diyAudio (
-   Solid State (
-   -   any limit on reservoir capacitance? (

lt cdr data 23rd December 2005 04:36 PM

any limit on reservoir capacitance?
I have a 10 yr old japanese amp, and excellent it is too, with 10,000uf 50 volt caps as the main reservoirs. I am going to replace them with new ones however can I increase the capacity?

Will it blow the mains transformer and/or brigdge rectifier ( or cause other problems?) if I significantly uprate the value of these? I have seen 27,000uf I fancy, but better ask first.

I should be able to get away with 12-15,000uf?


poobah 23rd December 2005 04:42 PM


All of those things can happen... do not increase the size of the caps unless you are willing to analize the entire power supply design.

If the capacitors are of proper size to start with, increasing the value will not improve the sound.


lineup 23rd December 2005 04:42 PM

You can add twice or five times or more caps.
Sometimes you can add 0.05-0.10 Ohm resistor before caps,
to limit inrush a bit.

Adding more and more caps, will not harm. (will harm your wallet though)
But question is if it will make anything better.
So much better that final performance improvement
can be noticed.

Kevinbd 23rd December 2005 05:45 PM

I think you need to read the earlier posts of ALW and Jan Didden about this. Basically, toriodal tranformers are wideband devices - they can resonate in the MHZ region (I think ALW measured this).

Increasing the capacitance can shorten the conduction angle, and consequently send out much higher frequency trash which is much harder for the amplifier to reject. The amps PSRR cannot reject HF trash anywhere near as good as rejecting LF benign 60 / 120 HZ ripple. I think you need to use massive toriodals to get around this problem.

I think it's the low impedance in the PSU your looking for. A regulated power supply of high quality is the answer I think. I think your amp still has to work like an ideal voltage source down to 2 ohm loads to really drive floorstanders. (irrespective of PSU). Double it's current into half the load impedance to say 2 ohms.

A couple of things, the diodes can generate copious amounts of RF noise. So slow recovery and RC Snubbers (properly designed) should be preferred.

I personally would go for the least amount of capacitance for the required ripple rating. Siemens Sikorel and some Elna caps can discharge 2-3 times the current for the same capacitance as others. Also computer grade capacitors tend to be quite good also. They tend to be much faster at discharging - good for transients I reckon.

I am going to go for an IE transformer (screened) for my power amps as opposed to a toriodal and will add decent mains filter. Lower interwinding capacitance etc.

Best Regards


AndrewT 23rd December 2005 05:50 PM

that's a big jump 10mF to 27mF.
The rectifier will be the second thing to be at risk, the first being the fuses.

To slow down the charging try a CRC PSU using 0r1 to 0r47 as suggested earlier.

You can even use a TC PSU (T=thermistor)
or CTC or TCRC. lots of options.

The last C should be the larger value.

Or try using the same budget to add on big plastic film caps//ceramic caps to bypass the electrolytics.

Upupa Epops 23rd December 2005 06:01 PM

Unneccessary work... don't do it... ;)

lt cdr data 23rd December 2005 06:11 PM

mines an el core transformer.

ok I will stick with around 10,000 muff :) I really need snap in ones, so are there any recommended ones like say the elnas? From what I can see on the sikorels, they aren't the right fitting.

thanks again

poobah 23rd December 2005 06:23 PM

Everybody, for once, is telling you pretty much the same thing.

Unless you hear hum in your output leave the caps alone. Now that is not to say that if you DO hear hum that those caps are to blame either. At ten years of age the caps are nearing their "rated" life. But, in reality, these caps typicaaly last much longer. So if it ain't broke don't fix it. You will not hear any astonishing change in "imaging" or "transparency" if you change the caps.

What you won't hear is the new CD's you could have bought with the money... :D ;)

amplifierguru 23rd December 2005 09:49 PM

Hi it cdr data,

long time no hear. Just an aside, was it you who posted a schema of a chip driven common source commercial amp some 6 months or so ago using IRFP's that you had probs with? The make/link?

On to your 10,000uF 50V. Large PS C's often become inductive at HF well before smaller ones, so you could replace the 10,000 uF each with 3 x 3,300 uF with better and paralleled HF extension/ESR and add small film bypasses. Or 5 x 2,200 uF or 3 x 4,700uF (as in my GB150ABS) although this is a 41% increase which, as said before by other posters, your fuses and bridge will need to accomodate higher inrush current.

Your Chemi-Con do some good ones discussed in an earlier thread.


quasi 23rd December 2005 10:21 PM


If I needed to replace the caps then I would go ahead and increase the capacitance. 27,000uF should be fine. I doubt the circuitry is complex or sensitive enough to know the difference.

This is because your not actually talking about driving your amp any harder, so the average strain on the rectifier & transformer should not change.

I would though upgrade the power supply rectifiers in case they were too minimal (the larger capacity will draw a higher surge current) and I would examine the fusing. You may have to change to a slow blow or increase to a slightly higher rating.

As Greg suggested if you have the room, go for parallel capacitors. This will give you a lower ESR and inductance than single units.

Oh...if there aren't any, place parallel 0.1 to 0.47 high voltage polyester (or MKT) capacitors in parallel with the main caps.


All times are GMT. The time now is 02:10 PM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 18.75%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio