Fixing a power supply for Chameleon Labs 7720

[leigh]

I read the tacked-on blue wire incorrectly - although it looked like it connected to the fuse, it actually doesn't - a PCB trace on the board has been (I assume, intentionally) severed, so appearances were deceiving. See photo.

Instead, the blue wire is part of the scheme that powers the lamp for the VU meter. The two leads to the lamp connect to the + and - sides of the power supply, (after two stages of RC filtering, but before the final regulators), where the DC voltage is about 23 volts per side (so it's 46 V total being put across the lamp circuit). The lamp is in series with a series pair of 640 ohm, high-wattage resistors, and when the circuit is fired up, there's a 5v drop across the lamp.

Anyways, that doesn't point to a solution of any kind, because even if the lamp had shorted (which it hasn't), those two resistors would still limit the current through the lamp circuit to about 36mA.

Which really only leaves the first two capacitors (the ceramic and the 1st electrolytic, which is 1000uF/35V) as the possible shorting points (with those 51 ohm resistors in series with the power rails after that point) - so maybe with that 50 ohm power resistor I'm using instead of the fuse, it's limiting current just enough not to have one of those borderline components fail.

Either that, or, it's an incorrectly labeled fuse, as PRR has mentioned. For what it's worth, putting the fuse in series with the 50 ohm power resistor does NOT blow the fuse.

Thank you all for your invaluable input. This seems like a whole lot of head-scratching for a lousy cheap power board, but I'm learning something, so that makes it worth it for me... glad you are all willing to participate in that process too!

 

[leigh]

I'm using these 1A fuses from Radio Shack... is there a chance I should be using slow-blow fuses instead? I asked the manufacturer what the replacement fuse should be (since it's not spec'd anywhere in the compressor's manual), and they said 1A but did not specify slow-blow.

Is the power transformer a torroidal type ?

Oh, and to answer this: I'm fairly sure it's not. The transformer is external, it lives inside a sealed "line lump" as part of the power cord, so I can't see it... but my assumption would be they would have used the cheapest option, because a) it's a budget unit, b) it's external anyways, so who cares about stray AC fields since they're not living next to audio circuits.

 

[leigh]

I'm starting to think the problem might be the fast-acting (rather than slow-blow) fuse.

After stepping down the series power resistor (that I was substituting for the fuse) to smaller and smaller values, and seeing no ill effects, I finally tried bridging the fuse clips with a current meter, with no power resistor at all.

The quiescent current settled down to about 200 mA. However, during power up, it looks like the current swings as high as 2 amps. There are a pair of 1000µF power caps in that first RC filter stage, one hanging directly off each side of the rectifier diode pair. When the unit is powered up, they draw in current limited only by their ESR, as far as I can tell.

Perhaps the external AC transformer had some ramp-up function that is no longer working as it should. Or maybe this unit should just use a slow-blow fuse, which, as I mentioned, the manufacturer did not spec, but I have emailed them now for clarification.

 

[JohnRoberts]

Modern Low ESR caps could exacerbate in-rush surge current. They actually make surge current limiters for use in products with massive reservoir caps.

When I build my old DIY amp back in the '70s I had lots of capacitance so used an old two position power switch (from an old tube PS). When first powered up, the current was limited by a series resistor. In the second switch position I shorted across the current limit resistor.

JR

 

[leigh]

I just heard back from the manufacturer - they say it should be a fast-acting, 1A fuse.

Since I'm seeing a 2A inrush current on power up, that's going to keep popping fuses.

I guess I'll try replacing the first stage of RC filter caps, and see if something's amiss there. Although I can't see how that would be - they're not shorting out, and I've only heard of electrolytic caps decreasing in value as they age, not increasing (i.e. if they had increased in capacitance, they would be causing a greater inrush current). But maybe there's some other phenonmenon I'm not understanding.

The only other thing I can think of is that something in the external AC-to-AC line lump transformer unit was supposed to be ramping up the juice slowly at first, and has now failed. Seems far-fetched, though.

 

[leigh]

24VAC ought to peak-out at 36V both sides. Which is suspicious close to the 35V rating on those regulators.

And, now I realize, suspiciously close to the 35V rating on the first electrolytic caps, the "reservoir" caps that come directly after the diodes.

After removing the first stage of caps, the fuse is no longer popping. Of course I need to replace them now to get smooth DC power back, and those replacements are on order (replacements will be rated for 50V). Then we'll see if there was some kind of cap breakdown over time that was causing too much inrush current...

(For what it's worth, the quiescent current before was 200mA, and with those reservoir caps removed, that current has dropped to 120mA.)

 

[mjrippe]

Forgot to mention, check the voltage rating of the caps. 

Yes, quoting myself here.  As I mentioned, Chameleon tried to pinch pennies by using caps barely (under)rated for the voltages they see.

 

[leigh]

Forgot to mention, check the voltage rating of the caps. 

Yes, quoting myself here.  As I mentioned, Chameleon tried to pinch pennies by using caps barely (under)rated for the voltages they see.

Yes, sorry I missed that. When I first read that, I figured, a 24Vsupply with 35V caps must be OK. But I had missed the peak voltage issue, which would seem to make it a much narrower margin (34V peak into a 35V cap leaves little room for momentary overages).

 

[leigh]

OK, a resolution of sorts...

I replaced the first stage of RC filter caps (the 1000µF ones) with a pair of 50V rated ones. The inrush current is now peaking at about 1.5 amps on the amp meter (down from about a 2 amp peak previously).

My last 1A fuse from Radio Shack still popped with this inrush current. Good riddance.

I then tried a 1A fast acting fuse, sourced from Mouser. It does not pop! (Although, you can see the fuse wire inside flex mightily on powerup.)

Thanks again all for your help!

 

[leigh]

Hmmm... maybe not so fast...

The fuse isn't blowing, but I've got +/- 34 VDC right after the diodes, instead of the 24 VDC I had before. How the hell did my DC voltage jump so much??

AC input still measures at about 24 V.

The pair of resistors that limit current to the VU meter's lamp are running HOT. It used to be 46VDC running that lamp circuit, now it's 60 some volts.

Arg.

Here's a quick approximate schematic of the power board, up until the voltage regulators (fuse and power switch also not shown).

 

[PRR]

> I've got +/- 34 VDC

That's what I'd expect. 24V AC Sine is 24V*1.414= 33.9V peak. Deduct a part-volt for diode drop, but add a few percent for transformer regulation.

 

[leigh]

That's what I'd expect. 24V AC Sine is 24V*1.414= 33.9V peak. Deduct a part-volt for diode drop, but add a few percent for transformer regulation.

Agreed, although I didn't record a measurement that high before changing the reservoir caps out. But I had started with a 100 ohm resistor in place of the fuse, and stepped it down from there to 50 ohms, then 25 ohms, before removing it entirely - so maybe I just forgot to record a reading once that resistor was gone.

Regardless, the power supply is not blowing fuses any longer.

My main concern now is that the two 680 ohm resistors (R111 and R112) in series with the VU meter lamp are running so hot. It's those two resistors and the lamp draped across the DC rails (pre voltage regulators), -32V to +32V. So, 64 volts total, and the lamp is only dropping 8 volts. Which, of course, leaves 56 volts across the pair of resistors (which make 1360 ohms in series). Ohm's law says that lets 40mA through, and a current measurement with the Fluke confirms that.

56 volts * 40 mA = 2.24 watt

So each resistor in that pair is dissipating over a watt, continuously. They *are* big-ish resistors, apparently chosen to handle some heat, but yet - doesn't that seem like quite a waste, to build a power supply that burns 2 watts just on those resistors?

 

[leigh]

Alright, I've been out of town, but back now and want to get this box closed back up and in service again. I am still petitioning Chameleon to release the schematics/tech notes (for just the power board, at least), but they have been unresponsive.

Now that it's not blowing fuses any more, my chief concern is the 2+ watts being burned off by power resistors in the lamp circuit. What I'm seeing: the lamp circuit is 64 VDC, the lamp itself drops that by 8V, leaving 56V across a pair of power resistors. Current draw for the lamp circuit is 40mA, so that's over a watt each for those two resistors, i.e. they're getting very hot.

If this had been a sturdier circuit from the start, I might not be as concerned. But I question the design choice of putting 35V reservoir caps in a spot that can easily see over 35 volts – and so likewise I question the heat being thrown off by these resistors, and wonder if I should either swap out a different lamp (to eat more of those 64 volts) or up the wattage rating on those power resistors that are currently getting so hot.

Thanks for your insight.