switch mode PSU's for audio.. pro's & cons, please discuss..

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
gswan said:
Nothing wrong with using a suitably designed switcher in it, provided you know what you are doing.

As I said the console sounds great withe only 2 channels in it.  But, I am not sure this psu was designed fior this consxole.  It looks like it's made with a bunch of generic switchers from ebay .............
 
Are the converter modules powered from AC?  Is there a part number or spec of any kind on the modules?  There has been no indication in any post of specs apart from output voltage, and some very subjective loading information.
 
Your proof is the down-loaded console.  Why not spend an hour measuring current requirements for the different voltages for the centre section and modules and do the math for the whole console.  It should prove that the logic and audio SM supplies from Velvet are insufficiently sized.

Mike
 
I was just talking with a friend who said he thought that maybe the electrolytic caps on the output should be paralleled with some small ceramics.  The reason being that any HF parasitics could cause a large current draw that causes the supply to start shutting down, when it shuts down enough the HF parasitics stop, & the power supply switches on again & the whole cycle repeats causing the motor boating.

 
Rob Flinn said:
I was just talking with a friend who said he thought that maybe the electrolytic caps on the output should be paralleled with some small ceramics.  The reason being that any HF parasitics could cause a large current draw that causes the supply to start shutting down, when it shuts down enough the HF parasitics stop, & the power supply switches on again & the whole cycle repeats causing the motor boating.

That sounds backwards...  Adding ceramics will lower the impedance (at HF) and could increase the current surge at start up, while I suspect the small ceramics to be inconsequential.

The evidence that the PS starts up with a small number of modules populated suggests that the caps alone are not the cause of the instability.

What needs to be inspected is what changes when you plug up more modules. This increases current draw but we ASSume the PS is sized adequately. This is OK for me to ASSume, but you need to check it out, either by finding a part number and data sheet, or testing with a dummy load.

It is possible, but seems unlikely, that extra capacitance on each module rail, incrementally pushes the total over some threshold making it unstable? It seems inspection of the modules could quantify this amount of added capacitance, and removing the large caps at the PS could experimentally determine if they are at fault.

I like to joke that consoles are the most complicated simple circuits... here is another example where even the power supply can require extra effort.

JR
 
I would strongly suggest to measure the current on the independent rails to see where you end up, kind of important to know anyway, no?

from your numbers each module would use up to 1.5A per rail (bipolar) and then run into current limiting.... maybe some less due to the current used for the main..... any real world numbers out there? I would guesstimate that 30W of bipolar is quite possible. do all modules use the same?

so that brings me back to the earlier suggestion; run a power usage test on all modules, starting with the master alone... testing one after the other...

my 2 cents.......

- michael
 
..generic switch mode PSU modules, each of which had an additional large cap strappped across its o/p..
What is your definition of large cap ? (+sum of rail bypassing caps per channel)
Looking at datasheets of some Traco switchers that might be in your used range, they say max.capacitive load 800uF for a 20W or max.6000uF for a 40W module. Other switchers might have similar limits. Just an idea.
 
Thanks for your responses.

Unfortunately it's rather difficult to check any of these things out because the console is 80 miles away.

Audiomixer I hear you about testing the individual current draw for each rail. but since I'm no where near the console, not easy to do.

The 24v supply which is for relays is 8.5A out so it's rather alarming that the relays are chattering.  I'm basing this on my MCI which has a hell of a lot more 24v relays & only has a 3A linear supply for them.

I think it may be that I need to completely dismantle the switching psu to try & find some data on the rest of them.  The supplies used are these & and look like the 100W range. However looking on ebay they have similar looking ones that are 50W & lower

http://www.stontronics.co.uk/Products/Switch-Mode-Power-Supplies/Internal-Enclosed-Power-Supplies.aspx
 
Don't rule out some obscure interaction, from say a floating ground. I would expect +/- rails to load up symmetrically, but something is going on..

it will be something obvious, after you find it... :)

JR
 
Rob Flinn said:
The 24v supply which is for relays is 8.5A out so it's rather alarming that the relays are chattering.  I'm basing this on my MCI which has a hell of a lot more 24v relays & only has a 3A linear supply for them.

My guess is that they're chattering because the logic that ultimately drives the relay coils is going in and out of reset due to the supplies going up then collapsing.

-a
 
One other thing could be that some smps do not like to be ganged (+/- supply)... maybe the tracking gets all over the place when current gets to something meaningfull. Regarding remote supervision and power metering: your client could put a simple powermeter in front of the supply, like the one you can buy to meter your household power equipment....

- michael 
 
Why don't you try the obvious things first.

Remove the big caps and see what happens.  The SMPSU probably has foldback current limiting which will give the 'motorboating'.  Having to charge the caps AND more than 2 modules at switch on might be the trigger.

Then ...

Measure the current taken by each & every module.  Then check that the +/- 15 supplies can deliver all that + a bit (maybe lots more) & still regulate.  That will tell you if the SMPSUs might be OK for the job.  There's a (small) chance that it might have worked.  If not, at least you know what new SMPSU to buy as a replacement.  That would be less risky and less work than designing & building a new PSU.
 
Hi Rob,

this is the Velvet facebook, gotta say, they look really pro:

https://www.facebook.com/pages/Velvet-Audio/186591948096401

contact info: http://www.musiktonstudio.de/


This post maybe of interest, again a Velvet supply spotted:

http://www.gearslutz.com/board/so-much-gear-so-little-time/368651-siemens-c4-console-has-landed-2.html

Andy
 
Interestingly that is the supply my friend has, and I think theyre talking about the same console, but it was never tested with this desk because I had to make a loom for it.

I just got qa mail fro,m Velvet audio, and Jurgen suggests that I connect all zero v rails apart from minus 15v.  I will try it when I next visit;
 
Back looking at the desk.  I tried different earthing config, which did nothing.  When I removed the 10000u caps plugging the psus. I was able to load the console much more, but it seems there are a few flak channels that will still cause the oscillation when inserted.  Waiting for diagrams to investigate further.
 
On further investigation this does still appear to be a load dependant problem.  This is because all the channels I had marked as bad apart from 1 worked when other channels were removed.  Also sometimes when I switch on the psus motor boats, but if I switch it off and on quickly it stabilized.
 
Rob Flinn said:
On further investigation this does still appear to be a load dependant problem.  This is because all the channels I had marked as bad apart from 1 worked when other channels were removed.  Also sometimes when I switch on the psus motor boats, but if I switch it off and on quickly it stabilized.

That symptom sounds like perhaps some fold back current limiting interacting with capacitive load. The first try partially charges the caps, then it can start the rest of the way.

Does the power supply drive a resistive load equivalent to the full channel load? I appreciate this is not easy to test.

A small R in series with the supplies could current limit the C load at start up as another test... the voltage drop would not be acceptable in normal use, but perhaps useful to diagnose what is going on.

JR
 
Question about the two switchers for the +/- rails.How are they wired? Is one of them a -15v supply, or are they two +15v supplies with one of the +15v rails tied to ground on the other?
Using two independent supplies also will not give good "tracking" like a linear dual tracking supply would.
Easy enough to build a +/- 15v supply using a dual tracking regulator with some series pass elements to increas current output.
 
HI John The 15v psu's are rated at 4A which is what I've been told the original linear supply is capable of.  Apart from that I can't tell you if it will drive 4A.  It's rather difficult to check these things out on site.

I'm not sure but there may be some trimmability on the 15v psu's which perhaps could make up for dropper resistors.

wkrbee

Its made from 2 positive psu's.  In practice it actually tracks quite well.  I hear what you & everyone else is saying about building a linear supply.  I've done it before, but it's not my console etc etc.
 

Latest posts

Back
Top