Phoenix Gold Phorum

OK, so i am revisiting a project that I got busy and put aside. It had blown FETs when I got it, and one had been repaired. I replaced all FETs with IRFZ48 and the 33 ohm resistors were replaced with 330 ohm resistors. I replaced the power supply caps too as they were leaking. The caps are good (tested continuity, all is good) but when I test the amp with power I smoke FETs. I tried removing all FETs and leaving only one per transformer and it still happens. I removed the rectifiers to separate the power supply from the rest of the amp, and the FETs still smoked. I have also made sure that none of the vias were damaged or shorted. I also keep using brand new FETs rather than reusing damaged FETs.

So I have narrowed it down to my power supply, and with the CAPs testing good all I can think of is to go back to the original FETs and resistors. Which ones were original. I cant find any of the original parts that I removed, so I hope that a collector will know or is willing to look and see what they are.

The SG2535 PWM controller may be damaged, and applying full on to the FETs, which will smoke them fast.

Often when FETs die, they overload the PWM controller and damage its outputs.

With no FETs installed, you should power the amp on and measure the gate pin locations for the FETs. Let us know what the DC and AC voltages are for these pins.

Another possibility is a short in your transformer somewhere. This too will smoke FETs almost instantly.

i have at least 6 zpa0.5's done with the irfz48 and 330 ohm gate resistors in the ps, and they work as fine as they can be......

So unless you putted in the wrong irfz48's, i can not believe you will solve the problem with changing those again.........

If you power up the amp for testing, witch fuse do u use ?? maybe it is better to power up with a 5 A (automatic) fuse

until you found the problem......

and the more you replace the same components, the more the traces there will get damaged.............

So for AC I'm getting very low in mv range, dropping to zero. For DC I get 11.75v at each gate, and 12.58 for constant power. I only have 2 bad traces but they were damaged when I first got the amp and I am aware of what I need to repair them.

And just for shits and giggles I checked the gate resistors, they range from 328.6 to 329.4 ohms.

All was checked using my Fluke 87 V. It was calibrated 3 months ago.

Where is the PWM controller? I don't see it, I imagine that I am missing something or maybe someone replaced it already, which would possible make it easier to narrow this down.

So I assume that this is the daughter board with the PWM because the traces go from this board to the gate resistors. It doesn't appear to be tempered with based on these pics in that the soldering all appears to be OEM.




But then I saw this and it may or may not appear too sloppy to be OEM


Either way, the chips on this board are ka3525a and ka339.

Also, there was a WIMA cap on the back of the board. Months ago when I was working on this I repaired some damaged leads going to it.





Now, upon further inspection, it looks like that was a factory repair. Here is why I say that
Here is one side


And the other side ( where the WIMA cap is on the reverse side)


Notice that there s location for it. Upon yet further observation I discovered that the WIMA is soldered to the rear side of the resistor rather than using those repaired leads to run it where the others are.

Question: should I move the cap back to where it looks like it should be now I have repaired those leads? Can anyone confirm that their amp has it on the front? I want to put it where it belongs but I'm way more concerned with getting the amp working, this is just a side job.

if you repaired you're self those two traces of the wima cap, i quess you went wrong there...;

i should open an amp again to be sure, but as i remember, most of my zpa0.5's had that wima cap

on the rear and both traces were interrupted..... although i should look again.........

most of the pics you post show that this amp never has been repaired before........

Yea you haven't seen this pic yet... This was also how it was when I bought the amp:




Ok, so I will put that lead back to how I found it. I didn't move e cap yet, at first I repaired those leads because I thought those were traces going to the cap.

i have indeed seen this before on some of my amps, and i always decided when one fet was burned (single repaired or not)

i replace all 16 of them... (with irfz48)

but did you check for sure if there was no short between upper and under side of the board on this place because of the burned spot ?

cause i would check that out to be sure..........



update, i just checked the pics of my last zpa0.5 amp i did and both traces were interrupted for sure......... so indeed bring it back

like you found them !!!!!!!! (talking about the red wima cap on the backside).......

i'm going to sleep now.......good luck with the amp................

Eric D wrote:The SG2535 PWM controller may be damaged, and applying full on to the FETs, which will smoke them fast.

Often when FETs die, they overload the PWM controller and damage its outputs.

With no FETs installed, you should power the amp on and measure the gate pin locations for the FETs. Let us know what the DC and AC voltages are for these pins.

Another possibility is a short in your transformer somewhere. This too will smoke FETs almost instantly.

AC=0 (very close to 0)
DC= 11.75 v

What am I looking for?

The KA3525 is equivalent to an SG3525.

I am confident yours is bad. With no FETs connected, the PWM controller should max out at a 50% duty cycle, which would be around 6V AC, and depending on the shape of the output signal, around 6V DC (both numbers assuming around 12V input to the amplifier).

If yours has no AC output voltage, and measures over 11V DC, then it is just turning the FETs fully on, which will fry them instantly.

Generally when I replace these PWM controllers, I install DIP sockets in their place, so any change later on will be easier to perform.

That makes sense. Yea I was considering sockets, glad to know that's something that is done. I Already ordered the PWM, FETs and new resistors. I'll also get some sockets (locally) and the wire needed to do the repair for the damaged leads. I am excited to finally get this project cometed.

Thanks for the input.

I have found it is best to clip the leads on the IC one at a time, then pull them out one at a time. If you try to get the whole thing in one shot, it is easy to damage the board.

So I got the board off. Wasn't easy but not that hard. I used a solder sucker and it made my life very easy. For the old PWM I clipped the leads and removed them one at a time.

my 4 zpas have the Wima cap on the bottom. I also have used the irfz48's, never a problem. Even with bigger caps on the rails..still no prob. No matter what I do I manage to damage the via's, from factory they are very thin.. so I used a via repair kit that has dif sizes in it, I drill out the bad via's, press the via tube in, swedge it in place for a tight fit to the boards traces, then reinstall/solder everything back together.. they are way thicker than og,and are a permanent fix. Factory via's is just a copper plating through the board and seem to be really only good for 1 disassemble. So now I use those repair via's on every high current component on my zpas,ms, orions etc..

So I replaced the PWM using a socket so it can be easily replaced in the future. I then checked the gates with power to the amp and got around 6 V AC, a little over 4 V dc. I then put a fet on each transformer and powered up the amp. I got life!!! The power lights worked for the first time for me on this amp.

So now I'm burning in the rest of the fets. I think I wil replace all gate resistors too considering I have blown so many fets with them now. I have the correct ones, including the tight tolerance. I was considering going with a tighter tolerance resistor but $8/resistor is insane.

Once that's done I will adjust DC offset. Then I should be good to test it in my car. Once it's up I will use it to power the dyns while I replace the fans on my Mac amp. If I like how this one sounds I might keep it in there and either sell the Mac or use it in my wife's car. I don't think the mc440 is worth all that much these days.

Thanks for all of the help guys. Any advise on setting bias and DC offset? I see two sets or yellow/orange dials, which ones set the offset? I could figure it out but if anyone knows off hand that would be great.

The controls mounted on the daughter cards are the DC offset controls, and those mounted on the main board itself are the bias controls.

If you have not changed any of the output transistors, or their driver transistors, I suggest you don't change the bias setting. It would be a good idea to adjust the DC offset for 0V though, as changes to the power supply might shift this slightly.

Be sure to make your adjustments after the amp has warmed up a bit. You can connect a DMM to the outputs to see where it is at on initial power up, and then tweak it once warm.

Thank you so much Eric. Is it easier to use my oscope or just the dmm?

I use my DMM, but you could also use an oscilloscope if you wanted too.

Well I haven't had time to do the DC offset, but it did put everything back together. I went high end on the thermal paste($10/syringe and 5 syringes used). For now this amp is staying in the white casing. I was considering selling the white, in which I would have swapped the original amp boards back.















Here it was in the black casing which now has the board from this white case. I swapped them because at first I wanted only the black one and I didn't think in would ever finish this amp.



I'll post pictures of the dmm when I do the DC offset.

Ok I got the DC offset fixed. I have one before picture, but both channels had 0.165v of offset when I started. The adjusters were set all the way in one direction so I think someone was playing with them or maybe thought they were gains or something.

Did you let it run warm or immediately took numbers? If not then your 1mv offset will drift once the amp has warmed up.

I let it warm for 15 mins. Everything I read said to give it 10 mins. I did hook up the dmm immediately and I watched to see if it would drift. It went up by a few mv, nothing drastic, but I waited for it stabilize before making adjustments.

trickyricky wrote:Did you let it run warm or immediately took numbers? If not then your 1mv offset will drift once the amp has warmed up.

Do you recognize the white case? Once this amp is up I will swap them back to their original cases before this one get used. I think it will look cool with 1 black ZPA and 1 white.

So I got the amp into my car, replacing the McIntosh. Its a great amp with lots of power. I do have to say that the McIntosh does have a better sound. My plan was to replace the fans of the Mac and put that in my wifes car, but im not 100% sold on having a ZPA0.5 to run my fronts and one for my sub. I really wanted it to sound better than or as good as my McIntosh. I will repeat, it doesnt sound bad, probably better than any other amp i have used, but its missing something that the Mac provides.

I was always the person that called BS when someone claimed that they could hear a difference between 2 good amps.

any input. were my expectations set too high?

Can anyone else hear a difference? I too have played with OG Mac's and I found the sound of a ZPA more desirable.. It definitely has more power, gob's of it. With it's discrete op-amp less design it's a purist amp.

ZPA amps only have a single power supply, and a common belief is any purist amp has to have dual power supplies to improve stereo separation. I think a lot of McIntosh amps have dual power supplies, which may give them an edge.

It is like two people worked independently to develop the ZPA or something. Or maybe one person designed it, and then marketing or purchasing stuck their requirements on it so the design was severely compromised.

Having no crossover, balanced inputs, and all discrete components are key features of an audiophile amplifier. But, having only a single power supply, different rail voltage capabilities, and fans built in are all key features of a bass only amp.

If someone is looking for the best class A/B amp out there for their subwoofers, I think the ZPA could quite possibly be it. However, PG has many better choices out there for running component speakers than the ZPA.