Replacing Filtersupply large can capacitors

Status
Not open for further replies.
I've had a large quantity of US made poweramplifiers from the eighties and nineties like Krell, Levinson, Threshold and Rowland.
Although they differ over circuittopology one thing they had in commen: a big powersupply.
My amps had toroids ranging from 1 to 2.5 kVA per channel and large can capacitors from Mallory, Sprague and CDE.
Designers like Nelson Pass all stated that those capacitors would last approx. 20 to 30 years. So Threshold Stasis(1,2 & 3, S/1000, SA/1 S/500) Krell KSA and KMA series, Levinson ML series (ML-2, 3, 9 & 11) and Rowland Model 5 & 7) all or almost are qualified for having a capreplacement. John Soderbergh performs those operations for Threshold amps standardly.
My question is as follows:
With todays capacitortechniques (better dielectrum I suppose) one can buy in the same size (length and diameter) and Voltage a much higher capacitance rating.
I now operate a pair of Threshold SA/1 from 1984 that had a capacitorbank with 4 Mallory's 80 Volt with 29,000 uF capacitance each.
They measure 6 3/4 Inch and have a diameter of 3 Inch.
The substitude I opted for was a CDE 75 Volts (Railvoltage is 60 Volt) 82,000 uF same diameter (so it fits in the clamps) and even a lower height of 5 5/6 Inch.
At the picture you can see that the Mallory is bigger and produced in last week of 1983. The CDE was produced in week 22 in 1998. So I gained less the nine years her ;-). With a variac I reformed the capacitors before I really start to use them in the amps.
I know that the bridges in the amp will be taxed more but are there any other disadvantages or advantages you can come up with.
All owners of the amps I summoned plus all I did not mention (Sumo, SAE, GAS, Bedini's, Belles, McIntosh, Pass Alephs, NRG, Classe etc..) will face this dilemma to stick with the same size and getting a much larger capacitance or just downsize to volume or increase the Voltage.
What is the best practice here not counting availability and cost.
Thanks.
 

Attachments

  • Mallory 29000uF and CDE 82000uF.jpg
    Mallory 29000uF and CDE 82000uF.jpg
    165.7 KB · Views: 1,163
The higher the voltage with the same capacitance you get a lower ESR, so for transient response you get a smaller voltage drop.When you increase the capacitance you get les riple but a larger peak current in the diode bridge and transformer ( shorter time the diode bridge is conducting). Goldmund use in his Mimesis 9 only 2 x 4700uF.
 
The higher the voltage with the same capacitance you get a lower ESR, so for transient response you get a smaller voltage drop.When you increase the capacitance you get les riple but a larger peak current in the diode bridge and transformer ( shorter time the diode bridge is conducting). Goldmund use in his Mimesis 9 only 2 x 4700uF.
What would you prefer and why?
 
What would you prefer and why?

I should prefer the old capacitance , so the diode bridge is not overloaded and the same for the transformer.I do not know the quality of the old American capacitors but there is a big difference between some types, not in capacitance but in ESR.

If you want good quality you can try the 4 pole of Jensen if it is possible to use them . They give a exellent separation between input and output.
I think the 4 pole capacitors give a major upgrade in sound quality.And I think you are looking for a upgrade
 
I should prefer the old capacitance , so the diode bridge is not overloaded and the same for the transformer.I do not know the quality of the old American capacitors but there is a big difference between some types, not in capacitance but in ESR.

If you want good quality you can try the 4 pole of Jensen if it is possible to use them . They give a exellent separation between input and output.
I think the 4 pole capacitors give a major upgrade in sound quality.And I think you are looking for a upgrade

The 4 pole of Jensen is a Snap-in type of capacitor.
I'm was referring to those large can capacitors like Nippon Chemicon (Sprague), Mallory's and Cornell Dublier (CDE). They're outfitted with screw terminals high or low post and mostly a diameter of 3 Inches. Most manufactures use clamps to fit them to the chassis. I found those in all old Threshold (mostly Mallory) Krell (first Mallory the Sprague) Levinson (Sprague) and Rowland (CDE) amps. There still must be thousands out there in those old amps that were to expensive (no one throws away a Threshold S/500 or Levinson ML-3 after 30 years of service) to disregard. So according to all the designers they need at least a capreplacement. Question remains what is wise if you choose to reatin the physical dimensions of the caps: an increase in capacitance or Voltage?
 
If you look at the side of CDE :
Aluminum Electrolytic Capacitors - Capacitors by Dielectric - CDE.com you can see the different lifetime and temperture (and quality )of the different capacitors. I Think every brand have there own list . So I should chose for a high temperture and long lifetime,and not which have the most capacitance.

I have the old goldmund mimesis 9 , also with very old capacitors from Philips, but stil no problem. I agree those old amps were very expensive but very good quality
 
I now operate a pair of Threshold SA/1 from 1984 that had a capacitorbank with 4 Mallory's 80 Volt with 29,000 uF capacitance each.

I know that the bridges in the amp will be taxed more but are there any other disadvantages or advantages you can come up with.

What is the best practice here not counting availability and cost.
Thanks.

You have a capacity of 58,000uF per rail voltage (or is it 2 different capacitor banks, one for each channel?) which should be OK.

Replacing the 29,000 with 82,000 will incur a very large inrush current which can blow your fuses unless your amps have a soft start circuit built in. In addition at this level of capacitance there probably is some resistor for discharging the caps when you switch off the amp. The wattage of this resistor must be increased otherwise it would not last much. Naturally, you must also change the bridge to a high current FAST bridge.

Moreover your choice of reducing the voltage is wrong. Electrolytic capacitors exhibit leakage, and the smaller the voltage rating - the higher the leakage. This effectively will reduce clarity in the bass performance.

My choice would be to replace the caps with slightly bigger ones i.e 40,000 max. (at 100V or more if space allows) and bypass them (i.e connected parallel to them) with 470uf film capacitors to reduce higher frequency interference. Low ESR is mandatory. To the best of my knowledge best smoothing capacitors for audio are ELNA SILMIC II. I would also change the bridge to a fast one of higher amperage.

If you are not sure how to select your bridge, stick at 30,000uF and simply increase capacitor's voltage and bypass them with film capacitors as stated above
 
I'd stop listening to the designers, who have no special ability to predict how specific capacitors age, and test the damn capacitors for value, dissipation factor, leakage and HF ESR. Replace the ones that are starting to degrade (way fewer than you probably think) with at least equal voltage and temperature ratings. Test those too, since it's kinda dumb to install parts that aren't any better than the originals- they often aren't, especially if the size has been reduced. The value should probably be between rated and 2X, but no hard and fast rule.
 
Are you sure they really need replacing? I have sa2's and the big mallory capacitors look like quality, compared to the smaller ones with same ratings from nowadays...
I am used to replace small capacitors, where you can tell they are bad because they are bulged on top, but how to tell here?
Another way to tell is if you have distorted audio on basses, but if that's not there either...
I have had Pioneer Spec and Nikko Alpha amps, which were about 10 years older then the Thresholds and capacitors were fine.
 
You can replace the 29000uf/80V capacitor with 42000uf/90V capacitor.
Tedss.com have the 42000uf/90V capacitors made by Cornell Dubilier.
This cap is 2.5 inches in diameter and 5.3 inches long. Tedss.com is located in usa. Price 66 dollars each.
 
Moreover your choice of reducing the voltage is wrong. Electrolytic capacitors exhibit leakage, and the smaller the voltage rating - the higher the leakage. This effectively will reduce clarity in the bass performance.

My choice would be to replace the caps with slightly bigger ones i.e 40,000 max. (at 100V or more if space allows) and bypass them (i.e connected parallel to them) with 470uf film capacitors to reduce higher frequency interference. Low ESR is mandatory. To the best of my knowledge best smoothing capacitors for audio are ELNA SILMIC II. I would also change the bridge to a fast one of higher amperage.

If you are not sure how to select your bridge, stick at 30,000uF and simply increase capacitor's voltage and bypass them with film capacitors as stated above

The leakage difference is trivial compared to the 100-120Hz charge wave from the transformer and that it is swamped by the parallel drain resistor loss, and by the effects of circuit impedance.

The recommendation for 470uF film capacitor seems strange, it would be extremely large and expensive. Perhaps you meant 4.7uF or 0.47uF, if not both?
 
Be aware that using modern caps, with lower ESR, your inrush current goes up. In an "ideal", "good" design, it wouldn't matter, but historically equipment has been built based on the specs of the components available at the time... and still is, today we don't design based on the idea that capacitors are always a dead short, but the ideal one would be... based upon what your ideal is.
 
Be aware that using modern caps, with lower ESR, your inrush current goes up. In an "ideal", "good" design, it wouldn't matter, but historically equipment has been built based on the specs of the components available at the time... and still is, today we don't design based on the idea that capacitors are always a dead short, but the ideal one would be... based upon what your ideal is.

Okay thanks.
I now have the latest version of Rowland M7's with only one bridge rectifier (CSB252, 25A) bolted against one of the cooling assemblies.
It has to feed six (probably CDE's) 22,000uF/100V large can caps now over 20 years old.
Jeff Rowland's vision about this is: 'If it ain't broke don't fix it'.
Another diy'r showed me his M7's also the latest version with ten of the twelve (M7 are two mono's) leaking after probably a long period if nonuse and then switching them on without reforming the caps.
He installs RIFA PEH169 type caps now. I thought why waiting till it happens to me although I use them on a daily bases.
 
Last edited:
Status
Not open for further replies.