"Old" electrolytics

[joCCe]

I'm building a rocket and I'm currently working out the layout for the psu. I have a handful of Sprague Atoms 47uf@500V that I plan to use.

I checked the date codes on them and they are from 2001. Would you use them or get some fresh ones?

[martin manning]

They are probably getting a bit old. You could try to reform them, which is easy to do when you get the amp built by temporarily putting a large resistor, say 100k 2W, between the rectifier and the rest of the power supply. Turn the power and standby on and let it idle like that until the DC voltage across the resistor stabilizes. If you do your layout for those caps, there will likely be plenty of room for replacements of any brand.

[joCCe]

Okay, thanks very much.

I'll give that a go!

[surfsup]

martin, interesting tip thanks. Just to be clear I assume there should be no power/preamp tubes in the amp so as not to draw tons of current through that 100k resistor.

[martin manning]

It probably doesn't make much difference if the tubes are in or not; the goal is to limit the current when the caps are brought up for the first time. As long as all of the caps are good for the full b+ then there is no over-voltage issue with tubes out. This may take a few minutes or a few hours depending upon the condition of the caps, so unless the tubes need burning in, it might make more sense to remove them.

[Structo]

A member named Larry put out some good instructions for forming caps.
He is from Germany so pardon the English.
This is for a 100w Marshall amp but the 100K resistor is the same and where to put it is the same.
The voltage will slowly go down as the caps form.
Once the voltage hasn't gone down any further the caps are reformed.

Quote:
Forming filter caps...

... isn't also a bad things for new caps and enhances their physically parameters, also when they've been (quick-) formed at the manufacturer.

But for new caps, which have been stocked on the shelf for more than 5 years it's obligation!

I'm describing the steps valid for a 100W Marshall, with the 45W or 50W models you can share some of this steps:

- pull out the power tubes and the preamp tubes

- disconnect the yello wire going from the PT between the first two filter caps, temporarily tape the unsoldered end for safety (100W only)

- unsolder each one lead of of the both 56K bleeder resistors across the both screen's filter caps and tape it temporarily for safety (100W only)

- disconnect the red wire from the rectifier diodes (the rectifier tube) to the first filter cap and swap in temporarily a 100K/1W resistor

The Resistor then is located between the rectifier and the first filter caps - temporarily, until forming is done.


Now fire the amp up completely (power & standby to 'on') and measure the voltage drop across the 100K resistor you've swapped in!

You'll notice, that first you'll have a voltage drop of about 200V, which within 15-30 seconds reduces itself to about 30-40V - then further on it is dropping slowly.

If so, all's ok and you can switch off your meter and do your other things, because now the caps require 4-24 hours time to form, depending on the cap's quality and the cap's age.

Only insure, that nobody else could get in contact with your 'open' amp - way the best is, to lock the room, where the amp is sitting and keep the key in your pocket. It's danger to the life for unskilled people!!!

From time to time you can check the process of the voltage drop - and as soon as the voltage drop across the 100K is less than 5V (five volts), your caps are formed.

Switch off the amp - and before you remove the swapped in 100K resistor, implicitly bleed the caps by connecting pin #1 of V1 with an alligator clip to ground for at least 2-3 minutes!

Now reset the temporarily tweaks and put in all the tubes again - you've done Fire up the amp and enjoy its (valorized) sound

Larry

BTW: Also when an amp hasn't been in use for longer than six months, you should form the caps, before you fire up this amp again! If not, the longevity of the installed caps might be significantly diminished!

The Resistor then is located between the rectifier and the first filter caps - temporarily, until forming is done.

[surfsup]

Martin,

It probably doesn't make much difference if the tubes are in or not

Wouldn't having tubes in, pulling 100mA current (assume 2xel84 amp), fry the 100k resistor pretty quickly (1000W dissipation)? I was thinking without tubes, much less current the resistor would survive.

[martin manning]

Martin,

It probably doesn't make much difference if the tubes are in or not

Wouldn't having tubes in, pulling 100mA current (assume 2xel84 amp), fry the 100k resistor pretty quickly (1000W dissipation)? I was thinking without tubes, much less current the resistor would survive.

No, the tail can't wag the dog, so to speak. If you put a 100k straight across 400V it will only have to dissipate 1.6W.

[surfsup]

Ok I just learned something thank you. Solving for current instead of power...I was using 100mA assuming the amp would pull that but it doesn't work that way.

400V*400V/100000 = 1.6

and I should have known this since I have been putting resistors across my standby switch.

[VacuumVoodoo]

Whoever suggested this cap forming procedure probably skimped a little on his homework.
Also, forming process at the factory is not a quick one. Cap forming is an exothermic electrochemical reaction, if it is allowed to run too fast the caps will simply heat up and blow up. They are formed in big batches so there could be a lot of heat. It takes the time it needs to take. Forming new capacitors that are younger than their shelf life (per data sheet) is a waste of time and energy.

Running just a 100k resistor in series with the cap and applying full working voltage at once can cause unwanted initial heat build up.
One should use a variable voltage, either from a variac or from your VVR aka Power Scaling. Start with 10% of caps nominal working voltage and increase in 10% steps in until you reach 105% of caps working voltage.
Have a DMM connected across the 100k resistor and monitor its voltage, this tells you magnitude of charging current (V/R). Increase voltage to next step when charging current has become constant. It can be after 15 minutes or couple hours or never (if that's the case just dispose of the cap in environmentally friendly manner).
The process is complete when at 105% WV the current in the resistor is constant and not higher than what the caps data sheet says the leakage current is allowed to be.
Of course this all done without any external load on the cap like tubes etc.

[martin manning]

I don't think this procedure is intended to be at all rigorous. It's a simple and reasonable way to go about powering up with potentially stale caps, and certainly much better than just hitting them with the full B+ voltage. Consider that if a 400V supply is connected through a 100k resistor it will pass at most 4mA (DC; a little more transiently at start-up), and this will be distributed over all the filter caps in the amp. I can't imagine that this would be excessive.

To get an accurate reading of leakage currrent one would have to remove all loads including tubes and all bleeder and balance resistors, and then since multiple caps are connected in parallel, there would be no way of knowing which one(s) have excessive leakage current if that is indicated. You'd know that there is a problem, but more testing would be required to isolate it.

[rp]

When powering up amps that have been parked for a long time I use the Gerald Weber way (I know, I know) - start at 40V on a Variac for a couple hrs and go up 10V per hour. Beats me, but seems reasonable. The other I'm thinking is to just use your DBL, you could start with a 15W for a while then a 40W, then a 60W then a 100W. Should work, no?

[VacuumVoodoo]

I described forming process in response to OP i.e. old caps not yet installed in the circuit.
For a complete old amp the Variac method as described by rp works. A small hand held IR thermometer is helpful to identify a cap going south. Check temp on the casing of each.

[martin manning]

Good tip on using a thermometer to spot bad caps, thanks Aleksander.

rp, I guess using the variac or bulb limiter on an amp with a vacuum rectifier you have to get the voltage up high enough to get current flowing to the caps. The advantage to putting a series resistor between the rectifier (ss or vacuum) and the filters is that you can then measure and look for stablized current draw more directly.

An additional thought on using the temporary 100k series resistor technique is that you could initailly place another 100k as a shunt to ground after the series resistor. This would limit the voltage to essentially half of normal. After running like that for a while, the shunt could be removed (carefully!) to allow the full B+ to be applied. Note that if a bleeder resistor is part of the supply (like the optional 220k in the Express schematic) it will have a similar effect unless the series resistor is installed downstream of it, like in the drawing Tom posted above.

[rp]

BTW this might have been posted already here's some more web info:

Strategies to Repair or Replace Old Electrolytic Capacitors
http://www.nmr.mgh.harvard.edu/~reese/electrolytics/