help repairing Hafler xl280

[Truckee_Tango]

Hi, I am trying to repair a friends Hafler XL280. One channel works fine, but the other is low in power and distorted. I have scratch built several amps but never repaired a MOSFET amp. Here is what I found:

- Power supply rails appear good at +- 64v and that voltage appears on output FETs both sides (drain)
- Power to the input stage from the single bridge for both channels appears to present +- 63.5vdc. That should be 75v?
- Fuses are OK and no parts look burned or damaged
- Output FETs on good side show 0vdc on case, but bad cannel shows votages varying from -1.4vdc to -2.3vdc on both n and p sources.
- Input to the output transistors with no signal is 0 on good side but varies from about -1.4 to over -2vdc on bad side on both the n and p gates.

I have downloaded the instruction and assembly manual, so I have the schematic, but there is very little troubleshooting information. The low voltage off the input stage bridge sure looks wrong, but I would think it would affect both channels equally, still the left channel sounds OK.


For checking other input board voltages, should I take the heatsinks off the chassis?

Any help would be hugely appreciated.

Marshall

[mrshow4u]

Sounds like the MOSFET's are open. Do they have "grid stopper" resistors on the gates? Check those. If you've got a scope, check if the gate signal is large while the Sources don't budge.

[anatech]

If those gate resistors go open, expect a ton of current draw. Generally resulting in a bunch of wire that looks exactly like output mosfets.

So power down, discharge the filter caps and check continuity on the outputs.

Watch for gate to drain shorts.

-Chris

[Truckee_Tango]

Thanks mrshow4u and anatech for such fast responses. Wow, this forum is great!

mrshow4u, the output fets do have gate resistors, I assume for load sharing. I've never heard the term gate stopper, but assume it is like the "grid stopper" for the tube gear where I have way more experience (started with dyna and heath before transistors were born!). I do have a scope and signal generator. I assume you are saying that I might see a good signal on the gate when I drive with say a 1k signal, but would not see one on the speaker output side. There is some signal coming through, but nowhere near what the other channel has. Could some but not all of the output fets be bad? I'm not sure I have equipment to test the fets.


anatech , The grid, oops I mean gate resistors are all OK, and there is no mass of wire melted together. By the way, after idling 10 minutes or so, the bad channel's heat sink was pretty cool, while the working channel's HS was just a little warm.

Can you tell me what you mean by checking continuity? Do I check for continuity between gate and drain? Should I be pulling the fets and checking continuity on each fet? I don't want to blow a fet that is good.

If any of the fets are bad, how do I get replacements? I understand they are very hard to get.

I'll dig out my scope and signal generator and load resistors and see what gives. Any hints would be appreciated.

Thanks a million guys!

[anatech]

Hi Truckee_Tango,
It sounds like your bias network is misadjusted or shorted.

Put your common meter probe on the output line. Measure to each gate line and tell us what you have for DC voltage.

The mosfet is a depletion mode device like a tube. If that resistor goes open, the gate floats high and the transistor conducts heavily, or oscillates. Either way, the transistor normally shorts.

-Chris

[mrshow4u]

The "grid stopper" term is borrowed from tubes. Actually I think the signal should be at the gates, large but ugly. The feedback loop should be trying to compensate for the low output by making the signal much larger at the gates. I think the signal will be ugly at the gates because the loop basically sounds open at this point, and even open loop will not satisfy the inverting input. I think you might get something very "squarish" at the gates. Oh, that reminds me. Check for clampers across the gates. It might be something like a zener and a signal diode in series wired across the gates. I don't have the schematic, but some sort of voltage clamper. Just another thing to check.

[Truckee_Tango]

mrshow4u There is a 1N4003 diode from what should be a higher voltage B+- to the front end to the lower 64 volts going to the output fets, but other than that no clamper that I see. The schematic is under the www.hafler.com site under tech support and archive tabs. I could send you the .pdf file, but it is a 4mb chunk. I have not checked those diodes, but it is suspicious that I don't see the 75 volts to the input section the schematic shows. The input section voltage comes from a common bridge and the outer tranny winding tap, while the 65V output fet voltage comes from separate heavy duty bridges one per channel. Output from both bridges measures within a volt of each other, but if those diodes are shorted, that could be a problem. However, the one channel sounds OK.

anatech , I measured from the speaker output terminals to the gates of each of the output fets and found the following:

good channel : one set of fets showed +.75vdc and the other showed -.45. I'm not sure without checking which fets which were p and n.

bad channel: for each fet, the voltage from gate to speaker + output was a +vdc on the output terminal and -vdc on the gate. The voltages on the gates (respective to the output terminal) changed on one set of three from -0.8 to -5.0vdc and on the other set from -0.3 to -5.vdc. Hard because of the fluctuation and erratic nature to tell what the min and max really were, but that was close. One of the fets appeared to be going from -0.13 to -1.3 while another in the same set went from -0.3 to 5.something.

Haven't dug out the scope and sig gen yet.

Thanks again for great responses.

Marshall

[mrshow4u]

Cool, thanks for the schematic link. For some reason, I thought they'd be chilly on posting that stuff. That string of diodes D1~D4, you'll want to check those for shorts. As Anatech say's , check bias. Measure from base to base of what looks like Q15 and Q16 (double check those numbers). With no audio, I'd expect 2~ 5 volts across those bases. That will depend on the Vgs of the output devices. Make sure you check the voltage on the Drains. Most older Haflers will put out low distorted sound with an open rail fuse.

[ifrythings]

I just spent 6 hours fixing a yamaha amp ...

Anyways I would check all the part in the picture that are circled red out of circuit (aka remove) exept resistors. oh and check the fuses.

BTW I would remove the primary fuse and use a 60W lightbulb for saftey of blowing up more stuff (cheap current limiting)

[Dick West]

Truckee,

Read the manual. It should have a table of voltages expected at different parts of the circuit. It also has a picture of the back side of the PCB. With this info and a Variac, you can measure voltages at different parts of the circuit. This will help in your trouble shooting.

I hope the trouble with the bad channel is not in the 4 input jfet transistors. They must be matched and are no longer available -- except at a high price from Erno Borbely in Germany. The other transistors are fairly common.

The PS for this amp has one bridge and small pair of caps for both PCBs but has a separate bridge and large PS caps for each channel. The PCBs run at ~75VDC, the output section at ~65VDC.