Switch-mode PSU for power amp?

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

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

ask

Well-known member
Joined
Feb 16, 2006
Messages
123
Location
Copenhagen, Denmark
Would anyone be able to recommend a switch-mode PSU suitable for a 2x400W power amp? I'm planning on building an amp from Hypex modules, but since they haven't made an SMPS yet, I'm looking for advice. I tried googling, but didn't get much. Please help. :wink:
 
dont use a switcher, they are too fragile and can get noisy.

they work electrolytic caps to death, so as soon as the electrlte dries out, you start popping fuses.

on the other hand, a large diameter copper wire around a nice piece of M6 grain oriented silicon will last 10 times longer, and give you halkf the headaches, but i do not you to have a better studio than me, so uy a switcher and be happy.
 
dont use a switcher, they are too fragile and can get noisy.

they work electrolytic caps to death, so as soon as the electrlte dries out, you start popping fuses.

on the other hand, a large diameter copper wire around a nice piece of M6 grain oriented silicon will last 10 times longer, and give you halkf the headaches, but i do not you to have a better studio than me, so buy a switcher and be happy.
 
dont use a switcher, they are too fragile and can get noisy.

they work electrolytic caps to death, so as soon as the electrlte dries out, you start popping fuses.

on the other hand, a large diameter copper wire around a nice piece of M6 grain oriented silicon will last 10 times longer, and give you halkf the headaches, but i do not want you to make the right decision and have a better studio than me, so buy a switcher and be happy.

whats wrong with a little attitude once in a while?

whats so stupid about peace, love and =understanding,

peacer, love and understanding companies dom not furhter the science.
 
Listen to CJ.

SMPS for a power amp can work well but there are a lot of gotcha's. For a one-off, don't do it. Unless you have to, for example, to run them off a car battery or if the law says you need it. Since a speaker is generally a reactive load, you can get rail pumping, where energy from one rail is transferred to the other rail. That is not a good thing. Boom. Properly designed power amps with SMPS inputs have a lot of thought put into things like that. One way around that is to have the normal switcher generate the + and - rails, and have another bidirectional SMPS be capable of transferring power from one rail to the other.

Thanks to certain regulatory bodies, we will be seeing more SMPS PSU's for power amps because you need to add a PFC and THD cancelling arrangement upstream these days. Try fixing a Carver, multiply by three and you have the complexity... :guinness: :guinness: :guinness: :guinness: :green:
 
Thanks guys. Perhaps I didn't make it clear that I'm not looking to design a SMPS, I'm looking for a drop-in solution that is tried and tested. I'd like to "go green" ie save 40% heat waste from an old school psu, but if it's not feasible then so be it. LAB Gruppen and other serious amp-makers obviously have good solutions, I hoped something similar would be accessible to DIY'ers, maybe this is not the case?
:roll:
 
I've got a pile of about ten broken SMPS (acronym for: Sh..tty Mother F'n Piece if Sh.....it) power supplies around here, and not a single broken linear one. I've tried fixing every one of them with no luck, except for the one that SVART was kind enough to fix for me. In my state of mind right now, I wouldn't mess with it unless you want to be messing with it every five years or so.

I've been replacing them with good old reliable linears when they go bad because, to me, it's not worth the down time. In case you didn't catch my thread on the Digi 96i/o, I got worked over pretty hard by one last week...
 
[quote author="CJ"]dont use a switcher, they are too fragile and can get noisy.

they work electrolytic caps to death, so as soon as the electrlte dries out, you start popping fuses.

on the other hand, a large diameter copper wire around a nice piece of M6 grain oriented silicon will last 10 times longer, and give you halkf the headaches, but i do not you to have a better studio than me, so uy a switcher and be happy.[/quote]

I reckon a great big YMMV is need here.
My "day job" is serviceperson for most of the big power amps brought into New Zealand - Crown, QSC, C-Audio etc etc.
All have flagship amps which use SMPSUs
All have "second teir" (read cheaper) amps which use conventional PSUs.
In over a decade of servicing these things, I have yet to see a significant number of capacitor failures. The opposite is true with small SMPSUs used in mixers, rack gear etc - typically a 3842 driving a fet in a badly ventilated box. Here cap failures are rampant (and a fine source of income :). It all depends on the integrity of the design, attention paid to cap selection, heat dissipation etc.
Go ahead, use lots of copper and iron - your chairopractor will love you for it :).
M
 
SMPS reliability depends mostly on the margins designed-in. You can make SMPS really reliable, or you can make them marginal, all by component choice.

Thinking about it, actually, it could probably work with a class B output, but if you have a switching power amplifier and a switching power supply, that's when you get rail pumping. In a class B amp, this actually isn't a problem. So you just need to find two power supplies that will deliver the required amount of voltage and current. You will probably need to add additional output capacitance. Remember that an unregulated iron PSU is sized essentially for average current, where a SMPS needs to be sized for peak current, so an SMPS needs to be oversized by the amount of reactive load you are expecting.

I still agree with CJ for a one-off, unless you have a specific reason for doing otherwise.
 
OK guys, you've (almost) talked me out of buying a SMPS. I don't wanna be messin with tings i dont understand, at least not at the possible cost of an amp. I guess it's bring on the iron! :?
 
Lets put it this way, you have Ray Charles sitting at the Steinway, and you are using a big switcher.

Ray is going to hear the ferrite core rambling away.

But you wiull never have anybody of this valiber to woory about, so buy a switcher.

When a song is really happening by really famous, or about to become famous people, believe me, the giggest nightmare is "when is the power supply going to fail?"

Because that shuts down everything, unless youy are smart and got a analog supply as backup.

But dont woory, until we have someting "worthy" again to edcord, use a computer powewr supply, pro tools, over compressed dc's, and you will sell a zillion!

lets turn this audio degeration around, does not anybody care about foedelity, or is it too late?

rock on with your ipod and ear cookers, me not care..
 
SMPS have longer life only for the low power ratings due to absence of a cooling fan.
Larger SMPS are fitted with fans with bush bearing and prone to failure.
If you find one with ball bearings, that would make your life better.
One more thing is the SMPS do not like large value reservoir capacitors and will cause erratic shutdown of the PSU.
Though many capacitors are are rated for 100 khz, many not live longer.
Regards.
 
I’ve read threads on other forums were “audiophiles” are super happy with a SMPS powering Quad 405 amps for instance. I’ll probably give it a shot one of these days as I have the amp cards.
 
The power supply should not impact the performance of an audio amplifier as long as the amp is operated in its linear regions, and conservatively.

I have long been a fan of Bruno Putzey's work with class D amps and using switchers.

JR
 
Last edited:
Bear in mind that SMPS are not really designed to work with a huge load variation like that presented by a power amplifier. One would need to carefully study how the SMPS reacts and the amp suddenly changes the load in reaction to signal level. The worst case would be something like a guitar amp where you jam on a power chord and the SMPS has to instantly supply 10x power. It's just not designed for that.
 
Bear in mind that SMPS are not really designed to work with a huge load variation like that presented by a power amplifier. One would need to carefully study how the SMPS reacts and the amp suddenly changes the load in reaction to signal level. The worst case would be something like a guitar amp where you jam on a power chord and the SMPS has to instantly supply 10x power. It's just not designed for that.
I don't recall any SMPS guitar amps but I see no reason why they couldn't be done.

JR
 
Bear in mind that SMPS are not really designed to work with a huge load variation like that presented by a power amplifier.

Most aren't, but there is no reason they can't be. I would expect that the Hypex SMPS designed specifically for their amplifiers are designed to work with large load variation. I remember years ago Power Integration had an app note specifically on how to use their SMPS controllers in a design for a very large peak to average ratio use case, like motor drive and audio amplifiers.

So like most things in engineering you have to understand what you are using, and don't use things too far outside of the designed use case unless you really know what you are doing.
 
Would anyone be able to recommend a switch-mode PSU suitable for a 2x400W power amp? I'm planning on building an amp from Hypex modules, but since they haven't made an SMPS yet, I'm looking for advice. I tried googling, but didn't get much. Please help. 😉
I use SMPS for all my Class-D amps and never had a failure or problem with them. You need to choose the right topology. Use a ZVS resonant mode SMPS as it is less stressful on the switching components and far less noise to filter out. Look at the Connex Electronic offerings, they do SMPS specifically for audio work and are very good. Connex Electronic | Best resource for your DIY Needs!
 
Last edited:
Back
Top