Power amp ok on resistive load, but not with speaker

[Mick Bailey]

One of these where everyone is going to say "don't you know anything you stupid boy - didn't they learn you nothing at school"

Anyhow;

Got a power amp module that works perfectly into an 8R resistive load with a bulb limiter. DC offset 14mv, nice sine output, +/- 54v power rails.

But, connect it to a speaker (and I've tried three) and the bulb flashes twice - once briefly as the caps charge, then brightly for a few seconds with DC on the output; it goes +10v then swings to -10v. It then settles down and all is fine. If the speaker is connected after the amp powers up, no problem. Seems like the inductive kickback from the speaker may be the issue. Any ideas?

[woodyc]

A hf load might help, like C14 and R15 in this pic...

[plec.22]

one simple idea would be to add a relay timer circuit with 3 second on between the o/p and the speaker that way the amp circuit will balance out nice. Try hooking up a relay plus 10 ohm 50w resistor as a tester. one thing to add here 55 volts for one pair is a little high bring down to say 45 per rails.

[plec.22]

that circuit looks like a mackie amp...

[g1]

Does it make any difference whether you use the bulb or not? Is this a project or an existing amp?
Sounds like the cone is in motion all the time? So more like AC at very low freq. like 1hz or less?
There are some amps that do this when turned on (cone moves one way then the other, whump! whump! sound) and sometimes it is normal for the design.
Is it so loud you are worried it will damage the speaker, or just an annoyance?

[Mick Bailey]

It now looks like it's fixed - C3 and C5 tested high ESR. Replaced them and the amp now kicks in just fine. I can't see why this would be affected by an inductive speaker load rather then a resistive load, though. Thanks for the ideas.

[Enzo]

I am with g-one here, if you had tried it without the bulb, it may well have worked.

The bulb is there to prevent the amp from burning up when something is shorted. But if the amp can power up without trying to blow a fuse, then is the time to get rid of the bulb. The resistance of the bulb in series with the amp primary can cause all manner of performance issues.

The fact you fixed it with a cap swap, may only mean something was marginal making the amp less stable when not fully powered.

[J M Fahey]

Agree and add: many amps are "stupid" for a second or two, at turn-on, until all caps charge to a reasonable value.

Plus many amps are quite unsymmetrical, relying on heavey feedback to correct this.

Without bulb, capacitors charge quickly, and amp atabilizes after a thump.

With current limited supply but no load, about the same thing.

But with load, DC (even if momentary) at the output causes high current through the load, which forces that rail down *a lot*, so it may latch towards that side and never recover ..

[Gtr_tech]

I've gotten away from using a variac (I know not a limiter but similar situation) to "smoke test" SS pwr amps because strange things can happen when everything is not up to proper operation conditions. I've actually had an old Standel PA amp not pass the smoke test because of the variac. Replace the parts, power up, set bias...happy as a clam.

[Enzo]

I only use the variac far enough to determine if the thing is going to draw huge. Once I think it won't kill fuses, off the variac goes.

[Mick Bailey]

I didn't try it off the limiter (150w) before changing the caps, but I'm happy to have checked them as they were over what my ESR meter will read - 30 Ohms. Probably the real culprit was the bulb and as Enzo suggests and the cap swap has made the situation less marginal. I still don't get why it worked with a resistive load, though - no problem at all. The only problem was with the speaker connected, so my theory is the speaker cone moves one way as the amp comes up with unbalanced supplies and creates a blip on the output. As the cone moves back it creates a reverse EMF that gets picked up through the feedback loop and momentarily swings the output the other way until both supply rails are stabilised by the supply caps charging fully. Or is that a 'flat-earth' theory?

Incidentally, a couple of years ago I had a 'fuse blower' with blown output and driver transistors etc. I thought I'd fixed it. When I re-tried with a bulb limiter, it it still had heavy DC on the output and the bulb glowed brighly. I spent hours with that thing until someone suggested powering it up directly. It was fine, but I took a chance. Shows how unreliable a bulb limiter can be with certain amps.

[J M Fahey]

Shows how unreliable a bulb limiter can be with certain amps.

Or rather, "Shows how unreliable certain amps can be with a bulb limiter" , meaning certain amps are marginally stable , in this case as far as DC working voltages.

Which is proven by the failure mode: latching up.