Okay, I've gotten an email from another BBS user pleading for help.

They have just had their Mk2a upgraded to 64MB RAM (PCA board), knob & button LEDs, internal FAN with controller, and a full version of the digital out board.

A few hours later the onboard PSU appears to have suffered meltdown, and let out too much magic smoke.

Pure speculation on my part, as I have not seen the unit (nor done any of the aforementioned work on it). I suppose there might be a wrong value resistor on one of the various add-ons, drawing way too much current or something.

Or maybe.. maybe.. all of this stuff has finally exceeded the current draw capacity of one of the onboard voltage regulation circuits of the Mk2a.

Patrick, Hugo: Any idea as to the "spare" current carrying capacity of a stock unit?

This failed one also has TWO drives installed, so I'm guessing there wasn't much juice left over to begin with..

Cheers

I've been running 64MB and lights and two drives for a year with no problems. Paul is running 48MB with lights, two drives and the digital out board on all his empegs.

I suspect that something else is at play here. Handle spring?

This unit also has a fan and fan controller installed. That could be just enough to tip a delicate balance.. Fans have high start-up current (not as high as drives, but perhaps enough).

Cheers

Could something have shorted? I would think that if it were drawing too much power, that you would either lose / could not spin up a drive or fan first.

Mark, a thought. Perhaps the drives would have SMART data on them that would show a power failure? They would be easy enough to transplant and read the results from.

Thanks Mark. I stopped in for lunch and noticed this. I'm still trying to sort this out.

Just for clarification, the unit does not have the digital out board.

After the upgrades, the player was "buggy" when first powered up. For instance, pressing the bottom button worked fine, except when holding it down. The visuals wouldn't change. But once I switched visuals via the menu, the long bottom button press DID work. Also, it had trouble waking up out of a standby, but turning the volume knob did bring up the volume display. Still asleep, but the volume "slider" was visible. My timezone wasn't set the same for some reaons either. A few reboots later and it was okay. I thought it was a quirk of the mem install. The memory seemed to pass during the boot process (still running 2.00 however). I noticed that pressing the buttons changed the brightness of the LEDs but I have no clear buttons to really know how much. Is that normal? Total running time about 1 hour.

The next day, I let the player run for 4+ hours. Thank God I was in the room when at some point, the "low battery" icon came on and the player did one of those quick reboots. Then the display faded in and out, it then was in standby mode as far as the screen looked, but I could still hear music. Then the audio glitched for a second or two and the next thing I know, I see smoke coming from the dock! I yanked the power away and let the smoke clear.

Maybe at 4 hours the fan finally came on?

As Mark pointed out, a short would not have taken 4 hours to show up right?

Valid point. There is no such thing as a time delayed short. Unless some solder melted with a short and dripped somewhere bad... But not as likely.

I think you will need to carefully examine your main board for any burn marks, other than near the PSU. Hmm... A low battery indicator while docked??? I know that the power feed while docked is through the loom, but I thought that recent hijacks recognized the dock configuration as AC power. Not that this helps, but it might be something to look towards.

Yep, the Low Battery icon is what made me stay in the room to see what was going on. Good thing too - I wasn't in the room and barely had the volume up the whole four hours! I thought the worst that could happen was a lock up.

That does certainly sound like the fan could be a culprit. Is there any chance that it could have been obstructed? A fan that is physically stalled would draw a large current.

Possibly. It's a fight fit in there. But the player was in a dock so I have no idea if it ever spun (or tried).

Sounds to me like your fan was jammed and it drew stall current (= a lot of amps). However, the PSU is fused and should have simply shut down until reset.

For smoke to come out implies something other than the empeg is burning.

Yeah, but the fuses just go immediately. At the point where the fan kicks in, then it should have tripped. Something else is going on here.

I think I would clip off the fan leads and try again. Be ready to yank the cable. Don't do it in the dock.

Mmm.. I wonder which way the installed fan tried to push the air.. ?

Pushing air OUT through the top of the player would be good, in this case, since that's the same direction as the dock fan moves the air.

Cheers

So it could have just been the fan that went up in smoke (or the controller)? That's very good (possible) news.

The power supply is good for about 2.5-3A at least continuously on both the 5 and 3.3v rails. If it's overloaded it should simply shut down. Certainly a short will make it die instantly with no damage.

The whole thing is also fused at a level which should prevent damage with resettable polyfuses. It's actually fairly difficult to damage the empeg psu, it's deliberately overdesigned to be very robust. If smoke came out after several hours operation, it implies something has slowly overheated. This seems to me to be unlikely to be the PSU, as under a constant load it would come up to it's maximum temperature quite fast, no more than 5 minutes or so.

The memory board presents a fair extra load to the supply, but well within usable limits. The button LEDs don't take more than a few 10s of milliamps at worst, and I can't imagine that the digital out board takes more than perhaps 100mA at the most.

The fan is a possibility, as if it jams (as Rob S suggested), it can end up drawing a horrendous amount of power. I have had one of those little brushless jobs get so hot it melted the solder holding it's innards in place, so them immediately became outards.

I would suggest that the first thing to do is measure for shorts between the dc in line and case ground, then between the 5v, 3.3v, and internal dc in points, to ground. With the fan both connected and disconnected.

pca

The whole problem was in apparent short between one of the fan module components and a corner of hard drive cradle. The most interesting is that it happened a couple of hours after the player was inserted in a dock. Initially at the installation I didn't have access to hard drives with cradle to test the proper clearance and adjust the components appropriately.
I had to replace bbq'd fan components and now everything is back at full swing.

You are right about one thing - it takes a lot to "smoke" empeg.

The intermittent initial misbehavior of empeg after memory board installation was probably due to the fact that I booted it many times with serial connection without ever attaching hard drives.

The actual measured current draw of fan controller with fan rotating was 0.05 A

I am running fan, lights, stacked memory and digital board in my main unit for a year now without a hitch.