Amp with high pitched static noise

[Diabolical Artificer]

Ditto. The solder looks more like paint, sorry but if a leg gets moved or looses contact whilst powered up, it's goodbye Mr FET into the valley below, cup of coffee or not.

Andy.

[G.Castle]

Perhaps the OP is trying to use lead free solder on that board, it was most likely originally assembled with leaded solder.

[spacebiscuit]

Hopefully this looks better - I'll try the big switch-on later!

[Radio Wrangler]

Better, but if it's tin-lead solder it still looks dry, or else it's still got plenty of the lead-free stuff in the alloy formed.

David

[Analogue man]

I have found that a number of lead/tin solders do produce joints like that, I spent a pretty penny trying to find one that produced the nice shiny joint we all remember from the past and have found it impossible to find, an iron hot enough for lead free solder seems to work best. I gave away my last reel of lead free solder to my son who must use it by law unless he works on military, medical or aviation based equipment where lead/tin is still mandated.

[spacebiscuit]

Powered it up via light bulb tester, the bulb came on and then dimmed so I powered off.

K2221 in tester now says unknown or damaged part, the 2SJ352 now a diode so I guess another 20 quid up in smoke. Should at least be able to raise a dispute for these.

[spacebiscuit]

An update, sorry for any repetition...

Powered it up via light bulb tester, just the single channel connected and the bulb came on and then dimmed so I powered off.

I then disconnected this channel which I had connected via the good channels earth, positive and negative. I then connected both channels via their leads, first time I had the bulb light at power on which then dimmed. I then tried a 2nd time and when I power up the bulb is on bright for a fraction of a second then goes off, same as when I powered up with just the good channel. I also observed this behaviour when I first switch on my soldering it on via the tester.

The new mosfet are hot though despite only being powered up for a few seconds, not so on the good channel so I guess that doesn't bode well.

When I try to test the mosfet they are reading as a diode and a capicitor now, although they are on the board.

Not sure what my next move should be?

[Analogue man]

MOSFETs are very static sensitive and possibly you may have damaged these when you put them onto the board.

[spacebiscuit]

I by-passed the bulb-tester and with the single pair installed the fuses are not blowing.

If nothing else I guess this tell me that I don't have a short and the previous fake MOSFETs were the causes of the fuses which were blowing?

Do I continue to add the other pair, I'm wondering why the first pair are so hot. Is the 2nd pair required, do they work in tandem or are they hot because they're already toast?

[Diabolical Artificer]

I'm wondering why the first pair are so hot.

Measure voltage gate to source quick, power off, be careful not to slip with your probes, write down, repeat for good channel.

Andy.

[spacebiscuit]

Does this setup look ok, I've used the retractable hooks from my component tester which are then attached via crocodile clips to the DMM.

Red to gate, Black to source. DMM set to 20v.

Not powered it up yet until I know this is safe

[G.Castle]

Looks like you're hooked onto the drain with the black lead there Spacebiscuit.

[Diabolical Artificer]

Looks like you're hooked onto the drain with the black lead there Spacebiscuit.

To me too. Mosfet pinout is GDS as my esteemed colleague says you have your black lead on the drain, move it to the left ( as in pic) to the source.

Well done for being cautious, Andy.

[spacebiscuit]

Seems in this case it is GSD which checks-out against the datasheet.

[Welsh Anorak]

When I'm running a power transistor away from the heatsink like this I find the old traditional bulldog clip (still easily available) can act as a heatsink for a brief period - long enough to prevent the transistor running away while your meter registers.

One for each transistor, of course, and keep them isolated.

[G.Castle]

Ooh, so it is! That's a little unusual.

My bad for not checking.

Do a sanity check and see if you have the transistors on ghe correct sides of the board

[spacebiscuit]

Yeah 100% they're the correct way around - checked against the other channel.

Good tip re: bulldog clip, got a few knocking around somewhere

[spacebiscuit]

Results are:

2SJ352
New: -0.38v
Original: -0.35v and -0.36v

K2221
New: +0.70v
Original: +0.33v and +0.34v

[Diabolical Artificer]

My bad too, pinout = GSD.

Ok, re Vgs of bad channel replacement's, not much there with those DC measurements I can see that would explain the fets getting hot; does the 2SK221 get hotter than the 2SK352?

.Next try measuring Vgs again but with your meter set to AC volts - 20v setting. Next measure the resistance of R17 to R20, amp off, then with the amp powered up briefly, measure the voltage across each R*, meter leads clipped onto either end of the R *( or just the resistors that go to the fets you have inserted.)

Andy.

[spacebiscuit]

As requested:

K2221:
DC 0.73v / AC 0.44v

2SJ352:
DC -0.55v / AC -0.13v

R17: 02.3 (off) / 0.23 (on)
R18: 0.23 / -0.58
R19: 0.24 / 0.24
R20: 0.26 / -0.50

Since only R18 and R20 are connected for the 2 mosfets, voltage:

R20: DC -0.02v (off) / AC -1.8v (on)
R18: DC 0.03v (off) / AC 1.9v (on)

One FET isn't really any hotter than the other, they're probably about as hot as a cup of tea that has cooled enough to drink - not quite too hot to touch.