807 vs. 1624 tetrode ?

[Miles Prower]

Originally posted by Tony
the datasheet for the 1619 and 1624 shows operating points for class AB2, are these tried and tested?

Yes. There are lots of 1624-based designs out there. Some good; some not so good. But, yeah, it's been done lots of times.

Also, like in the case of 4D32, class AB2 seems to shows lower distortion percentage, any comments please....thanks...

The 1624 is similar to the 807 in that it does produce some of the lowest THD figures of any pentode even before you connect the NFB. I would expect that the 1624 is like the 807 in another regard: that possibly it likes to make high order distortion. When I did a PP 807 amp, the Twin-T test I did showed that the advertised THD spec was dead on. However, the residual after nulling was considerably distorted, meaning that there was a lot of h5 and higher in there.

Listening open loop revealed that there was considerable pentode nastiness in the background. With some program material (i.e. Ozzy's Bark at the Moon) it rose to fingernails on the blackboard irritating. With all program material, there was inevitable listener fatigue even if the nastiness wasn't obviously audible.

After a week of open loop listening, I figured that the finals could benefit from local NFB (as O. Schade recommends: feedback 10% of the plate Vp back to the input side). That cleared things up considerably, and only another 6.0db(v) of gNFB was needed to make for a fine sounding amp.

By contrast, some types make a lot more THD, but it's all h3. I found this to be the case with 6BQ6GTBs: about 3.0% THD, but in the Twin-T test, it looks almost completely sinusoidal with maybe a trace of h5 or higher. Listening to these open loop doesn't give any pentode nastiness, just an overly "aggressive" or "edgy" sound (and, of course, sloppy bass due to woofer underdamping). That project wound up with 12db(v) of gNFB, adjustable. Sounds best with about 6.0db(v) dialed in since the whole 12db(v) sounds a bit "solid statey" (although good for some types of music, but hard rock definitely sounds "subdued"). No need for local NFB with 6BQ6GTBs (or 6V6s for that matter -- another type that produces mainly h3).

[Tony]

Hi Miles,

the GE 6BQ6G/6CU5 are on sale at a dollar each at Rogalski's....have you tried them in place of the GT's? thanks...

[Tony]

oopss, make that 12BQ6GA/12CU6...

[Miles Prower]

^^^^

I have a couple of the 6BQ6GTAs. These sound the same, but don't seem to be as robust. Running the 6BQ6GTBs with 50mA of Ip per tube gives a Pd= 17.5W (12W rated) and no red plate when viewed in the dark (RCA grey plates; the Sylvania black plates took up to 70mA without showing color, the RCAs glowed faintly at that plate current) The 6BQ6GTAs, however, red plated at 50mA of plate current. Sounded just as good, but probably wouldn't last as long running that hot.

When using the 'GTAs, probably best not to exceed the plate dissipation spec. Of course, that won't sound as good since it'll be closer to Class B. Running the GTBs hot definitely helped the sonics.

Can't say about those GAs. From the pic, those definitely have a fatter bottle than the GT versions. With the GTs, the glass fits completely in the black plastic octal bases.

Yeah, 6BQ6GTBs are cheeeeeeeeep for what you get. Although we keep talkin' 'em up and they'll get more expensive. The 12.6 volt versions will be even cheaper still due to the odd voltage.