Fun with Feedback

[Cressy Snr]

Well I’ve had the break and fixed the amp by replacing the offending part. Blown mains TX was the thing that had stopped it working; no further damage.

Glad you like the design Mike.

[izzy wizzy]

As I found out to great cost, fuses in secondary can save iron from destruction. Primary fuses not so much. It's a small thing as you don't want it to happen again for whatever reason ...

[Cressy Snr]

As I found out to great cost, fuses in secondary can save iron from destruction. Primary fuses not so much. It's a small thing as you don't want it to happen again for whatever reason ...

Cheers Stephen, I have a couple more things to do to tidy up loose ends and will incorporate fuses in the HT line.
Wouldn’t have thought to do it….until now that is.

[Nick]

Yep, a fault in a single secondary in a multi secondary transformer may draw enough current to kill the transformer but not enough to blow a primary fuse. As I found to my cost.

[simon]

Mmmm. So should we use fast blow fuses in secondaries? Or is a close tolerance to expected current draw more important?

[Nick]

I think slow is fine, its the thermal inertia of the windings that give the overload margin. Won't cover marginal overload I guess, but that rarely happens. Also won't cover a shorted turn, but its better than nothing. A thermal fuse in the primary side embedded in the windings would be the safe (as in CE, protecting the user, not the transformer) option.

[izzy wizzy]

Cheers Stephen, I have a couple more things to do to tidy up loose ends and will incorporate fuses in the HT line.
Wouldn’t have thought to do it….until now that is.

Just saying.... waiting until I tidied up a few things is how I lost mine.

[Cressy Snr]

Well yes, It appears there are one or two other issues that require further investigation of the power supply. Something has happened to the PSU, resulting in LF instability. OK on one pair of speakers but not on the other, which is no good either way. May have a rogue time constant in the output stage. Whatever it is, the test gear will have to come out again.

[Nick]

You have two about 1s time constants in the power supply decoupling.

[Cressy Snr]

You have two about 1s time constants in the power supply decoupling.

Yikes! Yes, thanks for saving me the calculation. That’ll need sorting.

[Nick]

Well, 0.75s and 0.82s, but them both being about the same is itself more likely to be a problem.

[ed]

Well, 0.75s and 0.82s, but them both being about the same is itself more likely to be a problem.

I'm scratching my head here, what RC figures were you using?

[Nick]

4k2 and 50uf for example

[Nick]

4k2 and 50uf for example

Though just released I was quoting the frequency not the period. So 0.75Hz and 0.82Hz. My bad.

[ed]

4k2 and 50uf for example

Though just released I was quoting the frequency not the period. So 0.75Hz and 0.82Hz. My bad.

that's better cos I made it 0.2 seconds