Switch mode supply again

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Chrisfromthepast

Well-known member
Joined
Mar 5, 2018
Messages
265
I started getting really annoyed by this problem when we started powering up gear that sat for 6 months during the pandemic.
No power… check the powersupply. Oh its a switched mode module thats crapped out. Sweet!
Lets see if the manufacturer stocks it.
Oh wow, ok, newark has stock though. Grab another because I can’t justify the bench time for a throwaway module.
I would love to learn to troubleshoot these little enigmas, like is possible with little knowledge about a linear supply. Instead, Im just building a pile. Is this course of action really more efficient than a linear supply in the lifecycle of the product?

ADF8C814-F494-4F16-A9D5-03D9556A57A2.jpeg
 
I’ve been through more than a dozen, in devices from seemingly good companies: Shure, Blackmagic, this piece is made by LOUD though…. Maybe this gripe is meant for the brewery, because I don’t necessarily have the time or overt need to start troubleshooting a module that is so available and affordable. Thats why the switched mode supplies pile up as E waste. My own guilt in this cycle is not overlooked.
Is the relative replaceability of these supplies an asset for equipment designers and end users, or are we overlooking the waste to a dangerous degree?
Is it always cheaper to buy another module later than it is to buy a transformer now?
I guess the market has decided, and Im just a stick in the mud.
 
Where is that supply on newark? Is it the big one (part number 0007796-03) or the little open frame one?

I have always thought Loud Technologies made good stuff. Their VLZ{3,4} mixers are pretty well designed. But they are NOT designed to be serviced and they are priced accordingly. So maybe the problem is your expectations.

If something is subject to a lot of heat and moisture, like you're gigging it outside and it's baking in the sun, then yeah, there are a lot of parts in an SMPS that can fail. Generally it's not worth servicing only because you can buy a new one for far less than what someone would charge for time to fix it. It just doesn't make financial sense. Don't make a pile. Just take it to the e-waste center and move on.

But of course this being DIY, and knowing that eventually you may not be able to find an off-the-shelf replacement, I would be happy to help you diagnose. Sounds like fun to me actually. There's probably a "90% method" for fixing those things where you just run through a check list and voila, it's fixed (90% of the time).
 
This particular unit is a poor example.
This one requires the fix described here:
https://design.eaw.com/knowledge/persistent-rebooting-on-a-ux8800Its not a cheapo mixer, its an older system’s processor, so its not one of the items thats only a couple of years old; third one I’ve done this to.
The latest meanwell was for a video switcher that lived indoors at a church. Its already been replaced and installed, but wonder why it would be internal to the device rather than on an external supply.

-EDIT I mean well, when I generically call this XP POWER VCT60US12 a meanwell.
 
Last edited:
What is the supply on newark that you mentioned?

I am interested because that combination of voltages is potentially very useful for audio.
 
I would not even call a corroded connector a power supply failure, I would call that a cable failure.
And that little XP Power 12V supply appears to only have 3 electrolytic caps, one of which is wedged in between two heat sinks. Seems a likely candidate for failure, and it would only take about 10 minutes to just change them out and see if that fixed the problem for one of those supplies.
 
Ive a couple of pairs of 20 year old active monitors , old style iron cored PSU still works ,
Ive seen pairs of very very expensive modern studio monitors with SMPS supplies die in less than 5 years usage ,
By the time you look for spares ,the product is replaced and the company involved , if your lucky might offer you a concession on their latest built to fail model ..... its the merry-go-round business model , like a vortex where its out with the old ,in with the new in an ever decreasing time span .
Sure its difficult and costly to make a quiet transformer based supply and if the product only has a lifecycle of a few years ,why bother .

I got a couple of Digidesign 02's a while back that were destined for the scrap heap ,
they have quite a good SMPS , 5V, +/-16V and phantom rails and plenty of xlr/jack sockets well worth taking the time to desolder .

Theres undocumented wars going all over the planet to feed this greedy business model , we all have blood on our hands Chris .
 
I would not even call a corroded connector a power supply failure, I would call that a cable failure.
Can we dig in to this?
These connectors have not ever been unplugged, the spring contacts are nice and springy. I see a little bit of oxidation, but its really modest by Florida standards.
Whats going on?
 

Attachments

  • 2433AF74-21BC-4D61-93F7-21F9D68E9F17.jpeg
    2433AF74-21BC-4D61-93F7-21F9D68E9F17.jpeg
    1.7 MB · Views: 0
I'm convinced. That's the "90% method" I speculated about!

I think it should be possible to create a high frequency ESR tester that can test electrolytics without removing them from the board. Just put a 100 ohm resistor in series with the output of an op amp but with feedback taken from the OUTSIDE end of the resistor. Then make a probe tip with some shielded cable with two thick bits of bus wire on the end that can be bent a little to match the pitch of the cap. Drive the probe with output and 0V of aforementioned driver amp circuit. Drive amp circuit with 100kHz [1]. Touch the probe to the cap long enough to, using an oscilloscope, capture the voltage drop across both the 100 ohm resistor and across the capacitor. Then compare. For example, if the voltages are the same, the ESR of the capacitor at that frequency is 100 ohms. If the voltage drop across the cap is about half as much as the 100 ohms, ESR is more like 50 ohms. And so on.

[1] It might be important that the the driver amp circuit be floating which is a little tricky if your oscilloscope and function generator is not.
 
Whats going on?
Nothing obvious on those two contacts. How does the mating connector look? Might be worth checking the other contacts and seeing if there are any that have obvious problems, then clean and reinstall the contacts and see if the problem still exists.
 
The connections were all beautiful.
When I first encountered this problem and asked EAW for help, I went down that rabbit hole of cleaning, brushing the contacts, but the processor kept sporadically booting.
Solder fixed it.
But shouldn’t the node at the power supply be very low impedance, and therefore the main constraint to worry about is the current capacity of the wire?
What am I missing?
 
the main constraint to worry about is the current capacity of the wire?

And the contact resistance of the connector. Perhaps a connector was chosen not appropriate for the use, either it required movement to keep the oxidation broken up, or required higher voltage or current through the contact to keep it conductive (see "wet" vs. "dry" contacts), or the two sides were not compatible (in the past you used to have to make a choice of tin plated or gold plated fingers on your memory modules to match the motherboard connector plating).

You would probably have to find the exact part numbers on the cable and the board side and go look up specs as well as any errata or bulletins regarding manufacturing flaws in some batches to figure out exactly what is going on. Apparently EAW just decided to give up and get rid of the problem point rather than trying to correct it.
 
Ahh, you’re saying two dissimilar metals and/or an unfortunate geometry aren’t exactly ohmic, so the resistance is much higher at low voltage or low current or whatever unfortunate circumstances create this non linear behavior at the junction.
And this makes the power supply hiccup because its not seeing a load that is linear?
Im down with that, or the UC is choosing to start reboot routine because the power rails aren’t predictable due to the same effects. Thanks!
 
Thanks for this reading material! The capacitor choices are so important, I feel for the bean counters. Cheap caps are so tempting

In short: To repair a faulty SMPS simply replace all the secondary side electrolytic capacitors PLUS the small electrolytic capacitor(s) on the primary side close to heatsink/resistor that gets hot.

This works 90% of the time?
I’ve replaced a fair amount of varistors on these dudes too. Should I worry about those units?
 
Thanks for this reading material! The capacitor choices are so important, I feel for the bean counters. Cheap caps are so tempting

In short: To repair a faulty SMPS simply replace all the secondary side electrolytic capacitors PLUS the small electrolytic capacitor(s) on the primary side close to heatsink/resistor that gets hot.

This works 90% of the time?
I’ve replaced a fair amount of varistors on these dudes too. Should I worry about those units?

I can only speak from my personal experience. I have fixed a lot of dead SMPS, probably over a hundred.

The most common failure is a shorted FET. This nearly always results in taking out many of the associated components: i.e. current sense resistor, gate drive components and the PWM chip. It smart to at least check the snubber too.

Next, is the auxiliary supply. The culprit is usually the small filter cap. An ESR meter usually finds this quickly

Then we have a shorted diode on the secondary followed by shorted bridge in primary.

I have seen more that a few shorted inductors in the secondary side of forward converters. These I rewind by hand.

Failed caps on the secondary are more subtle and tend lead to increased ripple rather than outright failure. I check their ESR.

Only once have I seen a bad transformer.
 
Last edited:
Back
Top