[Folsom]

Hey guys, though I'd through out a nice little upgrade that cost little. This is for those using small switching DC adapters primarily - or rather those with far from audiophile supplies.

You may be aware most small boards have a power cap on them, and your adapter will have one or two on the output. Their quality varies, but if you are thinking oh no I don't solder, well you don't necessarily need to in order to improve your little single DC source amp. All you need to do is get a few decent capacitors and connect them across the DC power in(negative marked side of cap, short lead to ground; and long lead to positive voltage). This can be done crimping or soldering. The capacitors can be hot glued together and to whatever with their low voltage. They can be connected anywhere on the DC cord between powersupply and amp, preferably toward amp. Only perform this with power supply unplugged.

For inexpensive quality capacitors I'd use panasonic FC or FM. However you should use 4 or double the amount used from there each time you decide to use more (this increases noise filtering). My test was with 4 panny FC at 1500uf. It's more juice than I need, but they are not only responsive to voltage change, but also load down AC on the line (noise).

My supply is 4.5a. It's for a laptop and shielded well, for what it is.

I think some of you will be impressed with what a few dollars can do for a few dollars amp. It can be as effective as getting a new amplifier.

I spent all of a minute and a half to bring you guys this riveting schematic.





You could consider adding something like a 330pf capacitor with a resistor, those two in series, but in parallel with the other capacitors, for RF noise from extra inductance and what not, if you're worried. (plenty of places on the web to look this up) 

[mboxler]

Hi

Sorry for these stupid questions.

Using your example (I'll assume your supply is 19 volts), the 4 capacitors should be rated at, say, twice the supply voltage (40V)?

Do the capacitors store more current, effectively making the 4.5a supply "appear" larger when the amp suddenly requires more current?  In other words...is it better to get a larger supply with the mods, or just do the mods.

Thanks for the suggestion. 

Mike

[Folsom]

Both. You want lots of current, but power supplies always have dips. You want to correct that with capacitors. The bigger the capacitence the better the correction, but big caps don't always perform as fast, so you use multiple smaller ones. Also multiple smaller ones provide for filtration. It's fairly common knowledge but I swear no one does it because they don't know the significance because they never do it.

Keeping current supply up gives a lot of body to the music. That's part of what makes class A so nice.

[Folsom]

You may want to try more capacitors and capacitence than I used. The exact value from different setups where no benefits happen by adding more may vary. I'm going to add two more out of curiosity. At some point there are diminishing returns.

[Folsom]

For fun I just added 2 more 1500uf. Results, positive.

[vlad335]

Going to give this a go. Genius idea.

I have a Dayton DTA-100a with a 24v/5a power supply. I am looking at these caps.

http://www.mouser.com/Search/ProductDetail.aspx?R=EEU-FC1H152Lvirtualkey66720000virtualkey667-EEU-FC1H152L

How many would be optimal? 12?

[Folsom]

Yes those will work. 12? Well, it won't hurt! But I think 9 might be sufficient. Who am I though? Not your ears, so maybe do a little R&D, (add a few at a time).

My supply puts out close to 17v measured, claims 16v. I think (not at home).

[Speedskater]

You lost me at the schematic!  Why all these electrolytic capacitors across the signal input terminals?

[Russtafarian]

You lost me at the schematic!  Why all these electrolytic capacitors across the signal input terminals?

Good point.  Salis should spend another minute and a half on the schematic to indicate that the caps are placed across the DC power input terminals, not the signal input terminals.  Good idea though.  A next step might be to add a choke for a CLC filter.  The Felix thread will have some notes on this.

Russ

[Speedskater]

Randomly adding electrolytic capacitors to a component is seldom a good idea and sometimes blows up the rectifier diode bridge.
A jr. member of our forum 'gootee' added a lot of information to a long thread in another forum:

http://www.audiocircle.com/index.php?topic=117488.msg1238397#msg1238397
long thread:
http://www.diyaudio.com/forums/power-supplies/240955-resevoir-capacitors-chip-amps.html

While replacing old electrolytic capacitors with new units that have a higher voltage or temperature rating and a larger value (but with a similar footprint) can be a good idea.  Just adding lots of caps is a bad idea.

[Folsom]

Randomly adding electrolytic capacitors to a component is seldom a good idea and sometimes blows up the rectifier diode bridge.
A jr. member of our forum 'gootee' added a lot of information to a long thread in another forum:

http://www.audiocircle.com/index.php?topic=117488.msg1238397#msg1238397
long thread:
http://www.diyaudio.com/forums/power-supplies/240955-resevoir-capacitors-chip-amps.html

While replacing old electrolytic capacitors with new units that have a higher voltage or temperature rating and a larger value (but with a similar footprint) can be a good idea.  Just adding lots of caps is a bad idea.

There is nothing to indicate this is a bad idea. First this is a recommendation for a switching power supply, with more than ample current to charge more than the capacitors and peak power for class D/T in these size categories. That said there isn't a rectifier diode bridge in this case to be concerned with. Even if there was one on the amplifier board (only some chinese units have some diodes for AC input) they are under 100% duty cycle with DC plugged in. Having two 60hz phases won't mean anything compared to 100% duty cycle.

Your links are not applicable to switching power supplies and class D/T amplifiers.

If there is a concern it's with added inductance building up some frequencies that people would use an RC network to deal with (no need here, and not worth the price as far as I'm concerned).

This isn't "just adding a bunch of caps" because anyone doing so is going to go over the rating, and this isn't talking about changing anything on the PCB board/topology of the class D/T amplifiers.

What have you written Speedskater? What do your ears tell you?

[Folsom]

Good point.  Salis should spend another minute and a half on the schematic to indicate that the caps are placed across the DC power input terminals, not the signal input terminals.  Good idea though.  A next step might be to add a choke for a CLC filter.  The Felix thread will have some notes on this.

Russ

I've been using digikey's schematic maker, it has very limited abilities.

[Quiet Earth]

When I first saw this post I also wondered why you would add a bunch of polarized caps to the signal inputs of an amplifier stage.  Now I see the little DC input designator, and I realize those are supposed to be the voltage input terminals.    Most of us are used to seeing the +voltage and ground designator coming off of the top and bottom of the block diagram triangle. So we goofed up interpreting your schematic. I think you should have clarified what you were trying to accomplish. Or maybe you did clarify and I had trouble interpreting. 

I will be honest and say that I still don't understand how this is any different than adding more capacitance across the DC line. Maybe you can explain it differently for those of us who don't understand what you are accomplishing. No offense intended in my request, I just think I might be missing the point of this tweak.

[Folsom]

I am not sure what you are talking about because this is adding capacitance across the DC line; but with smaller caps for speed and lots of it it so the drop is almost non-existant. Class D/T responds very well to this. In fact the quality seems dependent on the consistency of available (fast) current.

I'm trying to make a better one. But like I said the free program isn't great.

[Quiet Earth]

You mention current - lots of current (storage?), and "smaller caps for speed". You also mention enough capacitance to load down the AC line. All of these ideas seem to oppose each other, or mean different things.

When you say smaller caps (for speed?) I usually think of something much smaller than 1500uF. Are the FC and FM caps super low ESR caps? Is that what you are shooting for? Have you tried staggering smaller and smaller values to ladder the ESR ( and reactance etc), for a DIY super cap?

Sorry if I am missing the goal of this project. I am slow to catch on.

[Folsom]

Class D/T like to have lots of current. The way they work they need it in order to be responsive and act well in accordance with the demand of the signal. This is why people keep saying use a 5a SMPS on a .1-2a D/T amp. The differences between the SMPS's and their power reservoirs however is another affair; this has become noticeable because some D/T amps don't respond to a high amp SMPS.

The Panasonic FC/FM are very low impedance, and seem to be fast enough. There is a point of practicable vs. performance. Do you want to solder together 9000 1uf caps? I don't. My experience with FC/FM is that they are plenty adequate in this uf range for immediacy. In fact they provide well enough for some amplifiers to be a little on the forward side. I've used smaller on other things, yes. I think the key here is however more overall available current; at sustained levels that are more responsive than most SMPS's can offer (sort of, you're fighting AC fluctuations simultaneously). If it were a preamp or something only line level I certainly would run smaller caps, and at that wouldn't need as much power (typically).

The amount of capacitance here is at the voltage you are using to power your amplifiers, 12-24v, typically, and not 120v. You'll use around 25-50% of the SMPS's capacity a brief moment when you turn it on, to charge the caps. The capacitors in this setup on DC won't ever come close to discharging fully or half way, and even a quarter isn't likely. The intention is actually that they just stay charged and discharge for the extra bit of current the SMPS is too slow and/or incapable of at that moment. You need much more capacitance to keep the voltage stable - larger capacitors drop voltage much slower when you have a high ratio of capacitance vs. current need.

Large power supply sections of expensive amplifiers use much larger capacitors, but the better the quality the better the sound due to speed concerns. You can have 8x 10-40kuf caps in a unit, but the better they are the better the sound. You can easily spend $100 per cap to get this. But let's compare wattage here...

SMPS 12v x 5a = 60w , which is .5a at 120v. They're not that efficient so lets say .6a at 120v for 60w, for the average SMPS for a class D/T. Now the amplifier only uses about maybe at it's high capacity 2a of 12v, or 24w. The capacitors will charge once and barely discharge at all while the SMPS is on. The amount of capacitance area I'm recommending is roughly one to one for the amount of power usage, so at max when the SMPS turns on the caps will draw about the same as the amp could under full load, 24w. But because they are on DC, they only discharge with the SMPS lags in some way. However you have to remember that they're only going to discharge maybe 1-10% of their charge under fairly extreme lag from the SMPS. That means when the capacitors recharge in between lags, they're only using a couple watts. You are not loading down the AC. This is a very practical way that costs you almost nothing. A class A amplifier can easily burn 240-1200w of power continuously because of their design. Really, class D/T is super efficient even when you add enough capacitance to double the current usage it can use.

*capacitors are actually a bit more complicated due to how they work over time intervals, in-rush current will be higher with capacitors but it's nothing to be concerned with since it happens once, briefly. SMPS's can easily handle small peak demands, the average SMPS can handle a lot more load if it could keep cool, and in-rush current isn't enough to even get the heat going. It's not really anything anyone has to know to get very good benefits from a super simple modification, to low voltage power, for only a few bucks.

I think anyone who gives it a try will pretty pleased with the results. What it does for sound is something impressive that in the case of some very expensive class D units isn't happening yet.

[vlad335]

NICE! Definite upgrade.

I used nine 1500uf pannys in line with male and female plugs on each end as I didn't want to hack up my power supply cord. This way I can easily tell the difference with the caps inline and out of the loop. With the caps out of the circuit, the music simply falls flat and sounds smeared in comparison. Inline, sound quality noticeably goes up a couple notches for my Dayton DTA-100A. Soundstage is huge with increased definition between instruments. The premise of the OP is correct, it is like I bought a much better amplifier.

Love doing these kind of DIY upgrades. Thanks Salis Audio!

[Folsom]

Glad to hear it Vlad!

[raysracing]

A little off topic, but would making a capacitor bank for a sub that thumps when any other power source in the room is turned on or off filter that thump?  Sorry, very newbie!  Without buying a megabuck power conditioner just for the sub (soon to be two) I want to know options.

[Folsom]

I think your question would be easier to answer if we know what sub, amp, crossover, pre, source, happened to be going on.

It does not sound like power caps would change it. Some amplifiers have limiting devices to prevent this (common in smps too). It'd probably be easiest to modify the sub amplifier - if it's the culprit. I'd post in The Lab about it. The other concer that'll come up is if your abode is grounded correctly.