Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 28, 2024, 04:41:43 AM

Login with username, password and session length

Recent Posts

 

Sizing Power Supply Filter Caps for a Chip Amp

Started by JayFett, September 23, 2012, 11:25:23 PM

Previous topic - Next topic

JayFett

How does one determine how much power-supply-filter capacitance is required for a chip-amp based guitar amp? Is it as simple as using the chip-amp data sheet demo circuit values?

J M Fahey

The proper way is deciding on how much ripple you can accept, and then calculating what capacitor will meet that .
And that's a personal choice.
But since most guitar amp PSU are exactly the same, there's a couple classic values used by most, and which practice has shown to work.
One of such classic values is 4700uF per rail for amps with 4 ohm load.
Variations: real cheap people use 2200uF per rail (Marshall Valvestate and many others) and a few ones, in special occasions (such as big Bass power amps) splurge and use 10000uF per rail, for more solid bass.
We are talking about 4 ohm loads.
If the amp is stereo or is expected to regularly drive 2 ohms, double values.
For an 8 ohm only amp, this would mean you *could* use 1000uF caps per rail, but most nobody does, because the cost savings (considering the total amp bill of materials) between 1000uF and 2200uF are nil.
And yes, most datasheets agree with this.
Be careful on 1 detail though: *many* chipamp datasheet show a couple 1000 to 2200uF capacitors on the Pwr PCB itself.
These are *not* the main PSU ones (which are elsewhere) but extra ones, *real* close to the chipamp, for stability purposes.

joecool85

Quote from: J M Fahey on September 24, 2012, 04:24:43 AM
The proper way is deciding on how much ripple you can accept, and then calculating what capacitor will meet that .
And that's a personal choice.
But since most guitar amp PSU are exactly the same, there's a couple classic values used by most, and which practice has shown to work.
One of such classic values is 4700uF per rail for amps with 4 ohm load.
Variations: real cheap people use 2200uF per rail (Marshall Valvestate and many others) and a few ones, in special occasions (such as big Bass power amps) splurge and use 10000uF per rail, for more solid bass.
We are talking about 4 ohm loads.
If the amp is stereo or is expected to regularly drive 2 ohms, double values.
For an 8 ohm only amp, this would mean you *could* use 1000uF caps per rail, but most nobody does, because the cost savings (considering the total amp bill of materials) between 1000uF and 2200uF are nil.
And yes, most datasheets agree with this.
Be careful on 1 detail though: *many* chipamp datasheet show a couple 1000 to 2200uF capacitors on the Pwr PCB itself.
These are *not* the main PSU ones (which are elsewhere) but extra ones, *real* close to the chipamp, for stability purposes.

JM, could you post a PCB like you're talking about where the "main" PSU caps aren't on the PCB...I've not seen anything like this.
Life is what you make it.
Still rockin' the Dean Markley K-20X
thatraymond.com

J M Fahey

You have, only you didn't recognize it as such. ;)

I'm posting 4 PCBs, for LM3886 , TDA7294, where you clearly see they include a couple 2200uF (1000 in one of them) on the power board itself, at 1 inch or less from the chip.

The "real" PSU, with its bridge rectifier and main Filter caps, is "somewhere else", meaning in a different PCB, joined to the Power one with wires which can vary from 1" to whatever you want, even 20" or 40" if you wish.

Problem is Power Chips are stabilized too tightly (remember they all have *killer* specs) but this leads on they being on the edge of unstability.







The practice is "suggested" in the datasheets themselves, not with clear words but in the example layouts.

To comply with both "rules" (big caps both on the PSU and Pwr Amp) you can, say, use 2 x 2200uF ones in the PSU and another extra pair just by the chip.


JayFett

Thanks for the answers.
So... if the power supply caps are intentionally "under-sized", will the amp then exhibit voltage sag, and mimic the 'note bloom' that certain tube amps have when played at higher volume? Or will it just show up as weak bass in the chip guitar amp? Or just noise?

phatt

#5
Quote from: JayFett on September 24, 2012, 09:51:55 PM
Thanks for the answers.
So... if the power supply caps are intentionally "under-sized", will the amp then exhibit voltage sag, and mimic the 'note bloom' that certain tube amps have when played at higher volume? Or will it just show up as weak bass in the chip guitar amp? Or just noise?
Hi Jayfet,
Better minds here will have more in depth answers,,, but generally NO it no work that way as they are very different animals. :'(
Most SS stuff is DC coupled (Valve circuits are mostly AC coupled)
Add No output transformer, plus most SS power stages have a lot of feedback whereas a lot Valve stuff had little or none. The sonic fingerprint is very different.

With smaller supply caps LM3886 will distort a little more on big signals but there is no magic all singing and dancing sustain/sag/compression like old Valve stuff did. With most SS gear you do all the OD tricks with the preamp stages and most SS power stages run clean though you can play a sag trick in the supply rails by placing a small value series resistor between the first Caps and the second ones. Or put a 40 watt light globe in the mains primary wiring. :tu: I've done some experiments with stuff like that but you have to have the volume up loud for it to work.
Easier to work with a clean loud poweramp and use preamp tricks for OD effect.

Re Big Caps;
When talking guitar frequencies and a 40 Watt LM3886 Amp,, the difference between 10,000uF and 3,300uF is hardly going to be noticed. You only start thinking Big supply when wanting to design a 400 Watt Bass rig.

From the point of view of dollar per distortion produced,,The big advantage to SS tricks is you get better Level Control over distortion.
Yes you can do that with Valves but it will cost you much more so for a novice SS is far easier and cheaper to work with.
I use an old SS KB 50 Laney keyboard amp for guitar!! The power Amp is about as basic as one can build a SS Amp.
The power transformer is small and the main caps are only 3,300uF.
the preamp is about as basic as one could build and except for the built in reverb it's no magic smoke Soldarno killer Amplifier.

(The Laney Amp is still a lot more grunty than the LM3886 poweramp running from the very same supply.
My guess is chip amps don't have big TO3 metal case power transistors)

Anyway whatever you decide for power stage remember the Tone game, My setup still gets guitar players asking WTF is in the gear you use? Even though it looks like crap. :o

I play at a local muso club  on weekends and guys turn up with brand new Deluxe reissues,, Cyber deluxe,, Trademark 60,, Carvin Legacy,, to name a few but No one has yet said my sound sucks!!! So go figure??
To rub salt into there wounded pride I also often use my $150 elcheapo strat copy which is about the cheapest nastiest guitar you can buy in this part of the world. :)

Most of the mojo comes from my preamp setups.
From my observation it's ALL about TONE Shaping,,*It is at the heart of every great sound*,,
Get that part right and it matters little if you have Glass or Sand for amplification.
Far to many people focus on distortion and miss the point as distortion without tone shaping is just Noise x Wattage.
With most SS opamp powered circuits, Tone shaping before and after distortion is a big key to getting good results. If you are hunting metal type stuff you may may need even more.
Until recently I was using an old HiFi Graphic EQ right before the power amp but now I've refined it all down to a simple 2 knob notch filter. l figure I have enough knobs already. lol

The really hard part to explain is that there are just so many ways that lead to the same result it's hard to say which is best. For me it's all budget driven as I'm never going to be rich so I work with basic off the shelf stuff that is easily replaced if something goes bang.
Phil.

J M Fahey

FULLY
    FULLY
        FULLY 
            AGREE    :tu:
Phatt said it all.  :dbtu:


JayFett

Well... there you have it. Thank you for explaining it so well. I was just beginning to suspect that preamp tone shaping is where to put the effort, and you have confirmed it.  :tu:


joecool85

Life is what you make it.
Still rockin' the Dean Markley K-20X
thatraymond.com

Loudthud

Quote from: JayFett on September 24, 2012, 09:51:55 PM
Thanks for the answers.
So... if the power supply caps are intentionally "under-sized", will the amp then exhibit voltage sag, and mimic the 'note bloom' that certain tube amps have when played at higher volume? Or will it just show up as weak bass in the chip guitar amp? Or just noise?

I've been looking at this issue in an attempt to mimic the dynamics of tube guitar amps. Note bloom seems to be a function of gain change with power supply sag and recovery. Hard to do with solid state amps because they use so much feedback unless a deliberate compressor type circuit is used. Note attack is a function of power supply sag but the power amp has to swing to the rails.

There is a certain amount of power supply ripple that gets to the speaker when a tube power amp clips. I'm using a TDA2030 and can run off of an AC wall wart or a regulated DC power supply. The difference is subtle, but I would say the tone is a little more grunty when the AC supply is used. With my current proto I only get ripple on one rail, I'll have to go to a bridge setup to get it on both sides of the speaker waveform.

Sag also affects the preamp in tube amps. Most SS amps regulate the supply(s) to the preamp and with opamps there wouldn't be any gain change anyway. My amp is using a single rail with a capacitor coupled speaker and an unregulated supply to the preamp. The output bounces around a little and I like the sound. I'm trying to get something put together so other people can play it to see if they like it.