Toroid hi fi OT for guitar amp?

[jimboyogi]

Hi all,

just wondering if anyone here has ever tried using one of these
http://www.antekinc.com/pdf/AT-020L.pdf

for class AB1 guitar amp duties. Or if anyone can advise whether it is worth a try.

I guess I have 2 concerns -

1. Frequency response. These transformers have a much wider freq. response (10 - 50KHz) than is required for guitar, or than any guitar OT would have. Might this lead to high freq. instability, or just bad sound ?

2. DC saturation. As a toroid, it is likely susceptible to dc saturation due to imbalance in the PI and output section. I see some imbalance as inevitable (and desirable?), but I'm wondering if a reasonably balanced LTPI and matched output tubes will get me close enough to avoid saturation on one of these? Will I have to avoid 2nd harmonic distortion so much that it sounds awful ?

[Steve Conner]

I have no idea... Do it and be sure to post the results here!

The wider frequency response won't lead to instability, it'll have the opposite effect, making the NFB more stable.

Whether it hurts the sound is a matter of taste. One of my amps has an output transformer from a PA amp, that's interleaved about 8 times for a really good HF response, and I like the way it sounds. A guitar speaker cuts off everything above a few kHz anyway, so does it matter? But the big chunky bass you get from a hi-fi OT with a generously sized core, you probably WILL notice.

Even if the tubes were perfectly matched, it's possible for the preamp to produce an asymmetric drive that makes one tube conduct more than the other when the output stage is overdriven heavily. That's the 2nd harmonic distortion you were talking about, but it's not an issue before the output stage starts to clip, because in the absence of clipping, the coupling capacitors adjust the signal for a net DC component of zero.

A toroid might react strangely to that, but I don't know for sure. I did some work on saturation a while back and it made an interesting thread here, maybe you could check that out. Do Antek specify a maximum DC imbalance current?

[Merlinb]

2. DC saturation. As a toroid, it is likely susceptible to dc saturation due to imbalance in the PI and output section. I see some imbalance as inevitable (and desirable?),

The idle DC currents are what really matters with toroidal OTs; an unbalanced phase inverter is less important. I would highly recommend that you incorporate a bias balance pot or even a bias servo. If you let it saturate then it will overheat big time!

[jimboyogi]

Thanks very much for the replies Steve and Merlin. Once again you guys, and this forum in general, are generous with your time and knowledge.

However I don't think I'll proceed with this idea for 2 reasons -

1. I want to keep this design very simple and agricultural, so the idea of bias balancing pots or servo's don't really fit. This is steering me away from the toroid idea.

2. The design is to use with a pair of either EL84's or 6V6GT's, class AB1 PP cathode biased. Vp ~ 320V, Vs ~ 310V. When I ran a load line for the EL84's comparing 8K Za-a with 10K Za-a (close to the 9.6K for the Antek toroid), my figures showed significantly higher power out using the 8K. So I'm probably better off sticking with an 8K transformer, non-toroid, of which there are many to choose from!

[Steve Conner]

Merlin, I have a (hopefully relevant) question for the Valve Wizard. Can a bias servo still function when the output stage is overdriven?

[Merlinb]

Merlin, I have a (hopefully relevant) question for the Valve Wizard. Can a bias servo still function when the output stage is overdriven?

I think so yes, providing the integrator has a long time constant. It's not something I've tried under heavy overdrive though (makes more sense to use toroids in bass amps anyway). I suppose you could get a dodgy situation where you chug away for a while so that the servo provides very little bias (since the grid will naturally shift negative during overdrive), and then you stop chugging; it might take a couple of sends for the servo to respond and increase the bias again. A kind of anti-blocking?
Hmmm...