KISS 133
KISS 133
THE FAST CAP, THE FULL CAP, THE LITTLE HELPER CAP
by Andre Jute, Claudio Bonavolta & Thorsten Loesch
without the aid of fisticuffs!
How much capacitance does your filter need and how should it be disposed
OR
Is there such a thing as a fast cap or a fast power supply?
An innocent abroad asked a simple question on a newsgroup.
Several flame wars later he received two good answers...
and one off the wall.
I would like to know what is the correct method to "speed up" a 100uf X 100uf @500V cap. Someone told me that by soldering some small caps on the electrolytics cap, my amp would sound better in the high. What do you think? What can I do? Thanks.
A COMMON SENSE VIEW
from André Jute
First, inspect the circuit. Assuming a tube amp with up to say 500V B+, assuming that the existing storage capacity (if not its disposition) is correctly specified for a sound reason, and assuming the dual cap is in the power supply filtering:
--If the caps you mention are wired in series (stacked 100uF on top of 100uF) the total value is 50uF/1000V. It would be slightly odd if you had any requirement to "speed it up". 50uF, assuming that the choke is correctly specced, is assumed by many (including me) to be a maximum in considering the optimum balance between noise (ripple) reduction and bass/speed/cost/size etc. (The exception is single-ended--some guys on the cutting edge are down to single 8uF caps in their choke input or pi filters, sacrificing everything for speed.)
--If the caps are wired in parallel across the power lines, the value is 100uF/500V plus another 100uF/500V for a total value of 200uF/500V. This would, from empirical observation, slow down the sound of the generality of amps noticeably (and also the higher the B+ the more a big capacitance will slow it down, because the energy storage of a capacitance is proportional to the square of the voltage on it). The correct way to speed up the sound is therefore, if you can find the space, to fit 4x 50uF/500V caps in parallel for the same total value of 200uF/500V. You would have the same storage capacity in joules, but the cycling rate of charge/discharge would be very much faster and your amp would sound quicker and possibly better balanced. Replacing elecrolytics with Solens (which are pretty small as polyprops go--try The Parts Connection for supplies) would give you better quality sound too.
--Bypassing of caps has a certain low-level audiophile acceptance, mainly promoted by the makers and vendors of expensive caps for the benefit of their profits, but sonically I have always failed to hear a difference when the main cap is wrongly specced, and if the main caps are specced right, you don't need the bypasses. It is not surprising that those who can keep a perspective in the face of substantial sales pressure fail to hear a difference: back in the 1980s Elektor Electronics conducted both listening and measurement tests with all kinds of caps, and concluded that bypasses were a waste of money and time. You should consider that each cap has noise, so if you add bypasses (or even bypasses with their own bypasses as some do), you are simply adding more noise to your amp. That is another good reason to go to film caps--less noise in itself makes an amp sound quicker!
Copyright © Andre Jute 1995, 1997, 2006
IF YOU'VE GOT IT, FLAUNT IT
an entirely different view
from Claudio Bonavolta
"I suppose we're talking about hi-fi here ... "
If a "forte" occurs, the final stage of caps should be able to furnish the power without losing too much voltage. Just an example, if you have to furnish 10W during 10 seconds (well, it's quite a small climax...), you need 100 Joules that means, under 500V, a total capacity of 800uF. Well, it's a capacity needed if you discharge the cap completely ! But in hi-fi we need to keep the voltage drop of the final stage as low as possible, so imagine what should be the real capacity value for such a task. Sure, the caps are "refilled" by the previous stages, but aren't we talking about *fast* supplies ?!
So power supplies must not be large, they must be *huge*.
Music is mainly transients, nothing to do with the usual steady 1KHz sine wave used in power tests...
So power supplies must be *fast*.
The only storage component that are sufficiently fast are the capacitors.
Capacitors have the following caracteristics: -capacity - nominal and max voltage - physical dimensions
*but*
good specs should also add: - serial resistance, ESR - inductance - max continuous current - slew-rate in Volts per microseconds V/us
Electrolytics have a big capacity but a lower slew-rate, higher ESR and inductance.
Polypropylenes have a low capacity but a high slew-rate, low ESR and inductance.
The best idea would be to use only polypropylenes (or other fast type like oiled paper, ...) but that will result in monsters if you want huge capacities or in an unsufficient supply (quantity of Joules instantly available) if you keep the real estate demand of your amp within reasonable bounds.
Bypassing electrolytic caps with polypropylenes *does* work. The majority of commercial amps have poor supplies both in quantity (capacity) and quality (low and not frequency-related ESR, high slew-rate and max current). Because the components involved in a power supply cost a lot and they must keep the prices down, the manufacturers save on parts ... The power supply is a place where you can improve your amp a lot.
Here are some examples:
Modifying an existing full-range amp
As you're limited by space, it's difficult to add big components inside but you can generally add a polypropylene cap on the supply. I personally solder the polypropylene cap directly on the tranny. The cathode bypass capacitor should also be bypassed by a polypropylene of it is electrolytic, the normal case. To finish the "standard" improvement, you can also replace the coupling caps by high quality ones.
I've modified an existing amp (a 50W classic 6550 PP). Inside the original chassis, I added/changed the caps as mentionned above: 2.2uF on the trannies and cathodes and extremely fast, 1800V/us, polypropylenes as coupling caps. That improved the mids and highs in an impressive way, much more clarity, instrumental focus and an impression that the amp goes higher.
But I wanted to do it better. I built an external box with large electrolytics (around 2300uF per channel), changed the supply from a single to a dual supply, bypassed by polypropylenes caps used in pulse applications (32uF), bypassed these by high current polypropylenes (4.7uf/750V, 12A max, 80v/us) and bypassed the latter by small and faster polypropylenes (0.22uF). The bass changed a lot, incredibly more tight, giving more air to the mids and highs. Dynamics have increased a lot, there is much more difference between soft and loud sounds. No golden ears needed at all! Hum was not a concern (is it really in an amp ?!) but it has lowered further. Of course, if you're building such an amp from scratch, it's a lot easier ... All my listening sessions are done with my girl-friend who is absolutely not interested in hi-fi and understands nothing about technical aspects, just to make sure that what I'm hearing is the truth and not just an idea driven by technical convictions ... I insist: these differences have nothing to do with minor changes !
Multiamplification, bass amp
As all the power in music is located in that region, you need a huge amount of Joules, that means a big, big capacity. Impossible to obtain without electrolytics. The chance is that electrolytics are not so bad in the bass spectrum. Anyway, bypass them with some big polypropylenes and some smaller fast polypropylenes. In bass amps, "quantity" is extremely important.
Multiamplification, mid and high amps
You don't need a large capacity there so use only fast caps like polypropylenes, Teflon, oiled paper, ...
If you want to use huge supplies pay attention to some points: - keep a reasonable cap value [i.e. low, cf rectifier specs!--André] after the valve as increasing it will blow the valve quickly. Silicon diodes are less sensitive to such current peaks but it is often the fuse that will blow: increasing its value could be dangerous as a problem can occur without making the fuse blow ... - you can use large caps *after* the choke, as choke will limit the current to acceptable values. Chokes do not like quickly changing currents and react to them - use a two-step charging procedure through first a power resistor and bypass it after a while, to avoid these large current appeals - use a discharging [bleeder] circuit to avoid to keep high voltages on the tubes anode after it has cooled down.
Regulated supplies
I accept regulated supplies only if they are followed by a huge amount of caps, but then where is their interest? Keep them for preamps, bias and screen supplies ...
I've seen someone in this thread pushing switching supplies. Gee, high power 100KHz pulses have *STRICTLY* nothing to do in hi-fi !!!! Some audiophiles avoid to use silicon diodes because the commuting peaks they generate are difficult to filter, so imagine the difficulty of filtering such full power impulses... Leave such supplies to circuits where power consumption is the main concern ...
Copyright © Claudio Bonavolta 1998
AND NOW, FOR THE TRULY COMMITTED
Thorsten "EZEE" Loesch
explains how to do without the evils of AC altogether
AC Power is BAD. You MUST use batteries and bypass them with Polypropylene Film + Foil Capacitors of large value. Yes I know that you need plenty of Batteries (my Preamp will use 15 12V/ 1.1AH Lead Acid Batteries in Series for B+) but believe me, André, it's worth it....
Copyright © Thorsten Loech 1998
THE VOLTAGES IN THIS AMP WILL KILL YOU.
GET EXPERIENCED SUPERVISION IF IT IS YOUR FIRST TUBE AMP
All text and illustration is respectively
Copyright © Andre Jute 1995, 1997, 2006
Copyright © Claudio Bonavolta 1998
Copyright © Thorsten Loech 1998
and may not be reproduced except in the thread KISS xxx on rec.audio.tubes.