Identify this: cap leak residue? Heat damage? ISA 428 A/D Card Issue

[zmix]

Large electrolytic caps are often cemented in place using a brown glue, I've never seen anything like that ooze from one.....

[A27Hull]

If it was a cap leak, would the residue be hardened, and would it be conductive?

I did a quick test with a continuity meter to check, but it came up negative...

But if it was a serious cap problem in the PSU area, the whole unit would not power up...right?

[zmix]

No the the first 2 and I wouldn't make that assumption on the third point.

Is this a switchmode or linear supply?

[A27Hull]

After further investigation...laying out on the bench, I noticed a few things.

see picts above.

I removed the case cover and powered the unit up... ...

I plugged a beta 57 into an input to see if the mic channel would work.

I noticed that the Peak meters were lit, so some of the unit is obviously powered, but that none of the front panel buttons were lit. Strange...

Sure enough, I carefuly jiggle the connection seen at top of thread, (that looks burnt-brownish on two leads, and Wallah! all buttons light up.

Mic passes signal, Peak meter reacts... good sign.

Now I'll try re-installing the ADC card...

[SureShotStudio]

the caps that i have seen that had leaked and failed, looked like they had powdery crystally stuf around them or they were sticky and those elco's were at least 40 years old. But they never had hard brown stuff around them. The stuff around those caps on the pictures looks like brown glue.... and considering the age of your unit i don't think the elco's have leaked, but i could be wrong ..

also the burn mark on the power connector looks very suspicious. But it could be nothing.

The connector might have come loose and gave some sparks that could explain the brown mark. If it short circuited than the power supply should have failed. But that hasn't happenend according to your last post.

So i think your probably good to go.

good luck!

[Wavebourn]

Here is what really happens when they leak (I had this problem with my Nissan radio in hot day, the cap is near a tiny heatsink that supposed to dissipate power needed to drive 185W of speakers)

http://wavebourn.net/images/2005/nissanradio1.jpg

http://wavebourn.net/images/2005/nissanradio3.jpg

http://wavebourn.net/images/2005/nissanradio4.jpg

You may check contact between a wire and connector pin, it seems to be bad.

[A27Hull]

Originally Posted by Wavebourn

You may check contact between a wire and connector pin, it seems to be bad.

Yeah, I re-connected the ADC card after noticing the (connecton...e) between the PSU to PCB connector.

Slightly moving the connector previously had restored power to the whole unit, and when ADC was reinstalled, (with the connector connection previously restored,) it powered up as normal.

A very big thank you to you guys for your input.

I'm waiting to hear back from ff, and the probability of having to replace a board or something else serious is now lower than previously thought!

Any guess as to how common those plastic power supply connector are?

[David Kulka]

The failed Molex connector is a common problem in power supply circuits. After a few years the connection becomes slightly corroded and resistive. Resistive connection means some slight voltage drop between the male and female pins. Slight voltage drop + current = heat. The heat bakes the pins and increases the resistance, producing even more heat. Eventually the connection fails entirely. It definitely needs to be replaced and the male pins should be thoroughly cleaned. These are common parts, should be easy to find.

It's a little hard to tell from the photos, but it looks to me like at least some of those capacitors failed and began to ooze. Not 100% sure though. You might remove 1 or 2 and inspect them closely to get a better idea.

[A27Hull]

Originally Posted by David Kulka

It's a little hard to tell from the photos, but it looks to me like at least some of those capacitors failed and began to ooze. Not 100% sure though. You might remove 1 or 2 and inspect them closely to get a better idea.

thanks!

the connector I can handle, as for the caps...

I'll try to take some better photos tomorrow.

[A27Hull]

Here's an update to my dilemma:

I have taken the box to several people around town to get opinions on the power supply caps.

So far, no one can say for sure that the caps are leaking...My only conclusion to the cap issue is that it is remotely possible that the board itself has been burned due to heat, the caps have leaked underneath the glue, or that the glue itself is that particular color.

ugh.

----
The power connector can be replaced, however local electronics part shops do not cary the square-peg design. I guess its a british thing.

ugh.

The only guess I and others can venture is that a voltage surge triggered the burn-out in the Molex connector, which was already weakened due to heat stress.

The two burnt pins on the connector are recessed compared to the others, so perhaps it is possible that an arc occured between the pins and connector, in a situation where the connector and the pins were slightly separated.
---

Focusrite Technical Support is apparently located in England. This makes sending back to the factory out of the question. They have been helpful to the point of responding to my emails, so I have no druthers with them. They have one guy in the states that services focusrite gear, whose contact info they sent.

I do have questions about the engineers/designers, though...

Why should the unit run so hot? -high voltage, high current, many resistors?

Why use electrolytics in a PSU area, all rated for 85*, except for 2 rated 105*?

My guess is that its a cost/marketability ratio.

The ADC card has numerous 85* rated caps, several of which, by the way, are bulging at the top, a few with pin sized residue as well.

If through their rigorous testing, they knew the unit itself would rise in temp above the allowability of the components, why let it pass into production?

I have really grown to like the ISA range. I had even been considering a 220 or 430... Now the thought of having similar problems is killing that.

----

On a side note, one of my electronics instructors told me to devise a fan device for the unit, one that would draw out heat, and perhaps lower the operating temp. The theorized fan would probably be a 12volt computer fan. If I decide to attempt this, I'll try and document it on this forum.

Thanks to all who participated in this thread.

-Andrew Hull

[Wavebourn]

As an engineer - designer I can tell you that the socket may run hot because wires are not soldered to contacts; instead contacts are clamped around wires so if in this place contact between the wire and the contact is bad it is getting hot. Plastic is melting, contact between male and female parts of sockets get loose, and more heat is produced...

[A27Hull]

Thanks Waveborn!

[Monsterlab]

yeah - I usually cut those molex females out and solder the power connections directly to the headers - except where that operation makes later disassembly difficult - on every piece of gear I open up. No danger of high resistance and better long term reliability. Its only a matter of time before that nylon or whatever it is becomes wicked brittle and cracks from the inevitable hi resistance and associated heat - potential for short circuits/blown rectifiers/WORSE down the road. That crud on the board is glue. I bet if you heat it up gently with a heat gun, it will melt. If you aren't getting unusually high hum out of it, I wouldn't worry.

[A27Hull]

So, it would be safe to solder direct to the header pins as long as I make sure they can't short between/ie. insulate them?

I just spent about 3 hours hunting up the appropriate molex connector.

I swear, blind parts hunting is hard.

[Led]

YOu would be better off the desolder the header from the pcb and solder the wires into the corresponding holes. Keep in mind though, if you do this you will have to desolder the wires at a later point of you need to seperate the pcb's. Have you tried digikey or mouser for the molex connectors?

[A27Hull]

Well, I want to refrain from removing the main pcb board. It would require unhooking the whole thing.

Yes, both Mouser and Digikey offer the molex connector...or so it would seem, not knowing the actual part number and having to guess, there are several similar options.

I narrowed it down to one or two, they cost about $0.45 a piece.

The other issue I now have is replacing 5 small 220uF, 10V, 85* Radial SK Jamicon caps on the ADC card.

Jamicon only distributes through a co. called IDC.

Will an equivalent manufacturer suffice?

What brands do you suggest?

[A27Hull]

whether design or board related.

So I go to an electronics shop in the metro to get a molex connector and possibly some caps...

We get into a discussion about my situation, and I tell them that (aside from the burnt connector) the unit regularly hot during use...that the chassis gets so hot that you wouldn;t want to hold bare flesh to it very long.

They responded by insisting that something is definitely WRONG with my Focusrite.

They add that a solid state device, let alone mic preamp should never run so hot.

I reply that many people have complained about the same thing.

They say, perhaps it is design flaw, that it's possible that the unit PSU is producing too much current, and so forth...

They add that simply replacing the connector would be like put a new wheel on a car missing 3 others...

I mentioned also the caps going bad on the ADC card, too.

Bottom line is that they think what I describe are symptoms of a bigger problem with the unit. Namely, a component/board problem or a design problem.

What say you?

[neilio]

Originally Posted by A27Hull

whether design or board related.

So I go to an electronics shop in the metro to get a molex connector and possibly some caps...

We get into a discussion about my situation, and I tell them that (aside from the burnt connector) the unit regularly hot during use...that the chassis gets so hot that you wouldn;t want to hold bare flesh to it very long.

They responded by insisting that something is definitely WRONG with my Focusrite.

They add that a solid state device, let alone mic preamp should never run so hot.

I reply that many people have complained about the same thing.

They say, perhaps it is design flaw, that it's possible that the unit PSU is producing too much current, and so forth...

They add that simply replacing the connector would be like put a new wheel on a car missing 3 others...

I mentioned also the caps going bad on the ADC card, too.

Bottom line is that they think what I describe are symptoms of a bigger problem with the unit. Namely, a component/board problem or a design problem.

What say you?

hello, i had the unique experience to work on one of the focusrite consoles for a long time and a lot of albums, and to be fair that circuit is a toaster....in our studio control room during the winter(hollywood,ca mind you) the heater was never run,on a cold day coming in from the cold(marley song in my head) i could literally warm my hands by holding them over the board. same mic pre as your isa...they are current hogs,the trannies mostly.

as far as psu caps being rated to low,thermally you may be on to something there and if you can find some good nichicons(or whatever good quality) that do have a higher temp rating that may be avery good project for you to do,also dont worry about replacement caps if they have HIGHER voltage ratings as long they are biased properly.

correct me wise slutz if i am wrong on this.

in other words the isa series does run warm....

good luck.

[A27Hull]

the caps that are in there are all 85*C Standard jamicon caps

85*C =185*F

Is that right? does it have to run that hot to degrade capacitors?

I'm out o luck with it. Everything folks have told me points to a bigger problem than I can fix. Molex connector replacement is fine, but problems with the PSU power handling...board failure elsewhere....

Now I'd like to have it fixed, get rid of it, and be done with Focusrite. I really like the product, but I really really hate the thought of design flaw, and the thought of a reoccuring problem.

[Dan]

I've got several FF pieces, and yes the 428 runs hot. Especially with the ADC card in it. Does yours have the extra little heat sink on the back?

[A27Hull]

Yeah, the one that screws into the back below the ADC...

As an update, I sent the unit to Thermal Relief in LV NV.

I should have it back in a few weeks.

[JustDucky]

My ISA428 failed to turn on right before a recording session last night with the client waiting to sing. I found a workaround but I LOVE my ISA 428 and use it as my go-to preamp for vocals.

Anyway, I pulled it out and cracked it open this morning and it was exactly the same problem as others had noted above... the gray wires from the transformer to the power supply get too hot. In my case the Molex was brittle and the spring-contacts for those two wires just pulled out. I have the feeling I was not the first one in there because one of the wires was marked with a Sharpie.

So, I just took a pair of dykes and cut the brittle part of the connector off. I stripped (22 AWG) and tinned the wires, and carefully soldered them onto the terminals. The remaining part of the connector slips on right next to them.

There are some pitfalls here. You should be an experienced solderer to do this job (heat the parts, not the solder, wipe between parts, and don't drip solder onto the board!) I guess if you were unlucky when you cut the connector you could shatter the whole thing. It might be a good idea to write down the wire order before doing this.

Anyway, the ONLY authorized FocusRite repair in the US is Thermal Relief Design in Las Vegas. Their average repair cost on this unit is $435 USD (published on their website) and then don't forget your expedited shipping and extra-special packing so add on another $150 USD ($75 each way) so you can keep your client from finishing their emergency vocal part somewhere else.

Thanks, guys, I took pictures of the repair and will try to put them up here later... I am sick as a dog today and really didn't need this extra work.

This is a great community. Love it!

[Th3_uN1Qu3]

Originally Posted by A27Hull

the caps that are in there are all 85*C Standard jamicon caps

85*C =185*F

Is that right? does it have to run that hot to degrade capacitors?

Jamicon = cheap. 85C = cheap. Cheap cheap cheap. I would have expected better.

The temperature rating of capacitors is a maximum rating - depending on the electrolyte composition, they last between 1000 and 10 000 hours at their rated temperature, with 1000 to 4000 hrs being most common. When you do the math, that isn't much.

Every 10 degree drop in temperature doubles the lifetime of the capacitor, so let's say for example that we have a 2000 hour cap rated at 85C, running in a device with internal temperatures of 45C. That would be 2000*2^4, which gives 32 000 hours service life. At 65C the same cap would only last 8000 hrs.

With the current trend of making equipment as compact as possible, internal operating temperatures have risen, now couple this with low quality capacitors and connectors... and you've already seen the result.

[Radardoug]

A preamp that runs hot is badly designed. Period. Failure of Molex connectors is very very common. Often these connectors are used between the power transformer and the board, and the peak currents on these pins are very high. Designers who ignore this SHOULD NOT be designing anything.
Are you listening Focusrite? Anyone using 85 degree caps in quality gear these days is a cheapster. DON'T BUY FROM THEM!

[digitrax]

Originally Posted by JustDucky

Anyway, the ONLY authorized FocusRite repair in the US is Thermal Relief Design in Las Vegas. Their average repair cost on this unit is $435 USD (published on their website) and then don't forget your expedited shipping and extra-special packing so add on another $150 USD ($75 each way) so you can keep your client from finishing their emergency vocal part somewhere else....

For anyone interested: I will remove the Molex connectors and solder the PS wires directly to the PCB as well as replace the cheap Jamicon caps with Nichicon HE (10,000 hour caps vs 2,000) which also have a much lower ESR than the Jamicons. I will also add the rear heatsink (Focusrite P/N 605-K03172091) if not in place already. The price to do this is $200 plus shipping - average turnaround time is 2-3 days (plus shipping) and the work is guaranteed. I have done dozens of these units (including my own) so far and the improvement in reliability is substantial. Many have noted an improvement in sound as well.

In the interest of full disclosure, I will tell you that I am NOT an authorized Focusrite service center (so I can't perform work under manufacturer's warranty) but I AM a degreed Electrical Engineer and audio designer.