Advantage of multiple lower value capacitors in power supply

[bluegrass]

Hi, I am building a Quad 405 type clone, helped by a very kind member of this community. He sent me all the boards needed including those for a twin power supply.(Two separate boards which each accept 20 x 2700uf caps...40 in total ) I was just wondering what advantage this arrangement would have over 2 larger valued capacitors in origional Quad 405s . Thanks

[lineup]

One advantage is at switch ON
there are many wires to load all the caps
each cap is looaded with a smaller current
Having only one and load that cap can produce hurtful heat in the cap

[quikie22]

lower ESR, lower ESL

[mnturner]

ofter lower cost

[Chris Daly]

The audio synthesis ben duncan APS02 used this technique to very good effect, if i recall a bank of 10 caps with 4x 1 ohm resistors parallel which act as a fuse, followed by a bank of 5 caps 15 caps for each polarity 30 caps in all . The real issue with the 405 is not so much its initial power supply capacitance which is standard 10,000 uf but the current starving of the opamp supply rails. 3300/50 into a 15v zener supplies less than 1ma factoring in that the zener also consumes power. You will gain lots of improvement from using a pass transistor to each polarity on each opamp supply.

With the 405 Dual rail supply Quad just followed the rule of thumb rule for every amp of current a 1000uf would be required ie 5 amps on each rail on each channel, in fact historically the 405 name arrives from just this the amplifier could deliver 40v AC audio level at 5 amps. Quads use of the rule of thumb also protect the diode bridge.

I agree with your use of multiple caps but not so many I would use 4x2700 uf =5400uf per rail per channel ie 10,800 uf per channel 21,600 uf total as you risk damaging the bridge rectifier on turn on . Pay attention to the opamp supplies this is where you get the best return in audio performance.

Cheers / Chris

[bluegrass]

Thanks for the input. The 2x Quad type main boards are interesting as they use 6 x MJL21194 NPN and 2 x MJE15031 PNP output transistors per board. Im not so experienced with amp building, so don't know whether this would explain the larger number of caps in the power supply boards. I just know it will be a much more powerful amp than standard. I will be using op604 op-amps.

[Chris Daly]

If it has more overall power than the standard 100w per channel then you should use 1000uf for every amp of additional current for each polarity and each channel that the circuit requires. Have you a schematic or link to study this 405 amp clone?

The opa604 consumes 5.3ma minimum !! so if it already has pass transistors this is fine for the opamps, if not adapt the board to use pass transistors for the opamp supplies
Study the board leading to Pin 4 and Pin 7 of the opa604 if it has just a resistor and zener diode leading back to the supply respective rails, you will need to adapt it. If on the other hand you spot a TO92 transistor connecting to Pin 7 and another TO92 connecting to Pin 4 then the board has pass transistors.

50v in down to 15v is most easily done with TO92 transistors. Most designers at some point confront whether to use LM317 and 337 or 78XX regulators here, but the voltage in to voltage out differential is border line even for LM317 and LM337 and more easily done with a TO92 current setting resistor and zener diode with a small cap for decoupling.
Cheers / Chris

[SY]

Keep in mind trace resistance and inductance. This can often negate any supposed advantages of splitting up the cap- it's often easy to inadvertently cause ringing in the supply as well.

Never forget that these are physical objects that take up space, not abstractions.

[indianajo]

Originally Posted by mnturner

ofter lower cost

I'm recapping a 1994 amp that I would hate to have the power transformer burn out on. It has 4 ea 10000 uf 100 VDC illinois caps. Replacements that are 70 mm tall or less (height restriction) seem to be $22 ea for $88 total. Where as I can get 4700 uf @100VDC for $4.86 each, or $40 for 8. I've built a little board out of lexan and steel angle brackets, and am bussing the caps up with 18 ga wire to the original holes. I hope SY's warning about ringing does not come true - stay tuned for test results.
I see the designer of this Quad 405 kit also follows the cheapskate rule on output transistors.My amp burned or exploded 10 MJ15024 and 6 MJ15025 under previous ownership. MJW21194 NPN are $1.93 each, but there don't seem to be similar dealer bargains on PNP complements. I'll have to look up that MJ15031 trick, that PN didn't come up in my search. Best I found was MJ21195G for $3.73. Oh well, real TO3 metric saved me 5 min each on install cutting back mica washers and inserting plastic rings etc to get TO3P in a TO3 hole. Oh, no "rubber stamp" ebay bargain transistors need apply, I like dealer stock from real vendors.

[LineSource]

Hi bluegrass,

I have replaced large power supply capacitors in a few amps and I was glad I could just unscrew the connections for both leakage testing(1K ohm resistor in series to test voltage drop from leakage) and replacement.

The larger screw-connect capacitors are available rated at 105C and >100,000 hours MTBF reliablity. I believe very few small solder-in capacitors meet this level of durabiliby and longevity.

I typically use 4-8 large capacitors for my DIY amplifier power supply. "A few" capacitors in parallel adequately divides the surge current and lowers ESR.


If you put 20 small capacitors in parallel, the one with the worst statistical reliability will set your amp's MTBF.