JPL MPX300 - Short Circuit Protection

[nickb]

I'm running through the post-repair test & calibration for a JBL MPX300 - relevant sch below.

I'm hesitant to apply the short circuit test as written as I'm convinced that it will be damaging. Basically, you run to full power and then apply a short across the output. What worries me is there does not appear to be any VI limiting on this beast, it's current only. The limit would be set to about 22A at this point in the procedure and 65V (supply) at 22A ( 7.3A per transistor) would instantly destroy output transistors. I feel I must be missing something in the schematic that would prevent this, what is it?

I'm done this my usual way for now - set the output to a little higher than rated in this case 525W against 450W and adjusted the current limit to just clip.

Test procedure & sch follow:

[Dave H]

Does that relay on the right open and disconnect the speaker if there's too much voltage? I think it could do but I'm having trouble following the schematic with the way it's drawn.

[nickb]

Does that relay on the right open and disconnect the speaker if there's too much voltage? I think it could do but I'm having trouble following the schematic with the way it's drawn.

Thanks- I admit I didn't look much at that bit. I find it much easier to think of the reference as the usual mid point of the power supply. I believe R532/C517 form a low pass filter for the bridge that drives the offset sensing and ultimately the protect relay. I don't see any voltage dependant current limiting there. Also, the later notes say the supply current will go up meaning that no relay will open. Practically speaking, a relay would be too slow to afford any meaningful protection for the transistors anyway. It protects the power transformer and the speaker. I don't there's anything there that might help.

[Dave H]

I find it much easier to think of the reference as the usual mid point of the power supply.

Me too, it just doesn't look right with the 'output' of the amp grounded. I only had a quick look at it and confused the DC offset protection circuit with voltage limiting. It says "Apply a short to the output" but it doesn't say for how long. You could cross your fingers and do it quickly. I don't think I'd want to risk it though. Does it have an over temp cut out on the heatsink? I don't see one on the schematic and I doubt it would save the transistors from a shorted output anyway.

[nickb]

Me too, it just doesn't look right with the 'output' of the amp grounded. I only had a quick look at it and confused the DC offset protection circuit with voltage limiting. It says "Apply a short to the output" but it doesn't say for how long. You could cross your fingers and do it quickly. I don't think I'd want to risk it though. Does it have an over temp cut out on the heatsink? I don't see one on the schematic and I doubt it would save the transistors from a shorted output anyway.

R516 (lower left) is the heatsink themistor. 65V / 7.3A is well out of the DC SOA - it would not survive more than about 10-20ms. It's pretty clear they mean you to keep the short in place. Also the emitter resistors would be dissipating about 6W and are only rated for 3. It makes me wonder if the power supply has been designed to have a deliberately high source impedance. I'm not going to risk it.

[Dave H]

65V / 7.3A is well out of the DC SOA

[ATTACH=CONFIG]36708[/ATTACH]

It sure is. That SOA graph only goes up to 50mA. But it is after midnight.

[Enzo]

Where did that SOA chart come from? I looked up 2SC3281 and see and SOA chart going up to 30A, not 50ma. Clearly output transistors are rated far in excess of 50ma. Turn the legend on the let to amps instead of milliamps, and it makes a lot more sense.

The output of the amp isn't grounded, this is a grounded emitter circuit, also called a flying rail circuit. Except in this one, the collectors are grounded, and the NPN and PNP are on the opposite rails compared to most amps. Note the speaker is wired between ground and the power transformer center tap. This is very common, in fact QSC has done this for decades. By grounding collectors, they can mount the TO3 output transistors directly to the grounded heat sinks without concern for insulation.

In fact this looks like a lift from QSC. The short procedure is also like theirs. Note they did not say just short the output and go nuts. They said to set up a 2 ohm load at clip, and adjust the current limiters for that. THEN they say short the output. That is to demonstrate to yourself that the limits are working. Shorting should not increase current draw over the 2 ohm load limit already set. Once your 2 ohm limits are set, you don't need to short it very long to see if the current draw spikes. Also, the amp is not putting out 65v DC, you were driving it with a sine wave to clip.

See those jumpers over near the output transistors? And note the large rectangular empty areas above and below? Looks to me like they took one of the rail switching QSC circuits and trimmed out those extra circuits

[nickb]

It sure is. That SOA graph only goes up to 50mA. But it is after midnight.

I meant to say that the chart had a typo- mA instead of A. Here's the whole datasheet:

[nickb]

Thanks Enzo,

Where did that SOA chart come from? I looked up 2SC3281 and see and SOA chart going up to 30A, not 50ma. Clearly output transistors are rated far in excess of 50ma. Turn the legend on the let to amps instead of milliamps, and it makes a lot more sense.

The chart has a typo- I forgot to say - sorry

The output of the amp isn't grounded, this is a grounded emitter circuit, also called a flying rail circuit. Except in this one, the collectors are grounded, and the NPN and PNP are on the opposite rails compared to most amps. Note the speaker is wired between ground and the power transformer center tap. This is very common, in fact QSC has done this for decades. By grounding collectors, they can mount the TO3 output transistors directly to the grounded heat sinks without concern for insulation.

The point I was making was that you can call anything you want the reference. When it comes to the output protection circuit it's so much easier to understand if you use the midpiont of the supply as the reference. That way you don't have to mentally subtract AC signals. Aside from the heatsink benefit, you also get a little gain meaning you can use a small output range opmap to drive it.

In fact this looks like a lift from QSC. The short procedure is also like theirs. Note they did not say just short the output and go nuts. They said to set up a 2 ohm load at clip, and adjust the current limiters for that. THEN they say short the output. That is to demonstrate to yourself that the limits are working. Shorting should not increase current draw over the 2 ohm load limit already set. Once your 2 ohm limits are set, you don't need to short it very long to see if the current draw spikes. Also, the amp is not putting out 65v DC, you were driving it with a sine wave to clip.

That was also my understanding. But, it will drive the limit current of 22A into whatever resistance you have, probably about 0.1 ohm from the emitter resistors plus wiring. That means the peak drop will be 2.2V so unless there is quite a bit of series resistance in the PSU the peak voltage will be 62.8V. I'm assuming 65V for the PSU for the sake of argument by the way. The current waveform will be darn close to a square wave so the transistors will see this repetitively for half the time. Anyway that is how I see it, I presume the test procedure works, but I still just don't see how I take it you've actually done this without damage?

See those jumpers over near the output transistors? And note the large rectangular empty areas above and below? Looks to me like they took one of the rail switching QSC circuits and trimmed out those extra circuits

This amp comes in several variants, I think the biggest of the range (3600W) uses rail switching. Interesting about QSC, this is the third amp this week that is a QSC lift. They either suffer a lot of plagiarism or make a chunk from custom/licensed designs.

[Dave H]

The output of the amp isn't grounded, this is a grounded emitter circuit, also called a flying rail circuit.

I know that Enzo. I was just trying to say that I'm more used to amps where the CT of the transformer is grounded and the output is from the transistor emitters (or collectors). That's why I wrote it as "with the 'output' of the amp grounded"

[Enzo]

A lot of people struggle to conceptualize that form of circuit, when it really is the same as the conventional other than which side of the speaker is ground.

[Dave H]

A lot of people struggle to conceptualize that form of circuit, when it really is the same as the conventional other than which side of the speaker is ground.

Yeah, move the ground reference to the other side of the speaker in the schematic below and you have the conventional circuit. The current wants to return to the junction of the two capacitors. There's no current flowing to ground so it doesn't matter where the ground reference is

[nickb]

Bump!

Has anyone "... actually done this without damage?"