Tube substitutions in Heathkit W5M

[billinchile]

The specs for my W5M call for a pair of 12AU7 tubes. Since I couldn't find them here in Chile, I installed what was available: one NOS RCA 5963 and one JAN Philips 5814A. The Duncanamps page indicates that these 2 tubes are not direct replacements for the 12AU7. The high frequencies are very attenuated, and I would like to know if this might be related to using these tubes instead of the originals. Thanks for any advice.

[SY]

I don't have the W5M schematic at hand; IIRC, this tube was used as a diff amp/driver? In any case, the 5814 is a close replacement, but it does draw more heater current. If you've got a series heater string, this could be a problem. With parallel heaters, no big deal. I've never used a 5963, but I have a dim memory that it might have been a remote cutoff triode (not a good thing).

How ambitious are you? If you're handy with a soldering iron and socket punch, you can convert the 9 pin socket to an octal and drop in some 6SN7s. The distortion can be considerably lower. With some rewiring, the 6FQ7/6CG7 can also do an excellent job.

[billinchile]

This is the schematic for the amp: http://www.angelfire.com/vt/audio/heathw5m.jpg
and here is a spec page for the 5963: http://tdsl.duncanamps.com/show.php?des=5963

I would like to keep the amp in the ballpark range of original and see if I can get it to perform correctly, although your suggestion would no doubt be an interesting project, and I like the 6S7N (Heath used them in the W4AM).

Could the high frequency attentuation have to do with initial resistors in the signal path right after the input? I read somewhere that the values for these were often chosen with the intention of reducing LP background hiss.

[anatech]

Hi billinchile,
I'd look at the 4K7 and 270pF combination across the plate resistor of the input tube V1. If you look around, you may be able to find 12AU7's in test equipment (old stuff). VTVM's normally use one of these. They shouldn't be very expensive new either in the States.
-Chris

[SY]

Well, it might take me a while to do a complete pole analysis, but 15K is a bit big for a grid stopper. The first one may be there for RF suppression. Just as a quick experiment. short out the first 15K, then parallel the second one with a 4.7K and see if that brings the highs back.

[billinchile]

Originally posted by anatech
Hi billinchile,
I'd look at the 4K7 and 270pF combination across the plate resistor of the input tube V1. If you look around, you may be able to find 12AU7's in test equipment (old stuff). VTVM's normally use one of these. They shouldn't be very expensive new either in the States.
-Chris

Do you suspect that those values might be inappropriate? Should I eliminate (or change the value) of R5 and C11?

[billinchile]

Originally posted by SY
.... 15K is a bit big for a grid stopper. The first one may be there for RF suppression. Just as a quick experiment. short out the first 15K, then parallel the second one with a 4.7K and see if that brings the highs back.

Thanks. I'll give it a try and let you know what happens.

[billinchile]

Well, the "experiment" made a really big difference - a great improvement. The highs are much better and (curiously) the lows are too. For me, the change is a keeper.

A question: the negative feedback loop is connected to the 16-ohm speaker post - and I never use it. Should I move the connection to the 8-ohm post?

Thanks again.
Bill

[anatech]

Hi Bill,
Don't change the feedback point. It determines the amount of feedback for one. Although you don't use 16 Ohm speakers, the transformer is still loaded.
-Chris

[SY]

You can change the resistor to the 8 ohm tap as long as you reduce its value by about 30% to keep the gain (and the feedback) unchanged.