storage caps AFTER the SMPS - diyAudio
Go Back   Home > Forums > Amplifiers > Power Supplies

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 30th May 2011, 06:03 AM   #1
diyAudio Member
 
Join Date: May 2010
Location: was Chicago IL, now Long Beach CA
Default storage caps AFTER the SMPS

ref: Huge rail caps needed w/reg SMPS?

I keep running into advice that when you use a switching mode power supply for a power amp, the electolytics shouldn't be very big. Which makes sense if you only think about filtering out the RF switching frequency square wave, as opposed to old-school amps and suppplies filtering out the lower-frequency 60 HZ output of an unregulated transformer. And when I built bench supplies that's what I did.

On the other hand...when I made power supplies for audio amplifiers using unregulated power supplies, I always put in WAY WAY bigger capacitors (or capacitor banks) than were needed just for filtering. That was because of the nature of music signals I was amplifying. Adding a lot of capacitance at some point got almost like batteries instead of just filters, it allowed the power supply to handle the much larger occasional short-duration demands common in music. Of course they also helped correct for the poor "regulation" of the unregulated transformers in those primitive systems, and correcting that sag might be seen as fitlering from another perspective. Hirsch-Houck labs used to specify "dynamic power" and/or "dynamic headroom" for a specified pulse duration.

My amps with switching-mode regulators have very good continuous output rating, but no additional output capability for short-duration pulses. I'm not trying to handle an entire orchestral crescendo, but I would like to handle note attack and percussion impact with some recovery time. For some kinds of material I also sometimes use a dynamic range enhancer, EQ, and a digital sub-harmonic synthesizer. Combined, these increase this peak demand. I call it slam. My friends call it punchy.

To get the same punchy sound from a class "D" amp (or, unfortunately, conventional amp with a SMPS) it seems I have to use an amp with a much higer continuous rating. Which is not surprising to me. I've had 200 watt continuous / channel amps that handle short-duratoin 810 watt peak demands on both channels with only a little compression and distortion. But...My 800 watt / chan amplifer with switching mode power supply won't handle a short-duration 810 watt peak demand any better than my class-D 800 watt / chan amp.

The switching-mode component (whether amp or power supply) regulates nicely, but has NO additional headroom available for peaks. This doesn't seem to me to be a necessary restriction for a SMPS for a conventional amp if you add caps and don't mind violating the purist desire for the good regulation of the the SMPS. I'd like to add many large caps after the SMPS. They're not small or cheap, but they are cheap compared to a larger amp. It's not any more objectionable to my ethics or aesthetics than using class A/B. Good regulation within a power range, but when a demand exceeds that range what do I want the amp to do? Limit the peak? Attempt to handle the peak with a short subsequent recovery compression time? Attempt to handle the peak with inaudible recovery? It seems to me that a lot of capacitance should make it possible to avoid any audible "breathing" compression during recovery. Right now 2 channels share the same supply limitations, or each can enjoy any available excess power when the signal is very different between channels. I imagine any 'breathing" compression or reduction in volume due to supply sag during the recovery phase as the caps refill could also be a problem.

With the right soft-start circuitry and competent subsequent amplifier sections, is there any reason I can't add a LOT more capacitors after the SMPS? Some have said it might become unstable. Indeed under some circumstances it will occasionally increase the duration of full load on the power supply and push the amp output section harder for short durations with recovery time. I'd bet the power supply and the output sections would both live, especially since the specific amp I'm using is fan-cooled with a LOT of MOSFETs in old-school 2-screw TO-3 cases instead of today's TO-220 cases.

It seems to me I could have the best of all worlds. The "dynamic transient headroom" punchy impact slam of a huge old-school amp, nearly the compactness and power of a SMPS, nearly the continuous power of a class-d but less efficient.

I originally posed a similar question in the wrong forum:
Adding caps to commercial amp's switching power supply
I'm a genuine modder, hopping-up and experimenting is in my genes. I'm really inclined to build outboard capacitor banks in seperate cabinets which I can plug into these amps.

When I used to do semi-pro PA I used to add cap banks to a lot of my amps. The dynamic range of live material can be incredible. But I've never added caps to an amp with a SMPS. Yet.
  Reply With Quote
Old 30th May 2011, 06:07 AM   #2
diyAudio Member
 
Join Date: May 2010
Location: was Chicago IL, now Long Beach CA
Then there's the subject of the psycho-acoustic aspects of dynamic headroom and what that reserve "should" sound like. We are so used to the sound of some compression and non-linearity distortion in musical peaks that we associate it with a "loud" sound. Having some guitar amp background I am very aware of such things, but that's not really what I usually want in a high-fidelity reproduction system.
  Reply With Quote
Old 30th May 2011, 11:06 PM   #3
! is offline !  United States
diyAudio Member
 
Join Date: Jun 2005
Location: Midwest
Once you pick a power supply and know its rating for ripple, that is only a rating of the output at the connector for it. The rest of your (amp) powered device, the design, trace impedance, or even a need or desire for less ripple than the PSU is spec'd for, determines how much capacitance you need.

The goal is NOT to pick a lot of capacitance or a little, it is to have both! Bulk storage capacitors would be not too far away from the output transistors (or amp IC if the output transistors are integrated into it) but your low capacitance, decoupling capacitors with very low ESR would go as close to the chip (or transistor) pins as possible.

This assures the best electrical function and reduction of noise. It does not mean you will or won't like the way an amp sounds, lots of people do technically (electronically) incorrect or sub-optimal designs but find they like the harmonic effect it causes so for them it was the right choice.

Every now and then you'll find someone mentioning a resonant effect with more than one capacitor. Yes it is possible, but uncommon. If it happens just change one of the two caps that's in parallel.

There is the issue of inrush current if the capacitance is quite high, but personally I always suggest using a PSU with inrush current limiting. Yes if you turn up the amp to the point that there is more and more power output you come closer and closer to blowing a transistor. Will you blow one? It is possible, depends on how loud you try to go.

It is better to have a lesser drop in rail voltage and longer recovery time, the main thing to avoid is high frequency noise, that coming from the PSU and that created by the switching of the mosfets and driver subcircuits themselves. Of course, if these things matter a LOT then you are a candidate for a class A amp, since it will draw the same power regardless of what sounds are playing, and a candidate for a linear power supply since all switchers produce high frequency noise... though as mentioned above, some people like the sound of certain things so for them it is the right choice.

Last edited by !; 30th May 2011 at 11:09 PM.
  Reply With Quote
Old 30th May 2011, 11:37 PM   #4
diyAudio Member
 
Join Date: May 2010
Location: was Chicago IL, now Long Beach CA
Nobody else has any comments on using a SMPS followed by lots of energy storage caps followed by a conventional amp? The folklore (fostered by Bob Carver selling powerful amps) is that reproducing a scissors snip thru efficient speakers still clips with 4800 watts, and that it is impossible to accurately reporoduce the top of that instantaneous transient with any amp. That may be a 1-millisecond burst that is 100,000 times the average. Of course it's irrlevant if your source is CDs.

We use large cap banks to improve conventional power supplies and endow them with dynamic headroom for short transients, not just to remove 60 hz hum. Class D amps are not able to deliver dynamic headroom beyond when the pulse width is 100% but we revel in the less important continuous ratings.

But it seems to me we undersize the caps after a SMPS (when followed by a conventional amp) for economical and compactness reasons. We again revel in the less important continuous rating. Yes, it's great to always operate within the fully-regulated capabilitiy of a SMPS, but is it such a heresy to give it some additional headroom via a bank of caps, even acknowledging that the bank of caps is less than ideal and may sag for some recovery time after drained beyond what the SMPS can refresh in a few cycles (and the amp may operatoe outside of its most linear region?

Last edited by cyclecamper; 30th May 2011 at 11:41 PM.
  Reply With Quote
Old 30th May 2011, 11:50 PM   #5
diyAudio Member
 
Join Date: May 2010
Location: was Chicago IL, now Long Beach CA
I'm sure it's my guitar amp background that makes me think so much about the nature of how the amp behaves outside its ideal operating limits. Most engineers rarely address that, they just don't intend theri equipemnt to be operated outside its specified limits. Yet they often are.

What I'm proposing is to leave the decoupling caps on the output board, and the big filter caps mounted by the transformer, but adding another entire enclosure filled with capacitors, with the intent of adding significant dynamic headroom for handling very short transient demands. That's the spec where these amps fail to deliver now.

I've done this to many amps with conventional power supplies, with very good results. But I've never done it to any amp with a SMPS so I'm a bit leery, especially since these are unsuppored. But now I have an extra amp to experiment with, and not put everything at risk.

Last edited by cyclecamper; 30th May 2011 at 11:52 PM.
  Reply With Quote
Old 31st May 2011, 05:29 AM   #6
! is offline !  United States
diyAudio Member
 
Join Date: Jun 2005
Location: Midwest
Put the decoupling caps as close to the (chip or transistor) pins as possible. This is one of the most important factors (when talking last detail/tweaking) that remains except for keeping the signal path as short and ground plane shielded as possible.

We are talking about random things though, idealizing only goes so far and then it is a matter of splitting hairs if your design is good otherwise. Specifics like seeing the schematic and a good pic of the PCB trace side and component specs would be a better set of info going forward...

but in the end, it is always about budget and conscience, if not we would all have 5000W class-A linear amps and kill baby seals to power it.
  Reply With Quote
Old 31st May 2011, 07:36 AM   #7
diyAudio Member
 
Join Date: May 2010
Location: was Chicago IL, now Long Beach CA
I'm not looking for a discusson of decoupling or filtering, I'm really asking about additional energy storage caps for power to handle longer transients, practically batteries LOL.

To some people that means additional filtering to handle power supply dips that occur during longer transient peak demands, so it's really a matter of perspective and preferred terminology.

Some have suggested I put a scope with memory on the supply during some loud transient peaks and the period between such peaks, and that would give me a good way to change my own perspective and have something concrete and quantitative to address.

Last edited by cyclecamper; 31st May 2011 at 07:41 AM.
  Reply With Quote
Old 31st May 2011, 08:11 AM   #8
diyAudio Member
 
Join Date: May 2010
Location: was Chicago IL, now Long Beach CA
I stumbled into some agreement with my premise:
http://passlabs.com/pdfs/articles/powersupply.pdf

"This can be a deep subject, but suffice it to say that I believe that some
of the same caveats apply to switching supplies as linear regulators.
Again, they should be rated far beyond the nominal current
requirements of the amplifier circuit, particularly as the switchers I have
seen usually degrade badly beyond their ratings. Also, it helps if the
power supply capacitors before and after the switcher are very
substantial. This is typically not the case, since one of the primary
motivations to use switchers is to save money."

He's talking about 100.000 mfd parallel with a small film bypass for a 1000 watt amp. For my 1800 watt amp, 200,000 mfd does not sound unreasonable for a hobbyist enthusiast.
  Reply With Quote
Old 31st May 2011, 05:23 PM   #9
! is offline !  United States
diyAudio Member
 
Join Date: Jun 2005
Location: Midwest
^ Filtering is what they are doing, it is the same thing, you are simply filtering less frequent high magnitude ripple on the power line caused by the powered device rather than the PSU.

The point that you made about overspec'd power supply capability is true but the expense limits it happening in many cases. A properly paired PSU will itself be able to handle the momentary peaks, but your circuit impedance makes it important to have the energy stored adjacent the output chip pins for least impedance.

There is no need to measure with a scope at the supply, you should be most interested at the power (input) at the amp output component, transistors. If your supply lines are not staying where you want them, only at this point in the circuit will it matter. You could keep the PSU at an ideal voltage but it won't remove the issue of what the power rail is doing downstream at this critical point in the circuit.

If you use an off-board capacitor bank like 200,000 mfd, keep something in mind. Not only is there the inrush current which you can easily enough limit, there is another factor and one I don't feel you have focused on. Decent switching PSU will shut off from under-voltage conditions. This means that not only would it turn off during the longer transient power demand if the power rail dropped much, it also means that if you put lots of capacitance after the switching PSU, when first turned on it may not be able to get the voltage up to the set range before it shuts off because it took too long charging up the capacitor.

One way around this latter issue is to put current limiting on the input to the capacitor bank, run a parallel supply line to the power rails in the amp board so your voltage recovery is faster, and isolation diodes on the output of the capacitor bank to the amp so they remain current limited on their input instead of just charging up through the parallel circuit path... but then you suffer the loss of the forward voltage drop across the isolation diodes so it is far from ideal, probably not practical. Some people avoid anything in parallel thinking it's a matter of ground loops, but this shouldn't matter since it is isolated to the supply line, assuming they both connect on the amp PCB at the same location.

Last edited by !; 31st May 2011 at 05:31 PM.
  Reply With Quote
Old 31st May 2011, 06:00 PM   #10
infinia is offline infinia  United States
diyAudio Member
 
infinia's Avatar
 
Join Date: May 2005
Location: Californication
If your SMPS output is regulated , then adding significant > 50% additional output capacitance is problematic and will not add anything except a potential for creating feedback loop instability ( very bad). Peak transient power within the audio BW and tone bursts is mostly controlled by the feedback loop time constants and current limit. Short term transient behavoir (the rise and fall times of squarewaves) is mostly controlled by output caps (decoupling) and reducing stray inductance to the audio amp and/or load. The design tradeoffs of a SMPS is usually between the output filter bandwidth to attenuate ripple freq at the switching freq. and loop BW controlling line and load regulation. The problem with SMPS is the load current limit must be set for steady state maximums depending on the thermal design limits.
so, like anything it depends on the details of the power supply design, the amp design targets and exactly what you are trying to achieve. It maybe as simple as moving the SMPS current limit set points above the threshold determined by the thermal envelope knowing/ hoping the transients for music will exceed or approach steady state designed in thermal limits.
__________________
like four million tons of hydrogen exploding on the sun
like the whisper of the termites building castles in the dust
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
best way to recondition Rifa caps after long storage qusp Parts 2 6th December 2010 06:34 PM
High Voltage Huge Storage CAPS for that Ultimate Valve Amp col Tubes / Valves 0 11th December 2009 03:37 AM
Low ESR caps in SMPS Stuey Power Supplies 2 19th May 2007 07:35 PM
Huge rail caps needed w/reg SMPS? nine76chris Power Supplies 1 6th November 2005 01:07 AM
lp storage ironcitymadman Analogue Source 0 13th July 2003 03:14 AM


New To Site? Need Help?

All times are GMT. The time now is 04:44 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2