From what I understand, using many smaller caps is both cheaper, and they have a faster recovery time.
10 responses Add your response
I assume you mean the power supply filter capacitors. The function of these caps is twofold: to smooth the pulsating electrical current from the power supply and to provide energy storage for (temporary) high current draw resulting from low impedance.
Using the caps for filtering is the top priority. The energy storage is a secondary consideration - one that is provided by the power supply transformer. So why one, two, or more caps? One factor is cost. Audio grade caps are very expensive and to keep down cost, a good place to start is the power supply.
The other factor is performance. One difference between large and small capacitors is that smaller caps fill and discharge quicker. This enables the amp to be faster, which enables a higher bandwidth and all other nice parameters. Another advatage to multiple caps is that the energy storage is increased along with the advantage of faster charge discharge. The biggest advantage of multiple caps is that the output impedance is lowered when in parallel. This enables better voltage regulation at the outlet of the power supply. Voltage regulation is expensive and is one of the biggest factors is the amplifiers' sound quality.
The advantage of a single, larger cap is one of cost and lower resistance due to induction. One large cap can filter and store as well as a series of smaller caps but the charge/discharge time is slower as you go for more power from the amp. But a sinlge cap is only considered if the available sizes are what is required in terms of rated power. Sometimes, higher power amps require multiple caps because the total amount of capacitance needed cannot be supplied by one cap.
It's basically a cost/performance consideration. Why would an airliner have two engines instead of four - two can take you just as far and fast (but at a decreased payload) but obviously two are cheaper than four in terms of first cost and maintenance. But if you need to go farther with more payload, you go with the four. Same with amps - the need to maintain performance down to 2, 1 or half an ohm may or may not be a consideration unless cost is no object. But if it is, then you trade off.
Bigger caps typically have a higher ESR. To try and put that into English, a high ESR is equivalent to a higher series resistance / slower cycling of power delivery & charging. By using multiple smaller caps AND selecting specific values that are widely used in other industries, not only is the ESR kept low, manufacturing costs are reduced due to buying commonly available parts in greater quantities. After all, there aren't many applications that call for HUGE caps, let alone at the voltage that high powered SS amps run at. Sean
I wasn't aware that brands had to be 'unusual' or 'old' to use only one large-value filter cap per channel in the power supply. Amps I've owned from Classe, Conrad-Johnson, and VTL have done just that. Some even sound(ed) good. The McCormack DNA series, which I just bought one of, takes the 'modern' approach, hence its Distributed Node Amplifier moniker (the smaller caps are 'distributed' near their respective output devices, not clustered near the power transformer). I've wondered though, with this topology, do the power MOSFETS wind up foregoing any potential source impedance benefits arising from filter capacitor paralleling? Or is the power supply actually conventional and the 'distribution' just cosmetic?
First of all, the high-voltage filter capacitors have a certain size and in most cases even a low-capacitance/high voltage filter caps are quite large. It will probably require much larger space to mount 10 100uf/300V caps than to mount 2 x 500uf/300V + the price of first ones is technically much higher than in second case.
Higher collector voltage for SS or higher plate B+ voltage for tube amps dictates higher capacitance for filtering larger pulses and certainly higher energy storage to supply the output stage during peaks.
IME it depends, EG.
if we are retrofitting a clasic tube amp, we may have to go for only as many good filter caps in the original spot,s unless there is a lot of room elsewhere, but wa can often raise the value uf wise while staying near the original dimensions, if the rectifier/s are up to it, (rectifier tubes generally aren't), and if we have a choke, or can fit one in.
you CAN get most of the ESR benefits of paralleling up small caps, with some modern large electros at least, AND their is always bypassing.
All the filter caps in my 2 modded baby Leak St20's, are bypassed with ladders of MF polyprop and polystyrene caps.
Boy, am i glad that this thread came up again. I re-read my post above and it is backwards. Gs had it right i.e. bigger caps have a lower ESR. I can remember posting this late at night, which is when i do most of my posting. I must have been very tired to confuse the issue / reverse the situation. I apologize for posting incorrect info and hope that this hasn't caused any inconvenience for anyone. The other comments pertaining to cost and manufacturing considerations are still valid, even if i did screw up the technical stuff : ) Sean
PS... As i tuck my tail between my legs, hang my head and wander away slowly.....