Ampeg SVP 1600 POOF!

[drewl]

I guess the P in SVP stands for POOF!

Channel A blown big time, all ten mosfets, and accompaning .47 ohm, 47 ohm and 1.5K resistors all fried.

I can just install a pair of the mosfets (one N one P of course) and the associated resistors to test it after rebuilding, right?

Just double checking

Looks like the loose washer, meant to be used for the power trans, did it in.

poof.....

[nevetslab]

Well, I think by now you know the answer on that one. This looks like the same output stage in an SVT4-Pro, though this only having four pairs rather than five pairs of MosFET.s They need to be batched and matched so you get even current distribution into the load. Thus far, I've not been so lucky in buying 25 or 50 pairs and finding no need to match parts from the parts received from Mouser. I do get a range of parts that yield sets, and those that fall outside of the min/max curves of Source Current, I set aside after marking them on the heat spreader to my recorded charts.

I've found even though the Vgs readings of the P-Ch and N-Ch aren't equal/opposite polarity, that doesn't appear to matter. I do try to get similar current curves of the two sets, though have used dissimilar batches without any failures.....I think. I have had one amp before fail after replacing the full compliment of N's and P's a few months later. What caused it? No clue for sure.

[drewl]

I'd guess many are blown up by using Bridge mode incorrectly!

[nevetslab]

Loose washer? One of the many used to hold the MosFET's to the heat sink? Ouch! And, of course, driving 2 ohm loads in Bridge NOT a good idea!

[J M Fahey]

I've found even though the Vgs readings of the P-Ch and N-Ch aren't equal/opposite polarity, that doesn't appear to matter.

They are called similar names (240 - 9240) probably for commercial reasons and are regularly used together but they are NOT true complementary (as in "mirror image") by any means.

Nch and Pch are very different, on many parameters, and in fact Pīs should use "one more extra transistor" on that side to truly balance the Nīs.
Nobody does because in any case they are way over rated for their job, as far as voltage and current goes, so they get away with it.

[nevetslab]

They are called similar names (240 - 9240) probably for commercial reasons and are regularly used together but they are NOT true complementary (as in "mirror image") by any means.

Nch and Pch are very different, on many parameters, and in fact Pīs should use "one more extra transistor" on that side to truly balance the Nīs.
Nobody does because in any case they are way over rated for their job, as far as voltage and current goes, so they get away with it.

Good point. I do always get nervous when I see a set that's been in use for years, and the idle current thru each device is all over....nominal idle current should be 25mA per device, but finding one running two or three times that current along with others at 1/3 that, only a matter of time before the current hog part goes poof. The N-Ch IRFP240's are rated considerably higher at 25 deg C...20A or so, while the P-Ch IRFP9240's only 12A @ 25 deg C. But, as you said, way over rated to survive the task (until they fail, of course).

[J M Fahey]

The N-Ch IRFP240's are rated considerably higher at 25 deg C...20A or so, while the P-Ch IRFP9240's only 12A @ 25 deg C

That.
As long as none either polarity is actually asked to deliver 12A or higher, it will work, but that always scratches me the wrong way.

I typically use 4 TO247 MosFets for 300W into 4 ohm , +/-65V rails, BUT since forever used quasi-complementry outputs.

Normally 4 x IRFP250 (which have NO P Ch equivalents and are beasts) but last time I ordered 200 IRFP240 from Mouser because thatīs what they had at the time and for distance/freight/customs problems I order as many as I can in a single purchase.

Pay and suffer the same whether I buy 8 or 200 so always try to be stocked for a long time.

Oh well.

[drewl]

Yeah, we have a ton of 250's and 450's at work, but no P channel equivalents.