Reforming old electrolytics

[thermion]

Got a box full of old electrolytics off my house clearance pal the other day, mostly new old stock sprague powerlytics 400v working, 590, 790 & 1200 mfd.
big blue things the size of a coke can. If im reading the date code right they were made in 1977.
Re-formed in April 1983, & been in storage ever since. They are taking a couple of days each to reform on my current limited, 2m/A max reformer. They seem to be settling to a milliamp or so at just over working voltage. Would you consider this normal? My plan is to use them in series on valve amp power supplies downstream of the choke so all they’ll see is half line voltage plus surge at switch on, say 660v across the pair.

[dave_n_t]

In a word - yes (completely to be expected).


To amplify: I've reformed many, many electrolytics over the years and 1mA leakage on caps this large is actually very good going. (Most formulas I've seen involve the product of the capacitance and the working voltage, together with a 'constant' which is working voltage related, and an offset of anything up to 0.5mA).




dave

[Leon Crampin]

I would consider it to be good practice to fit balancing resistors when using electrolytics in series on a high voltage supply. Equal value resistors (observe voltage rating as well as wattage) fitted across each capacitor will minimise he possibility of one capacitor in a series pair suffering excessive voltage if its partner suffers excessive leakage. The ohmic value of these resistors is a compromise: the higher the value, the less dissipation and worthless loading on the supply, whilst lower values are more effective in balancing the voltage across each capacitor. As a rule of thumb, if practical, I arrange for the resistors to pass about 10x the expected leakage current.

Leon.

[Uncle Bulgaria]

My reformer I made from a Silicon Chip kit has a table of acceptable values. I redrew it in an UNO stencil typeface as attached for the front panel. By that measure, your values are very good.

[m0cemdave]

Originally Posted by Uncle Bulgaria

My reformer I made from a Silicon Chip kit...

That looks neat. How did you package / house your former? And what did you use for a display? (it may be worth opening a new topic)

[Uncle Bulgaria]

Thanks. I have a thread on it from some years ago when I was making it here. I ought to post some finished pictures, I guess.

[retailer]

For those that are interested the PCB ($AU20) and pre-programmed PIC chip ($AU15) are available and allegedly in stock on the Silicon Chip Australia web site.

[trobbins]

Depending on the load circuitry arrangement, and time-constants, it's not uncommon to see one cap in a series string end its discharge (after power off) going to a negative value. That situation may vary over time. I will typically add a 1N4007 in parallel with a balancing/bleed resistor to avoid any risk of negative biasing.

Modern resistors should last much better than vintage resistors - restoring vintage equipment can show up quite a drift, or failure, of such balancing resistors, and is well worth checking on even though it can be a pain to disconnect one leg of such a resistor.

[Radio Wrangler]

Originally Posted by Leon Crampin

As a rule of thumb, if practical, I arrange for the resistors to pass about 10x the expected leakage current.

In addition to this, you might also want to consider these resistors as safety bleeders, discharging those capacitors to a safe voltage (each and overall) within a reasonable amount of time.

50v is generally considered a safe voltage, and 30 seconds a reasonable discharge time, so your resistors could be aimed to get the whole lot down to 50v total in 30 seconds from whatever scary voltage the lot was at when running. This would seem a reasonable rule of thumb.

Olde Worlde high value resistors used to drift high, often to open circuit, so bleeder resistors could never be considered to be a complete safety scheme with capacitor banks of this level of voltage and stored energy, even today you still ought to probe around with a voltmeter before fingers go near. But having good bleeders gives you a good chance that if you wait a minute before going in with the voltmeter, the readings ought to be reassuring.

1200uF at 400v is 96 joules each ! (want a skull and crossbones emoji !)

[G6Tanuki]

I have used bleeder networks across electrolytics where, because I am cynical about the peak voltage rating of modern 2-watt resistors, I used three series resistors across each 400V of electrolytic.

I wired a LED in series with each three resistor chain too, arranged to be visible through the perspex insulation plate over the smoothing panel. Seeing four red glows when the case was removed acted as a quiet reminder that there were lethal voltages around, and watching the LEDs gradually dim before I started on any interior work on the amplifier was reassuring.

[Craig Sawyers]

You have to be really careful about safety bleeders in a reformer.

I have a Lafayette TE-46. It has all sorts of functions, but it also has a reformer. And it has a 1k 2W safety discharge resistor. But it is (was) a carbon film resistor.

I was reforming a high voltage high value capacitor, and when I "discharged" it I heard a quiet click. I disconnected the capacitor, and put a DVM across the leads - still at 500V! Eek!!

The 1k resistor was no more. Could not cope with the discharge current.

I replaced it with a small wirewound , which is more able to cope.

In hindsight I should have used one of the new-fangled ceramic composition resistors. A 2W one can cope with currents of thousands of amps for 100ms.

Craig

[cmjones01]

One trick I've seen used in posh power supplies with large high voltage electrolytics is to put a little neon relaxation oscillator across the high voltage capacitors. I attach an example from the Tektronix 7912AD power supply which has 2x2000uF in its switch-mode power supply's primary side (and they pack a serious punch, as they made clear by the violence of the explosion of all four chopper transistors when a fault developed!). That way there's a little flashing light whenever there's a hazardous voltage on the capacitors. Not perfect, but a useful reminder.

Chris

[Craig Sawyers]

It'll come as no surprise that I have one of those. It has every shred of technology from the 80's, including a bit slice processors, and circuit cards crammed into everywhere.

It pulls its near to 1GHz bandwidth by writing onto a specialized internal CRT fast (onto a diode array), and then scanning the diode array slowly. Hence the name Transient Digitizer.

It is a hefty beast.

I have the specialised calibration fixtures, and a NOS internal CRT.

Craig

[Bob_Moss]

I use Paul's capacitor reformer, as detailed in the 'Projects' section on this site. It seems to work well for me. I have made it in a compact box using external multimeters to monitor progress.

[Uncle Bulgaria]

Originally Posted by Craig Sawyers

In hindsight I should have used one of the new-fangled ceramic composition resistors. A 2W one can cope with currents of thousands of amps for 100ms.

That's a really good tip, Craig.