Hope everyone had a good Christmas 😉
I have been looking how to calculate balancing resistors for a power supply using 2 capacitors in series. I found this article which seems very comprehensive, and wondered if anyone has any other opinions on the magic equation that might be used without knowing or measuring leakage current on the capacitors in use.
Thanks!
I have been looking how to calculate balancing resistors for a power supply using 2 capacitors in series. I found this article which seems very comprehensive, and wondered if anyone has any other opinions on the magic equation that might be used without knowing or measuring leakage current on the capacitors in use.
Thanks!
I agree with the article's general comments, and can add some.
It can be worth forming up new e-caps and confirming their leakage current, especially when the history or datecode indicates that years have passed since initial forming by the manufacturer.
Some applications such as in test instruments mean that e-caps aren't powered for months, years or even decades. That can mean that initial turn-on can be quite stressful if current limiting is not applied (eg. a variac start), as the leakage current may likely exceed the charging current and take some time to reduce. This process can be somewhat clouded as the actual cap leakage current can be obfuscated by variations in balance resistor values. I would seriously encourage individual reforming for e-caps older than say 20-30 years. Some instrumentation (eg. HP and Fluke) used high quality e-caps in the 1960's and 70's, and reforming large chassis mount cans can still be worth the time to confirm they attain low-leakage at rated voltage, as well as acceptable capacitance and ESR values.
I'd also recommend adding a reverse biased 1N4007 across each balance resistor in a series stacked arrangement, as age can sometimes force a negative voltage across an e-cap in a string when discharging after power is disconnected, which is not going to help e-cap service life.
And remember that each balance resistor should preferably be rated for 110% the rated capacitor voltage (as that is the max allowed voltage across a high voltage e-cap), otherwise the resistor could fail or degrade and hence take out the caps (and then more). So best to confirm the resistor voltage rating by finding and checking its datasheet.
And lastly, don't forget that modern day mains voltage may well vary by 5 to 10% between locations and time of day, so that is an aspect requiring some margin.
It can be worth forming up new e-caps and confirming their leakage current, especially when the history or datecode indicates that years have passed since initial forming by the manufacturer.
Some applications such as in test instruments mean that e-caps aren't powered for months, years or even decades. That can mean that initial turn-on can be quite stressful if current limiting is not applied (eg. a variac start), as the leakage current may likely exceed the charging current and take some time to reduce. This process can be somewhat clouded as the actual cap leakage current can be obfuscated by variations in balance resistor values. I would seriously encourage individual reforming for e-caps older than say 20-30 years. Some instrumentation (eg. HP and Fluke) used high quality e-caps in the 1960's and 70's, and reforming large chassis mount cans can still be worth the time to confirm they attain low-leakage at rated voltage, as well as acceptable capacitance and ESR values.
I'd also recommend adding a reverse biased 1N4007 across each balance resistor in a series stacked arrangement, as age can sometimes force a negative voltage across an e-cap in a string when discharging after power is disconnected, which is not going to help e-cap service life.
And remember that each balance resistor should preferably be rated for 110% the rated capacitor voltage (as that is the max allowed voltage across a high voltage e-cap), otherwise the resistor could fail or degrade and hence take out the caps (and then more). So best to confirm the resistor voltage rating by finding and checking its datasheet.
And lastly, don't forget that modern day mains voltage may well vary by 5 to 10% between locations and time of day, so that is an aspect requiring some margin.
Thanks for the guidance, the caps I am going to be using are from 2004 and have been in storage.
Is there a good link/article on forming the caps?
I do have a autotransformer Variac that I can use and I also have an isolation transformer I can find that I have somewhere too.
Don't worry about leakage in those PP film capacitors.
However, differing uF values will cause the DC voltage across each to vary, according to C1 x V1 = C2 x V2.
Or rearranging, C1 / C2 = V2 / V1
Last edited:
This article suggests leakage current with PP Film is very low.
PP Caps, is the reforming only applied to electrolytic capacitors maybe?
Last edited:
Yes, only electrolytic caps require reforming as a consequence of inactivity, or a long duration at a lower voltage.
My other hobby is restoring large vintage computers at a museum. The soup can sized capacitors from the 1970s and 80s still fare pretty well when doing a leakage test. Some machines have dozens of them and I test each one individually for leakage before applying power. Usually if the leakage is under 2ma after a few minutes I consider the cap good. Most settle down in a few minutes but some from Data General took over a half hour. Old wet tantalum caps get automatically replaced because they leak acid that eats traces and even the leads themselves.
Yes I could imagine that large e-cans for vintage computer systems would have considerable leakage current, but still be fine. Large e-caps can be a pain to replace, especially if you don't have to. I am nearly through a batch of 10x 160uF 450V Spragues dating from 1963 - very consistent capacitance and ESR and so far all get to below 100uA at or very near 450V rating with just hours of reforming for each one as the onerous aspect. These caps are in series pairs with 470k 2W balancing each, and although only nominally have 250V across each, reforming to rated 450V is an indicator of ok internals, and some confidence that the HP Klystron power supply can work for many years/decades to come if called upon.
I rarely come across vintage wet tants, but must admit that every one I tested recently was ok in an early HP5165A Rubidium - and some of those had to sit at oven temperatures continuously - but yes they seem like time-bombs.
I rarely come across vintage wet tants, but must admit that every one I tested recently was ok in an early HP5165A Rubidium - and some of those had to sit at oven temperatures continuously - but yes they seem like time-bombs.
I just looked at the spec sheets of the film caps in my HV amp, and they state that they can endure 130% overvoltage for a minute, once per day.
So, I assume that if I switch them on once a day or so, they can handle a very short overvoltage while they settle to the balance from the resistors.
As noted, leakage is pretty much non-existent.
Jan
To be honest I tend to just bung in whatever decent resistors I have around the 100k to 200k range. By decent I mean 3w metal film rated for 250v or more.I have some 180k blue bodied jobbies which as AFAIK are rated for higher voltage applications but as I've no idea what they are exactly I could be wrong. This cavalier approach hasn't caused any problems so far. That said I fuse my HT/B+ rails after the transformer winding and all my amps that use higher voltage HT (over 400v) have a soft start.
Andy.
Andy.
Thanks,
So the voltages involved are 500v on each and I was thinking of 2W Bradleys that have a sufficient voltage rating it seems as here.
I could consider a soft start at 1000V but I would need some guidance on this.
🙂
That's serious voltage so needs great care. When I built my 1250v PSU I wound my own tfmr so used 5 supplies stacked on top of each other keeping the voltage between caps etc within reason, you only have two but just do the same as with your caps use several resistors in series, say four 100k's.
For the soft start I used a 6N30 delay valve/relay, two big 1k 100w power resistors and an Omron relay to short out the resistors after 25 ish seconds on my last amp which was 500v HT/B+. For a 1kv supply I'd fettle something to drain the caps when power is switched off. I'd also use micro switches on the PSU chassis to cut the mains should someone open the supply up in the future, an anti tamper type thing. I'd also use some protection on the OP valves as well as whacking on a few "Fecking Lethal Voltages Inside Numbnuts - Beware!" stickers. Can't be too careful.
Andy.
For the soft start I used a 6N30 delay valve/relay, two big 1k 100w power resistors and an Omron relay to short out the resistors after 25 ish seconds on my last amp which was 500v HT/B+. For a 1kv supply I'd fettle something to drain the caps when power is switched off. I'd also use micro switches on the PSU chassis to cut the mains should someone open the supply up in the future, an anti tamper type thing. I'd also use some protection on the OP valves as well as whacking on a few "Fecking Lethal Voltages Inside Numbnuts - Beware!" stickers. Can't be too careful.
Andy.
Tonescout, you are being a bit minimal on technical detail and scope. It seems that you are making a 1kV power supply without any schematic, detailed design, or specific part identification (eg. the link to metallized poly caps just says 200-500Vac voltage rating), which makes for a feeling of unease in providing any advise.
There are important safety issues with any B+ supply, but those issues get stretched out even larger with series'd and stacked capacitors and bleeders. If you don't have an experienced "spotter" with some experience and a strong safety ethic, this really isn't something to be tackled lightly. A good overview can be found in mid-1960's Radio Amateur handbooks, and they're surprisingly still pertinent. Insulators, both inherent and builder-added, become significant components.
I do not wish to discourage you, but some safety issues are not at all obvious. If you have an available "spotter", some geezer with experience and a strong safety ethic, lean on the old buzzard for some corralling. Then build it so that nobody can defeat its safety design. Don't do any less than this.
All good fortune,
Chris
I do not wish to discourage you, but some safety issues are not at all obvious. If you have an available "spotter", some geezer with experience and a strong safety ethic, lean on the old buzzard for some corralling. Then build it so that nobody can defeat its safety design. Don't do any less than this.
All good fortune,
Chris
One thing I build into my HV amp is an interlock switch in series with the HV xformer.
These are mounted against the case side and are pressed down (activated) when the cover is put on.
Remove the cover, and the HV gets interrupted.
Has saved me more than once.
Jan
These are mounted against the case side and are pressed down (activated) when the cover is put on.
Remove the cover, and the HV gets interrupted.
Has saved me more than once.
Jan
This is the one I used: https://www.mouser.be/ProductDetail/Omron-Electronics/D2D-2000?qs=esLHqywZQsh8%2BgoKAWATGQ==
If ypou remove the cover and still want the HV on, just pull up the plunger..
Jan
If ypou remove the cover and still want the HV on, just pull up the plunger..
Jan
"One thing I build into my HV amp is an interlock switch in series with the HV xformer." Ah, that's the correct term. That's what I was referring to in post #15, cover off, power is cut. I agree 1kv + voltages needs caution, that's a lot of potential and stored energy. A shock or a belt off a 500v supply hurts, a 1kv belt might put you on your back permanently. That said old radio transmitter engineers are thin on the ground so as long as your careful you should be right.
David at El Paso on Youtube has built a few EHT supplies, his video's are worth a watch, there's also the GEC An approach to Audio Frequency Amplifier Design PDF floating about on the web that's worth a look. It has a few big linear supplies in it.
Lastly a variac, lamp limiter & capacitor discharge doohdah are essential for mucking about with high voltage linear supplies, enabling you to power up your supply slowly and discharge the cap bank.
Andy.
David at El Paso on Youtube has built a few EHT supplies, his video's are worth a watch, there's also the GEC An approach to Audio Frequency Amplifier Design PDF floating about on the web that's worth a look. It has a few big linear supplies in it.
Lastly a variac, lamp limiter & capacitor discharge doohdah are essential for mucking about with high voltage linear supplies, enabling you to power up your supply slowly and discharge the cap bank.
Andy.