MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Maxing out your PSU: How far can you go?
MUFF WIGGLER Forum Index -> Eurorack Modules Goto page 1, 2  Next [all]
Author Maxing out your PSU: How far can you go?
R.U.Nuts
Hey there, this is my current rack:

As you can see, it consumes 949 mA on the +12V rail. My PSU is a Doepfer DIY kit with 1200mA on both 12 and -12V rails. It works rock solid. However at the moment I'm thinking about swapping out the Echophon for a 4MS DLD. The DLD consumes 110mA more on the +12V rail than the Echophon which means my configuration will draw 1059 mA from the positive rail, which is pretty close to the max 1200 mA of my PSU. Do you think this could cause me issues?
luketeaford
Just summarizing advice I've seen before until someone with more experience weighs in: I think you will be fine as long as you are below the rating, but when you power the case on, it needs to draw a lot of power all at once, so it is nice to not absolutely max it out.
DSC
If in doubt, there is no doubt, change it out!
moremagic
luketeaford wrote:
Just summarizing advice I've seen before until someone with more experience weighs in: I think you will be fine as long as you are below the rating, but when you power the case on, it needs to draw a lot of power all at once, so it is nice to not absolutely max it out.
i have never read any advice saying pushing the mfr's limit is o k, everything ive read says 80% or less *per rail* to leave enough juice for power on/booting and such

op, the DLD should be o k but if you get a high noise floor or anythiing funny it could be from pushing the power supply so hard. if it works and you dont need your echophon no more, pm me, could always use another
R.U.Nuts
Well, 80% of 1200mA is 960mA. My current setup is right below that threshold. With the DLD it would be above. Since that 80% advice is probably more or less a rule of thump I guess I should just try if it works with the DLD...
And what about that new WMD Softstart for the possible extra load on power-up?
SunSpots
R.U.Nuts wrote:
Well, 80% of 1200mA is 960mA. My current setup is right below that threshold. With the DLD it would be above. Since that 80% advice is probably more or less a rule of thump I guess I should just try if it works with the DLD...
And what about that new WMD Softstart for the possible extra load on power-up?


Here's advice I actually did for myself. Buy an intellijel PSU. One of their big ones. Make sure it fits in your case. 3k on +/- 12v rail and 1k on the 5v. then never think about it again unless you're like me and loading up on digital modules like a robot.

It's worth the money... and it's easy to do.
R.U.Nuts
SunSpots wrote:
R.U.Nuts wrote:
Well, 80% of 1200mA is 960mA. My current setup is right below that threshold. With the DLD it would be above. Since that 80% advice is probably more or less a rule of thump I guess I should just try if it works with the DLD...
And what about that new WMD Softstart for the possible extra load on power-up?


Here's advice I actually did for myself. Buy an intellijel PSU. One of their big ones. Make sure it fits in your case. 3k on +/- 12v rail and 1k on the 5v. then never think about it again unless you're like me and loading up on digital modules like a robot.

It's worth the money... and it's easy to do.


Naah, I won't do that. The Doepfer PSU works great and I have no plans to expand this rack. If the DLD should stress the PSU too much I will just return it and stick with my Echophon and get a looper pedal that can be synced to my modular because audio looping is what I'm after with the DLD.
Mungo
R.U.Nuts wrote:
my configuration will draw 1059 mA from the positive rail, which is pretty close to the max 1200 mA of my PSU. Do you think this could cause me issues?
It might or it might not, more information is needed. On paper it seems to be worringly close to the specification but thats not always a problem. From the replies so far we can pull together more details.

luketeaford wrote:
I think you will be fine as long as you are below the rating, but when you power the case on, it needs to draw a lot of power all at once, so it is nice to not absolutely max it out.
This is usually only a problem for switching power supplies, they don't always have a lot of extra capacity or are configured to shut down when overloaded to avoid damaging themselves. Most (but not all) linear supplies on the other hand will provide much more current than specified for short periods and only heat up a little more.

more magic wrote:
i have never read any advice saying pushing the mfr's limit is o k, everything ive read says 80% or less *per rail* to leave enough juice for power on/booting and such

op, the DLD should be o k but if you get a high noise floor or anythiing funny it could be from pushing the power supply so hard.
80% is a very "finger in the air" amount, it can vary a lot between different supplies (even from the same manufacturer). There have been several manufacturers just put the theoretical maximum current the components could deliver under ideal circumstances as their specification, when it is clear that inside a warm case there is no possible way that could occur for more than a few seconds. Then there are other manufacturers who have tested their supplies at the rated load for long periods in realistic situations to verify that the rating is accurate.

This occurs even with manufacturers of the bricks/transformers, even with the same rating printed on the casing a cost reduced supply might produce more heat or less stability at the rated load while a more conservative manufacturers product could continue working to 50% or 100% more than the official rating. The only way to know is testing. For instance the brick on the Doepfer DIY case is rated for 2500mA AC, which would be less than needed for the full capacity of the supply:
http://www.hammondmfg.com/pdf/5c007.pdf
But that specific brick may be able to comfortably provide the extra current and heat from imbalanced power rails, do we trust Doepfer? Certainly, and will they provide support and assistance if the product doesn't work as claimed? Absolutely.

The other half of the current draw is the specifications of the modules again with examples going in either direction of real modules drawing more or less than their specification (or not having the information entered on modulargrid). So again its a trust thing with the specific manufacturers. Ideally the advertised power consumption would be worst case maximum so that the user can size their supply correctly but some figures are typical (some times noted with extra power needed at warm up) or just guessed from the design and not tested like with power supplies. And just as with power supplies there is competitive advantage in reporting "better" numbers so its tempting to specific a power supply right at its limit for the biggest number possible, and a module consumption as low as possible. Together these make it hard for the user to size their power system correctly as measuring the current draw per module is somewhat difficult and time consuming.

For most supplies and bricks as they approach their limit the noise and/or ripple will increase. This is usually measurable with a multimeter by checking the AC voltage on each rail which should be much less than 0.5V (many cheaper multimeters can't measure small AC voltages accurately). But this is not always true for the switching supplies which will often simply cut out suddenly when the current draw is too high.
R.U.Nuts
Mungo wrote:
R.U.Nuts wrote:
my configuration will draw 1059 mA from the positive rail, which is pretty close to the max 1200 mA of my PSU. Do you think this could cause me issues?
It might or it might not, more information is needed. On paper it seems to be worringly close to the specification but thats not always a problem. From the replies so far we can pull together more details.

luketeaford wrote:
I think you will be fine as long as you are below the rating, but when you power the case on, it needs to draw a lot of power all at once, so it is nice to not absolutely max it out.
This is usually only a problem for switching power supplies, they don't always have a lot of extra capacity or are configured to shut down when overloaded to avoid damaging themselves. Most (but not all) linear supplies on the other hand will provide much more current than specified for short periods and only heat up a little more.

more magic wrote:
i have never read any advice saying pushing the mfr's limit is o k, everything ive read says 80% or less *per rail* to leave enough juice for power on/booting and such

op, the DLD should be o k but if you get a high noise floor or anythiing funny it could be from pushing the power supply so hard.
80% is a very "finger in the air" amount, it can vary a lot between different supplies (even from the same manufacturer). There have been several manufacturers just put the theoretical maximum current the components could deliver under ideal circumstances as their specification, when it is clear that inside a warm case there is no possible way that could occur for more than a few seconds. Then there are other manufacturers who have tested their supplies at the rated load for long periods in realistic situations to verify that the rating is accurate.

This occurs even with manufacturers of the bricks/transformers, even with the same rating printed on the casing a cost reduced supply might produce more heat or less stability at the rated load while a more conservative manufacturers product could continue working to 50% or 100% more than the official rating. The only way to know is testing. For instance the brick on the Doepfer DIY case is rated for 2500mA AC, which would be less than needed for the full capacity of the supply:
http://www.hammondmfg.com/pdf/5c007.pdf
But that specific brick may be able to comfortably provide the extra current and heat from imbalanced power rails, do we trust Doepfer? Certainly, and will they provide support and assistance if the product doesn't work as claimed? Absolutely.

The other half of the current draw is the specifications of the modules again with examples going in either direction of real modules drawing more or less than their specification (or not having the information entered on modulargrid). So again its a trust thing with the specific manufacturers. Ideally the advertised power consumption would be worst case maximum so that the user can size their supply correctly but some figures are typical (some times noted with extra power needed at warm up) or just guessed from the design and not tested like with power supplies. And just as with power supplies there is competitive advantage in reporting "better" numbers so its tempting to specific a power supply right at its limit for the biggest number possible, and a module consumption as low as possible. Together these make it hard for the user to size their power system correctly as measuring the current draw per module is somewhat difficult and time consuming.

For most supplies and bricks as they approach their limit the noise and/or ripple will increase. This is usually measurable with a multimeter by checking the AC voltage on each rail which should be much less than 0.5V (many cheaper multimeters can't measure small AC voltages accurately). But this is not always true for the switching supplies which will often simply cut out suddenly when the current draw is too high.


Hey Mungo! Thanks for the in-depth reply. So given that the power specs of PSUs and modules are possibly not really accurate the conclusion is that I just have to try if the Doepfer PSU can handle the extra load of the DLD... Thank god for German return policies (hoping that I won't need to return the DLD).
Short Circuit
Hi Folks, absolute newbie here and didn't know where to start posts but at the moment the Doepfer a-100 diy supply is what I need to sort out so thought I'd start out here. Hope that's OK.
I've just recently purchased a second hand rig 84HP 9U with around 22 various modules all supposedly powered by the a-100 supply.
When I got the rig there were about 6 modules plugged in and it worked but when I plugged in another module the fuse in the a-100 blew. The fuse was a 400ma and there were 3 spare fuses with the rig. I plugged in another fuse and nothing worked. I found my multimeter and after an hour found my multimeter wasn't working properly and that the fuse holder was rather loose hence the original owner using sticky tape! The original owner told me the fuses were correct ratings. Finally I fixed the fuse holder but nothing worked. Anyway after a lot more stuffing around, finding another old multimeter, internet research, I found 2 of the spare fuses were dud.
The third fuse worked so I decided to plug in the Pittsburgh Box with the mixer and voila, the Box worked and finally some fun. So then I try plugging Pittsburgh Filter but the fuse blew again. I had some old 500ma fuses laying around so started trying them but they all blew immediately, even after unplugging everything including the buses. I'm now waiting for the mailman to deliver some more fuses.
What I'm mostly concerned about is the a-100 can apparently handle 1200ma so why would I be advised a 400ma fuse is correct, can I not use a 1A fuse? My old 500ma fuses I'm sure were fast blow and probably no good for the initial current burst for the Pitsburgh Box but I thought they would be at least adequate for the a-100 with no load, what gives? The Pitsburgh Box is rated 125ma but does anyone know what the start-up current would draw? Is it safe to use a 1A fuse in the a-100? and, if I want to have all the modules plugged in should I be looking at upgrading the power supply and if so to what?
This forum is a wealth of information and I've been searching for some time but easily lost.
Thanks in advance.
PS: I like cats!, gemstones, electronics, building, renovating, reading, gardening and of course noise making. Used to be pro drummer until injury got the better and built the ETI4600 in the late 70's.
JohnLRice
Welcome to Muff's, Short Circuit! w00t

I've never had that power supply but looking at the docs you should be able to use up to a 2.5A fuse? hmmm..... Someone please correct me if I'm wrong, here's the link:
http://www.doepfer.de/a100_man/A100_DIY_kits.pdf

When the fuse blew the first time are you sure you plugged in the module the correct way? It's a mistake probably a majority of modular users have done at least once as it is quite easy to do. And if you still have module plugged in the wrong way and keep changing fuses they will keep blowing. Before you try any more fuses you should triple check all power connections, or better yet unplug ALL modules and then put in the proper sized fuse and use your meter to check for correct voltages and stability. Then once that is done, plug one module in at a time and recheck the voltage rails each time.
Short Circuit
Thanks for your quick reply JLR.
Checking for the correct way to plug the module was something I did with great care and I feel positive they were plugged in correctly. The thing is the supply still blows the fuse with nothing plugged in, although the fuses are fast blow, rather than the recommended slow blow. Still, I wouldn't have thought they'd blow with no load attached.
The link you provided is where I'v been getting the info but I've been a bit confusing as they say the supply will deliver half of the transformer, or brick, current rating, so with 2.5A transformer the a-100 will deliver 1200ma. Jumping from a 400ma to 2.5A fuse is a big jump so I just want to be sure that it's correct.
Cheers
JohnLRice
hhmm, blowing fuses with no modules connected sounds like something is shorted on the distribution board or in the power supply (or some other power supply failure) Check for something shorting the rails above and below the buss boards like a loose washer, screw or nut. Also check to make sure the buss boards are wired correctly to the power supply.

EDIT: and check the headers on the distribution board to make sure some of the pins aren't bent and shorting the rails together.

Actually which of the two setups below do you have? I was assuming the first one but if you have .4 amp fuses maybe it's the second one?

Pighood
Dieter D personally told me 1200 means 1200, not 80% of 1200.
Short Circuit
JLR, I have the first version #1 and I see now clearly from that diagram, which I hadn't seen before, that the correct fuse is 2.5A. I've got 1A on order so I'll order some bigger ones.

As I mentioned before, it blows fuses without any load, including the busboards. I have everything disconnected with only the 15vAC connected and it blew the fuse again.

I've checked the board for mechanical faults including dry joints and it all appears OK. There is solder resin on the regulator connectors and it could be that they were replaced at some time. I tested the diodes in situ and they appear to be OK. I can't get any clear reading on the regulators while they are in place.

I can only guess at this stage is the regulators are faulty. I haven't found a circuit diagram so am not sure what they are without unsoldering. That's probably my next step, unsolder the regulators and get some new ones.
Not sure whether I should just rebuild the whole thing while its apart.
Mungo
Short Circuit wrote:
JLR, I have the first version #1 and I see now clearly from that diagram, which I hadn't seen before, that the correct fuse is 2.5A. I've got 1A on order so I'll order some bigger ones.

As I mentioned before, it blows fuses without any load, including the busboards. I have everything disconnected with only the 15vAC connected and it blew the fuse again.

I've checked the board for mechanical faults including dry joints and it all appears OK. There is solder resin on the regulator connectors and it could be that they were replaced at some time. I tested the diodes in situ and they appear to be OK. I can't get any clear reading on the regulators while they are in place.

I can only guess at this stage is the regulators are faulty. I haven't found a circuit diagram so am not sure what they are without unsoldering. That's probably my next step, unsolder the regulators and get some new ones.
Not sure whether I should just rebuild the whole thing while its apart.
Until you have the correct (slow blow: http://www.doepfer.de/faq/a100_faq.htm#Fuse) fuses don't go any further, there is a large startup transient/glitch that can occur and blow smaller fuses which is entirely normal for linear power supplies.
papawasarodeo
Recently I thought I had a safe margin at about 50 left on the minus rail in one of my cases. But then I had a problem with modules freezing up. Turned out that I needed a larger margin, sometimes modules can spike momentarily in power consumption. And for that you need margins.

There’s no way to know beforehand what those margins need to be since every case is it’s own eco-system.
pugix
Beyond the question of a power supply coming up and every module running, remains a question of how do you know it's still not overloaded? An overloaded supply can start to produce more ripple and noise. So how would you know? I'm asking the question. I don't know the answer. However I remember reading about one test that involves putting a frequency meter on every VCO and looking for small wavering in the frequency, which you can also probably hear if you listen carefully. A VCO that is stable when the power supply is not overloaded can start to exhibit this problem. But this probably applies only to analog VCOs.

So, how could we tell if a supply is being overloaded? What sorts of test might we do?

Thanks,
mskala
Three thoughts:

If you can't run the power supply at 100% of its spec, then the manufacturer lied to you about what the spec really was.

Trying to run the power supply at more than 100% of its spec is probably not a good idea. There will be a safety margin built into the number, but if the power supply could safely run at a higher current, the manufacturer would have said so.

Modulargrid's numbers are for estimating only. You don't know how much current your system is really using unless you have measured it; and that is one reason to buy a power supply with more capability than the Modulargrid estimate for your system. Not because you need an overpowered power supply, but because you don't know the real power consumption of your system when you haven't measured.
Sync
Just because a PSU is rated for a certain power level doesn't mean that under the conditions you are operating it, it can do that. In particular, if you enclose it to the point of impeding necessary airflow, it can overheat and shutdown or fail. Depending on what heat sinks and fans or air vents the manufacturer chose to use, components may or may not live up to their ratings, and even with good heat sinks if you've blocked the airflow by putting it inside a tight DIY case with no additional fans, you could have a problem. And that's not the maufacturer's fault, necessarily...
mskala
Mungo wrote:
The other half of the current draw is the specifications of the modules again with examples going in either direction of real modules drawing more or less than their specification (or not having the information entered on modulargrid).


I do "maximum current in normal use." Most of my modules are capable of delivering a fair bit of extra current at the outputs if you short them out or pile on the "passive" modules, and that current has to come from somewhere; but if you built your system to supply true worst-case current for all modules all the time, you'd be wasting a lot of capacity. If someone else specifies modules according to "maximum possible current under any circumstances" and a third manufacturer specifies "typical average," it's not at all clear how to add them up and get a meaningful result.

It's similar to the issues surrounding "voltage range" specifications on inputs and outputs - are we talking about what's expected in normal use, what will produce some effect but not normal intended operation (such as audio distortion), or what should not be exceeded because it will cause permanent damage?
mskala
Sync wrote:
Just because a PSU is rated for a certain power level doesn't mean that under the conditions you are operating it, it can do that. In particular, if you enclose it to the point of impeding necessary airflow, it can overheat and shutdown or fail. [...] And that's not the maufacturer's fault, necessarily...


It is if they didn't specify the supply's cooling requirements.
JohnLRice
pugix wrote:
how could we tell if a supply is being overloaded? What sorts of test might we do?
Hi Richard,

Do a Google search on "testing power supply under load" or similar. I'd try to explain but this is a newish topic for me as well and I'm not knowledgeable enough to try to explain it myself! oops Basically you just need an accurate voltage meter and the power supply and you compare the voltage output with no load and max load, I think.

I recently purchased a dual channel electronic load to make testing easier, probably an unnecessary and excessive purchase but I have over 15 different modular power supplies here now, I wanted to check one that I had been accidentally overloading for a while, and . . . I just really like test gear and tools! Mr. Green This is the one I got, pretty low and as far as dual DC loads go but I couldn't justify a better one:
https://www.amazon.com/Happybuy-Electronic-Overcurrent-Protection-Cont roller/dp/B06Y281P5N
Graham Hinton
pugix wrote:
Beyond the question of a power supply coming up and every module running, remains a question of how do you know it's still not overloaded? An overloaded supply can start to produce more ripple and noise. So how would you know? I'm asking the question. I don't know the answer.


When a PSU drops out of regulation like this the ripple will usually be rectified mains frequency, 100Hz or 120Hz, and its effect should be very noticeable.


Quote:

So, how could we tell if a supply is being overloaded? What sorts of test might we do?


Measuring current in situ is difficult. A multimeter will do it, but it requires disconnecting the power lines and inserting it in each line in turn and it may effect the operation. All the techniques for measuring current involve either inserting a resistance or measuring the magnetic field . You can get current "clamps" which do the latter, but good DC ones are expensive.

However there is a way that I used on the FullPower2 PSUs to monitor the currents which only introduces a very small resistance (3 milliohms). This is now incorporated in FullPower3s before the regulators so it doesn't effect the outputs even that tiny amount and I have a limited number of the old pcbs available. PM me if interested.

There are two outputs: a set of LEDs that indicate 50%, 90%, 100% and 110% of the rated current, and an analogue voltage than may be connected to a meter or an oscilloscope to find the exact current or its dynamic waveform.

Sync wrote:
Just because a PSU is rated for a certain power level doesn't mean that under the conditions you are operating it, it can do that.


Actually that is what rating is supposed to mean. If you look in the small print of a properly rated PSU's data sheet there will be information regarding how to derate it under defined conditions, like lack of ventilation. It is just that for the Eurorack market in particular there are some badly designed and incorrectly rated PSUs around.
Short Circuit
Mungo wrote:
....there is a large startup transient/glitch that can occur and blow smaller fuses which is entirely normal for linear power supplies.

Thanks, I was wondering about start-up transients and I guess the slow-blow fuse is designed for such situations.
But I still can't understand how the original owner was running the rig, with at at least 4 modules, on a measly 400ma fuse. It even worked for me until I plugged a sixth module in.
Anyway I've pulled out the lm317 & 337 and ordered new ones so I may as well wait and put the new ones in.
Does anyone know of a schematic for the a-100 diy#1 supply.
MUFF WIGGLER Forum Index -> Eurorack Modules Goto page 1, 2  Next [all]
Page 1 of 2
Powered by phpBB © phpBB Group