember, voltage = potential difference!), and it's not a problem.

Gordon, you likely have a fairly hot set of KT-66s. I run them in at 465 volts on the plate/screen and -50 on the grid. I don't want tubes over a certain Ik figure, and there are quite a few of them in a given batch.

They have no runaway problems at all in my experience. You are going to need about -55 to -58 on the grid from the looks of it. 50 ma Ik would put you at 23 watts combined plate & screen dissipation. That's fine for the tubes, if the OPT worries you go for 45 ma.

 

[Kegger]

Yep didn't pay close enough attention is was a cathode bias amp, carry on...
(yes in a cathode bias amp you subtract the cathode voltage from the plate)

 

[GordonW]

So, I guess I'll just try dropping in a pair of 470 ohm cathode resistors, and go from there. Should theoretically make the cathode bias about 4v more negative. At least that should get me in the ballpark, to tune it with the grid resistor voltage divider.

And yes... effective K to P voltage (cathode/plate) is 510 - 57 = 453v. So, realistically... it's not over-voltage that's the problem. The PS caps in the unit are all rated at plenty voltage (two 500v in series, with bleed/balance resistors) to handle it as well...

Just felt like I was going off the deep end there. I'll make sure I give back the original parts, in case he decides he wants to run 7581s again, sometime...

Regards,
Gordon.