Increasing Filter Capacitance on the 2270?

I routinely buy these from AK sponsor "Parts Connexion" up in Canada. Good folks to work with and the shipping cost isn't much more and the time to get here either. At least compared to Digikey deliveries to the southwest!
 
Thanks. I just happen to have a Parts Connexion not too far from where I live. I could place the order online and pickup on my way to work one day.
 
I have always had luck getting those caps from Hifi collective over int eh UK. A bit more for shipping but I order things in batches so its not so bad. I have a set of Nichicon KG 10Ks sitting on my desk right now waiting for an amp I'm custom designing for myself.
 
A change from 9000µF to 10000µF isn't going to cause any ill effects that I know of.
Many electrolytic capacitors in the 'old days' had a tolerance of +100/-20%, a 9000µF capacitor might have been as much as 18000µF, new.

Tom
This is interesting and informative for me. Recapping old gear I noticed that actually those 40 year old caps had higher capacitance than stated, sometimes still 50-100% higher. But generally not big filter caps, many 10000uF I replaced were 6500-9000uF, while smaller caps were above their specs, 1uF were 2.5uF sometimes, 2.2uF were 2.8uF, 330uF from power supply of 2270 were really around 400uF, most caps in my 2270 were 20% and more above stated markings. I think the only one below markings was the one 470uF/63V that was 390uF, so still within today's -/+20% allowance. New caps are usually around -10%.
 
Last edited:
Well good thing I'm in no big hurry for them. They just emailed me a while ago to say they are on back order. Waiting to find out how long before they expect them. I'm fortunate in regards to their location because it's not far out of my way to work. I've placed an order there before and had a buddy I work with , who lives around the corner from them, pick up my order for me and bring it to work.
 
Well good thing I'm in no big hurry for them. They just emailed me a while ago to say they are on back order. Waiting to find out how long before they expect them. I'm fortunate in regards to their location because it's not far out of my way to work. I've placed an order there before and had a buddy I work with , who lives around the corner from them, pick up my order for me and bring it to work.
Do you guys think they are better than GP 10000uF caps or is it just about the solder lugs? I am a bit put off by lifespan of those audio caps. I don't want them to go quickly. I measured them, they had excellent ESR (which surprised me for audio caps) but my ChemiCons are even better in this department and they are 80V on top of this :) , price is roughly the same as I remember.
 
Parts Connexion currently has a 25%off all caps sale going on till the end of April.

(I just saved $5000 on a $20000 Silmic II order.)

Bob
 
The 25% off is one of the reasons I ordered them now as well. Wow that's a substantial savings when ordering in bulk like that.
 
Here is what I'd do before going nuts on capacitors. Put a scope on the capacitor and crank hard on the amp, look for ripple and dc level sag try a few different frequencies. Do the same on the power amp where that line goes onto the board and compare the results. If there is enough capacitance to supply the current demand, a larger value should not improve anything. I've played around with it and truly huge changes can have almost no effect, you are fighting physics. Once you have enough capacitace its all about lowering the impedance, get that power to the user. Also if you go overboard with huge capacitace it can cause problems. The transformer is the most expensive power supply component and I'd say it is likely the limiting component in that system (maybe not on one of these marantz units they have huge transformers). Also I see people increase the size of the main filters but not the smaller ones on the power amp boards, the point of the cap is to act as a low impedance source of current. The less wire length (and heavier guage) the lower you can get the impedance, so you can have a lower impedance source of surge current right on the amp board (as frequency raises this becomes more important, being physically close). The wires will act as inductors and resistors, the current damand is transient. This will create a voltage drop under load hence a source close to the amp. I would not be surprised if heavier wire and better connections would have a larger effect than hot rod caps assuming the capacitors in place are healthy and not too small. However as stated before I dont think modification will make it sound better if there is no problem. I'm working on a 2275 lately perhaps I'll do some experimentation. See if I can measurably lower the THD with some power supply "improvements" and if i can put my money where my mouth is. Ill take improvements out if quotes if I can prove them.
 
So I did a little testing today. On this 2275 I've got here running one channel at 25 volts on an 8 ohm load.

First unloaded
B+ at main cap, 47.7VDC and 25mVAC of 120hz ripple.
B+ at amp board 47.7VDC and 26mVACof 120hz ripple.

Loaded 25 volts RMS on 8 ohms,1 kHz.

B+ at main cap 43.71VDC and 217 mVAC
B+ at amp board 43.7VDC and 220 mVAC.

The drop and ripple across the wire was lower than I had expected but I did note that the measured frequency at the amp board was unstable, not locked on 120hz. Need to look at that with the spectrum analyzer, see if i can see the 1 Khz.

Next I added 15000uF to the 13000uF already there so 28000uF. About 2.1 times stock. Unloaded it hardly made any differance and loaded my ripple is down two about 140mVAC. About a 1/3 reduction in ripple. Id like to see the effect at high frequency, might try that next. I want to see if the V drop from the load can be seen getting higher as you approach the load.

I did not test THD yet because I was only looking at B+ so very well could have mode it worse by causing an offset. Might try a bank of many smaller caps to see if lower ESR can kill some more ripple without going insane with huge capacitace, I doubt it but why not. That and cleaning the tarnished of the connections on the amp board.

Sorry to hijack the thread. If I continue with this I will start my own thread In DIY.
 
There are small film caps (C717,C718) on each amp board situated from the power supply rails to ground. These provide a easy path for high frequency trash to ground.
Remember, the power supply rails are at a high DC potential, but are at AC ground.

Tom
 
So if those poly caps work well enough and are sufficient in size not much to gain there, I did not look at the drawing before, the whole point was to just see it for my self. No more storage capacitors untill after a resistor, this makes sense I guess, the wire is not that long and not much impedance to AF. However if the 2270 is anywhere near the power of the 2275 and it has 6800uF caps I bet the ripple is much worse under heavy load, looks like its rated at 70 watts and runs the same voltage. I don't want to solder all over this amp but I'd like to see where the sweet spot is with capacitace (looking at distortion, not just ripple, technicaly you don't care about the ripple if its not causing the output signal to degrade why chase it down the rabbit hole). VS the cost it does make sense (to me anyway) to go up a little if you are replacing them but how much? Generally I just price the original value and the next one or two standard values above it and select depending on cost. Or max the tolerance if they list it somewhere. Here is an idea, dont pitch the original filter caps if they are large value (expensive enough you don't want to replace them as preventative maintenance), still strong and balanced. Just parallel new ones of some value to reach whatever that sweet spot might be.
 
RM, I believe the 2270 comes stock with 9000µF filters, 10000µF is a common value, so I use those. I think the 2275 comes with 13000µF so I generally use 15000µF for those. I try to match the diameter of the original ones so that I can use the original mounting hardware.
I think it's great that you are trying to find the weak points and make things work better.
I have spent a lot of time looking at the output of power amps using an audio spectrum analyzer with THD+N measurement capabilities. When I see THD that increases dramatically as frequency increases, it is usually due to failed (or missing!) power supply bypass capacitors. On some models, the manufacturer even eliminated some or all of them to save a few pennies. I put them back in the way they were before the bean counters got involved.

Tom
 
20170422_190007.jpg
13000uF for B+ and B-, about 25v @ 1k on 8 ohms. Big meter is on B+ and B-.

20170422_185905.jpg
Same but 28000uF on both rails. THD was 0.012 on both with the keithley. Inconclusive because thats down in the floor on that thing I think. Go figure. Ill try at higher frequency I guess. Need to finish the hp 3580.
 
0.035% THD on both normal and huge cap at 10 khz. On both the meter frequency count was bouncing around so I bet the meter is seeing noise from the 10khz on the supply. Ill try 20khz and then fancy bypass caps. Unless my test equipment just sucks its looking like this all original 2275 is not lacking in the power supply department... doh! I should load both sides. I've been loading just one side. Ok I think I'll start a new thread. Good luck with the 2270!
 
Pat, What is a good source for these Nichicon KGs? All the KGs at Mouser seem to be snap in. Also the only 10,000 mfd they have are 50V or less, so not much use in the 2270 (OK for the 2245).

Also, as I'm sure you know, there are 3 types of the KG with the type 3 being the best. Can't find those at Mouser either...

So a source recommendation would be appreciated.

The Nichicon KG (Gold Tune) solder tab caps are usually in stock at Michael Percy. That's where I got mine. Here...http://www.percyaudio.com/
 
Back
Top Bottom