EP2500 fault in bridge mode

[Brian Swatton]

I have an amp here that has me scratching my head a bit. I wonder if anyone else has seen this? Both channels work individually, I've had about 800w out of each of them (at limit of my loads) without any problem.

However, when running in bridge mode, each amp produces only a tiny portion of waveform peak, -ve or +ve depending on the channel. Otherwise each output sits at zero volts. So, on a scope you have a largely flat line with just the very tips of the waveform appearing. These were measured unloaded. As soon as any load is across the bridge outputs (no signal in), the "A" clip light comes on too.

Apparently, this fault appeared after the unit was knocked over, having been stood on it's side, so there's a chance of cracked SMT I'd guess. Visual inspection, unsurprisingly, hasn't picked up anything obvious. I've been trying to get my head around the bridge mode switching, but am getting a bit confused by it. It's not easy making powered measurements either. I've confirmed the switching is ok to the amp end of the ribbon cable (I've had to replace one of those before in one of these).

If anyone's familiar with it, am I right in thinking that the inversion for the 2nd channel is happening by feeding to R151,44,31,46 into p5 of IC4b? And, has anyone else noticed that there seems to be some mixing up of channels between parts of the drawing and the unit? Channel 1 on the drawing feeds into the half with IC4 for input, yet it's the other half that warms up when I drive it.

I'm not usually inclined to go too far with these units, given their value and difficulty for working on, but this is largely working. Any tips would be gratefully received.

Thanks
Brian

[Mark Black]

Did you ever get the correct schematic? If not I could email you one (I'm assuming there are still problems with attachments here on the forum), along with the user guide. Just shoot me a PM.

Looks like the inversion happens by feeding IC6 pin-5, and the mode switches don't disconnect the chan-2 input circuitry. The manual does say to keep the Chan-2 input attenuator all the way down when running bridge mode.

[Jazz P Bass]

That is really odd that the A Clip Led illuminates with a load but without a signal.
That may be the clue.

[Brian Swatton]

Thanks Mark, I think I did get the correct schematic, Enzo pointed me at Elektrotanya. That was the first one I found that had circuits for both amps and the power input too. It's revision G (11/07/03), which appears to be what I have from the board markings.

I think my confusion about which side was which was due to the individual amp drawings not corresponding to the individual PCB's. For instance, IC4 & 6 are on different drawings, but are on the same PCB.

I have had ch 2 gain at zero during bridge tests btw.


Yes JazzP, I thought the clip light was a clue too. It didn't make sense. There was very little DC offset on the two outputs to cause much current to flow when the load was connected, from memory about -17mV and +22mV. Its a shame they're so difficult to make working measurements on. I'll have to attach wires and refit, but, I will have a go with a few around IC4 & 6 later on.

I have a big SMPS to rebuild today first, so can't get back to it until later.

Thanks for your input guys.

[Enzo]

To the uninitiated it can be confusing. You tend to expect the two channel boards are the same, one per channel. Then you find after beating your head agains the wall a while that some of the circuits for both channels are on one of the boards.

[Brian Swatton]

Yep, you got it Enzo. This is the first time I've been into any real depth on this particular unit. It wasn't until I got the full drawing I realised how it was arranged, so many thanks for the earlier info, It minimized the headbanging

[Brian Swatton]

Back on it this morning and I've found something that sure isn't going to help.

During cold meter checks around the bridge combining areas I found T39 is wrong (PSU dwg). Instead of an NPN A06 it has a PNP A56 fitted, this would feed volts out of it's base and into BR_CUT and BR_BAL lines when the switches are set. T39 has clearly been replaced, a little untidily, and there is other evidence of a previous repair.

I don't have any of these in stock, so I'll post result when its sorted.

This doesn't quite chime with the customer's report of knocking the unit over and things then not working, so I've asked for clarification.

[Jazz P Bass]

Pretty neat find.
Keep us posted.

[km6xz]

I would not take the customer's word for how the problem happened, unless you see evidence. A stronger chance of the problem was a bad attempt as repairing the amp, by themselves or by a friend, and they got in over their heads. The "fell over" excuse was probably to save face. A fall that would damage SMD circuits would have physical evidence like bent rack ears or broken pot or something. Those surface mount parts and their connection to the pc board are a lot more rugged than the mechanical chassis.

[Brian Swatton]

As suspected, all is now well with the correct device fitted for T39. I usually think of finding weird faults as ammo for the future, but I doubt I'll see that again. It's helped me get a bit more of a handle on the circuit I suppose.

km6xz: Yes the customer's word is only the first clue and not relied on. After speaking with him again about it, he now he thinks he probably hasn't tried to run this one in bridge before. It was bought S/H not long ago from a studio, who probably didn't use it in bridge either.

I have found SMD C's and R's to be a bit susceptible to cracks if boards are flexed by very much. They're certainly not as resilient to some kinds of stress as the old leaded types. It used to come up quite a bit when I was involved in production testing for SMPS that used a lot of SMD.