Behringer MX9000 - odd power issue

[Deekaph]

Hi guys, new here and just in search of a solution to my pickle.

Just bought a Behringer MX9000 as a blemish item off eBay (chip in the side paneling, no prob), arrived on Friday and everything was fine, very impressed. Saturday the problem started, then no problem on Sunday, and then Monday it was constant, as it was today. Here goes..

Regardless of what kind of signal is running into it (or any at all, it might be quiet and it'll do this) it will randomly and intermittently crap itself, you don't have to touch anything, it just seems to decide for itself:

1. Meter bridge blanks out.

2. Sound drops - you can still hear it faintly in the background, but it's distorted and about -60db (and the bus routing ceases to work)

3. The +17 light on the power supply field on the main board goes out. I THINK THIS IS THE KEY, based on my readings on the net these things seem prone to PSU problems, but I'll continue..

4. The MAIN SOLO light comes ON

5. The activity lights (green '-20' lights) come on for channels 21, 22, 23, 17, 12, 10, 7 and 4. NOTE: They flicker as though they're reading activity, but there's NOTHING running into those channels.

6. Turning the power on the external PSU off and then on resets it ... for a bit... again, intermittent, might be a couple of hours, might be 10 seconds before it craps itself again. ALSO, 'flicking' the power (turning it ALMOST all the way off and then releasing it) works, but again, for a 'bit' ...

At first I thought it might be a current problem (I have a *LOT* of juice running through this room), so I ran an extension cord from a seperate circuit up through the basement into the room, but like that rotten cat, the issue reoccurs .. I even repositioned the PSU, thinking that maybe magnetic fields from the 15" sub it was sitting on was causing problems, but no joy.

I even took some video of it acting up (pardon the coarse language, this is really frustrating for me! it's at :
http://youtube.com/watch?v=8T-DfI1urjY

I'm no electrician but I have dabbled a lot over the years and have a pretty good operating knowledge of electronics and can follow instructions. If anyone has dealt with this problem before and knows how to fix it, or sees the answer staring them in the face, or ANYTHING, I would be forever indebted to them and would wish a thousand happy good karma points at them. I'm desperate!

TIA

 

[Eva]

It seems that like a bad solder joint leaving part of the console without one supply rail at times... It may be also a supply regulator transistor or IC with one of its legs losing contact to the die internally (caused by shocks or thermal stress).

 

[Deekaph]

Hey thanks for the quick reply

If it was a bum solder or something, wouldn't the occurances usually happen depending on the juice running through the board - ie, sudden spikes or drops in input levels?

I am just doing some research and I can't help but wonder (maybe someone can correct me or acquiesce with me) if maybe it's a current problem with the juice coming into the PSU that's causing something in the PSU itself to crap out intermittently? I own an older home with older wiring and we're rural so I'm not convinced the grid is very strong out here. Would purchasing a power conditioner help, maybe?

Would anyone reccomend a power conditioner anyway? I'd never owned one before but I've seen them in a lot of studios and wonder if it might help.

 

[megajocke]

Bad joints usually react much to temperature change and vibrations. The current through them might matter, but usually they are worse at low currents. I'd expect it to make horrible crackling noises if there are bad joints in the power supply lines somewhere.

The metering bridges though will draw more current the more leds light up. If it is possible to disconnect them then try that. Now I wonder, is this with an external PSU? Does that have lights? And if so, does +17 go out on the PSU too?

The power supplies for Behringer desks are known to not be of the best quality. I think they use LM350:s running on the edge of what they can take. It might be overheating or just the regulators getting tired due to constant high temperature. If one regulator goes into foldback current limit it will not be able to come out of it due to the desk still pulling current through it when the -17 is still running. This seems likely as cycling the power makes it go again.

 

[Deekaph]

Actually, after a lot more research I've come across some other MX9000 owners who've encountered this issue, and you actually hit it there: Behringer used LM350's, which peak at what the unit is rated. Replacing with LM338's resolves the issue.

Thanks for your help!

 

[megajocke]

That sounds like a plan, they should drop right in. Make sure you get the same case style as there are both a plastic TO-220 (LM338T) and a metal TO-3 can version (LM338K). Good luck!

 

[Deekaph]

Ok, I posted yesterday but I guess it didn't post...

I've replaced the regs but the +17 is still dropping out after about 10 seconds, lose the meter bridge and audio, other various wierd things on the board. I pulled out the volt meter and took some readings, and the caps run steady but the +17V rail itself goes from 17 to -7 and the regulator jumps from 10 to 35.

Like I said, I'm not an expert but know enough to try things out and can follow directions. Heh. I throw myself at your mercy: what am I missing? Is this foldback? I've actually replaced the reg a couple times now, but I wonder if I got it too warm when installing it? I dunno. You guys seem to really know your stuff so I thank you in advance for whatever help you can offer. Thanks!

 

[megajocke]

Did you replace it with a LM338 or LM350? What is it that jumps from 10 to 35? Regulator tab? Which one? Does the -17 drop too when this happens? Another possibility is that the filter caps are bad. Measure the voltage over the filter capacitors, both in DC and AC modes on the multimeter. On the AC mode take measurments both with reversed leads and without. If the two AC readings differ by much then the value shown doesn't mean anything.

You could check if the regulators get hot when it freaks out. If it takes ten seconds then it's possible the regulator isn't mounted correctly to the heatsink. Make sure you mount it correctly on the heatsink, use either mica + thermal grease or a silpad.

Don't run it too long with -7 on the 17V rail, the electrolytics in the desk won't like that at all. They should have installed a diode to protect against that! But of course, that would have cost them like $0.1 extra... I'd install polarity protection diodes on the outputs. 1n540x or something like that.

Be careful, if the regulator is hot then it will probably be really hot! Watch out for the high voltage parts. Even the phantom power voltage might sting a bit.

 

[Deekaph]

Ok, here goes:

I don't run the desk more than a second or two once it poops its pants I shut it off right away.

Replaced the LM350's with LM338T's. The reading I took was with the voltmeter set to DCV500 and the reading from the regs was taken from the outside pins. It's from those pins that I got the readings from 10 to 35.

The regs get VERY hot, like burn to the touch within about 10 seconds. Which brings me to my other kerfuffle... I initially replaced them and mounted them to the heatsink with thermal gel, but silver particulate, which I think conducts, which is why when I powered it up after that it went down IMMEDIATELY. So I kicked myself in the butt and ordered new LM338's. Now, for heatpads I used a bit of rubber I shaved off a set of screwdrivers. I KNOW, I suck. It melted right away. Do you think that could be it, that it's not dispersing the heat into the heatsink? Or is it getting so hot because of whatever's going wrong?

Now, I don't really know which is the filter cap? I've uploaded a pic of the interior of the psu here: http://www.flickr.com/photos/[email protected]/2281131736/

Thanks so much for your help, I'm digging in my heels and not buying a new PSU because I think I can hammer this baby if I just stick to my guns and the only way you ever learn is by doing. heh. Thank you thank you I really feel like I'm close to nailing it once and for all.

 

[megajocke]

Hmm, that rubber is probably not good enough. You need all the heat transfer performance you can get. If you don't want to bother with thermal paste then you can try silpads - the grey and flexible stuff. The regulators have overtemperature shutdown so it's possible they are shutting down because of overheating. Do both regulators get this hot? Did both regulators have insulators from the beginning?

It's hard to see from your picture, but did you install the plastic bushing under the nut holding the regulators? It might be mounted under the heatsink too. If it's not installed you will have a short from +17 to 0V. It looks like the left regulator in the picture is for +17 and the right one is for -17.

The filter caps are those big ones in the picture. The group of four diodes above them is the rectifier. Measure from the anodes of the left diodes to the cathodes of the right diodes for each group. Be careful so you don't short something.

BTW, where did you buy your LM338s?

 

[Deekaph]

Well I still have thermal paste, but am I not supposed to keep the back of the reg, which is metal, from coming into direct contact with the heatsink? That would be the point of having the plastic bushing wouldn't it? (and yes it's installed there, but not on the bottom).

The regulators did originally have the silpads under them, although they were in really rough shape and I chucked them.

In the picture the one on the far right is the +17V -- although, I should point out that on the green board inside the PSU it's identified as +18V. Hmm. And it's the +17 that's having the problem, although when I replaced the -17V the first time and used the thermal paste it cut out too. Now it's just the +17, and it gets HOT, whereas the -17 gets pretty warm, but not as hot.

I'll have to take the reading from the filter caps tonite when my wife's around to turn it on and off for me.

I got the LM338's directly from National Semiconductor... "samples" for free, just got to pay shipping ($15, and they came pretty fast for coming via UPS)

 

[Eva]

You should get some mica insulators for TO-220, like these:

[IMGDEAD]http://www.questwww.com/EXCESS%20INVENTORY/MICA%20T0-220%20Insulators.JPG[/IMGDEAD]

You can also cut bigger ones to the proper shape. They are intended to be used with that white thermal compound, with a very fine layer on each side.

Don't forget the nylon shoulder washers on the bolts or screws to prevent them from shorting the metal tab of the regulators to the heatsink:

[IMGDEAD]http://www.superiortattoo.com/images/products/5852.jpg[/IMGDEAD]

Once the regulators are mounted properly, finding the fault should be easier...

 

[Deekaph]

OK well I managed to get some non-conducting thermal paste today, but no mica insulators. The reg gets REALLY hot really fast.. 10 seconds or so and it'll burn your finger. I got some readings...

Reading from the filter cap is 25DCV, 59ACV.

From the rectifiers, I got 25V and it goes up to about 27 when it crashes.

I keep saying that I'm no expert, but I'm sure that the reg overheating is the problem. It's overheating even with the paste.. it's heating up too fast for it to dissipate into the heatsink, within seconds it's almost combustible. SO, I guess the real question then is WHY oh WHY is it heating up so fast?

Suggestions?

TIA, as usual, I'm lost without you .

 

[megajocke]

Are you still using those shaved off rubber pieces? I doubt they have good enough heat conductivity... Thermal paste is just for providing better contact between the surfaces, if the insulator insulates the heat too you are out of luck.

 

[Deekaph]

Heh nope removed the rubber, now it's just the regs mounted as close as possible to the heatsink with paste inbetween.

 

[megajocke]

What does "as close as possible" mean? You are not going to get near enough heat transfer without screwing them down with the proper pressure. And I think that might be a bit hard to do without inulators if you don't want a short circuit.

 

[Deekaph]

Yeah thats the thing, if you crank it right down then it's gonna short. *SIGH* The problem is there's no real electronics shops here... there's a repair place, but they don't seem to have ANYTHING, so whatever I need I've got to order out of the US (and I'm in northern alberta)... time to place an order I guess. Is there anything else you think I should order while I'm at it (filter caps, etc) because it means another couple weeks, another $20 shipping for $0.50 in parts.

 

[Deekaph]

*SIGH*

And so, the battle continues...

Got myself some of insulators (mylar) and used with the white paste. Now I get 28 seconds before the +17 or -17 goes down. Twice as much time, but certainly no use for recording studios.

Exact same issue as before: it gets HOT after a few seconds and by 28 it dies. The rest of the regs in the PSU don't even really get warm, but like I said before, these ones are getting sizzle to the touch after a few seconds, not enough time for the heat sink to dissipate the heat.

I'm about ready to buy a new PSU because it's been almost two months now with this going on, my gear gathering dust, and I've got some time on my hands coming up and I'd really like to do some recording.

So, we're back to that old question: why are the regs getting SO hot SO fast? I don't know enough to say what it is, and so I once again throw myself at the mercy of my more skilled friends at diyaudio. Help!

Thanks as usual ....

- James

 

[megajocke]

In that case I'd guess there is a problem with the desk drawing excessive current if the mounting of the regulators is good. Is it the +17 and -17 getting hot? There isn't really much that can go wrong in the power supply itself: Partially shorted capacitors (should get hot if this is the case), bad isolation between regulators and heatsink, bad overvoltage protection zeners if there are any or shorted cable are the only ones I can think of.

I'd try measuring the current drawn by the desk from the +17 and -17V.

 

[Deekaph]

How would I measure the current being drawn by the desk? Open it up and measure the current where the power comes in?