capacitor reforming

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

daArry

Well-known member
Joined
Jun 3, 2004
Messages
722
Location
londaaan
Hey,

So I need to reform the electro's for the psu caps on me la2a project and came across this PDF: http://www.vmarsmanuals.co.uk/newsletter_articles/big_bang.pdf

I've redrawn the ckt for a 6pos switch:

cap-reform.JPG


Thing is, I can't find any 16uF/350v caps - any ideas?

Cheers :thumb:
 
Hi daArry,

I wouldn't worry too much about the values- 16uF is an "old value" cap- (this betrays the roots of this article- 1969!) 10uF or 22uF would be the nearest preferred value today. Make sure you run this circuit from the secondary of a mains transformer- it'd be a lethal device run straight off the mains :shock:

The original article states that you want about 25mA flowing through the potential divider chain- you've got 5 x 2k2. This means a total R of 11k. Therefore I = V/R, I = 500/11000 = 45mA. If you made those resistors 3k3 or 4k7 you may be better off- and they could be lower wattage. Those 2k2's would have to be 5W or greater- they're each dissipating 4.5W in this circuit! You may even get away with 10k for each step. This would reduce the current in the chain to 10mA, dissipation in each R would be 0.22W, so 1 or 2W types would do. And it would be less of an electric heater!

The easier option is to get to know someone who has a variac, and connect it to the mains input of your HT supply. You can get 0.5A variacs surplus for £30 or so if you hunt around.

Mark
 
Hey Mark,

Cool, 10/22uF 350v's are much easier to locate ;)...

Make sure you run this circuit from the secondary of a mains transformer- it'd be a lethal device run straight off the mains

ehehe - no doubt! I was planning on using the la2a power transformer (hammond 369JX) for this reforming task...it has 250VAC secondary so should be ok no?

You may even get away with 10k for each step. This would reduce the current in the chain to 10mA, dissipation in each R would be 0.22W, so 1 or 2W types would do. And it would be less of an electric heater!

Thanks for the info there. I've got some 10k/2w'ers 2hand so'll give them ago...

The easier option is to get to know someone who has a variac, and connect it to the mains input of your HT supply. You can get 0.5A variacs surplus for £30 or so if you hunt around.

The HT is the heater supply right? (new to valves ere soz) I thought variacs just connect straight to mains power (new to variacs too! :roll:). I don't know any local bodz into theyre diy so finding someone with one is unlikey - I thought with this ckt tho it'd be easy to do for an amature like me - s'got an indicator bulb n everything! :grin:

cheers,
 
Yep,

Fine with 250VAC secondary- with a circuit like this you have to have "galvanic" isolation from the mains (i.e. no direct electrical connection- the connection is magnetic from primary>secondary)

The "HT" supply is the "High Tension" supply- which is normally your "B+" in valve gear. The way I reform caps is I use my HT supply (a 250V:250V transformer in a box with a rectifier and smoothing caps) and connect it's mains input to my variacs mains output.

This way I can bring the HT supply from 0 to around 350V by slowly increasing the control on the variac- it's basically a variac followed by a 1:1 mains transformer, followed by a rectifier and smoothing filter. This means that it's a 0-350V variable DC supply. I have a separate AC and DC supply for heaters- you don't want that varying with your HT!

Your circuit above will work fine, but it's just easier long-term to have a variac on the bench. You just build a standard LA2-A style PSU using that Hammond transformer followed by rect/cap/resistor/cap etc. and you put it in a box. Then you mount the variac in another box with a mains plug input and mains socket output. You can use the HT supply "standalone" by just plugging it into the wall, and use that as a 350V DC supply, or by plugging it into the Variac's mains output, you've got a variable HT DC supply.

The variac soon pays for itself as you can use it when 110V equipment comes in for testing, and even for testing Vxx stuff which needs 220V. I don't reccomend variacs for long-term voltage adjustements for installed gear- too easy for the knob to be knocked accidentally, and gear will fry.

When old valve gear comes in you can use the variac to bring it up slowly to full mains voltage, after doing basic electrical safety tests, to make sure that any electrolytic capacitors have a chance to reform. So many people buy something "old and groovy" off evilbay, and then plug it directly into the mains. That can cause a sticky mess as the electrolytics in the PSU get hot and decide to take themselves out!

One thing to remember though is that a variac is NOT and ISOLATION TRANSFORMER and will NOT give galvanic isolation from the live side of the mains supply!! A variac is an autotransformer (i.e. a single winding with a sliding "tap") An isolation transformer has two electrically isolated windings, like a traditional power transformer.

For experimenting, you can use a 12V AC or DC "wall wart" for the heater voltage. Most ECCxx's etc can be run with series heaters at 12V AC or DC, depending on your requirements.

Let us know how this reformer goes.

...never mind caps, I've known people who could've done with a few hours hooked up to a "reformer"!

:wink:

Mark
 
I like to heat the caps up to about 80C before I reform them.

I would place the caps in a car or something that will heat up in the sun.

I believe I read in a old Audio(the audio mag) article from the late 70's early 80's (when adding lots of caps to the power supply of a solid state power amp was a thing to do)
That you could use the supply that you had

What one would do is calculate a current limiting resistor.

Knowing the unloaded B+ voltage calculate a resistor that would limit the max current to 1ma(.001 amp) IIRC. The caps voltage rating will need to be greater than the max B+ voltage

say 330V B+ unloaded 330V/.001A = 330K ohms. Max power would be I^2 x R or V X I 330V X .001A = .33watts

One would hook the current limiting resistor in series to the + of the cap and the cap - to ground. forming a RC netork the cap being the C

The current limit resistor voltage drop will be the most at the start and then drop down to the LEAKAGE current of the cap x the resistor value V=IxR

If the cap is shorted or shorts when forming or hooked up backwards the max current is the unloaded B+ / current limit R

Before anyone uses this I would like to know if people think this is safe. When reforming caps I like to place them behind a blast shield

One could use a ma meter in the resistor leg.
 
It's hard to blow up caps with a milliampere, even if the current times the voltage across were all going into heat. But a safety shield is always a good idea. When caps do explode, even if they aren't doing so in someone's face the mess to clean up is usually prodigious.
 
Thanks again there Mark - detailed info and easy to follow, melikesy :)

I think i'm gonna go with the posted ckt and leave the variac box till i'm more flush - need all available cash to complete some projects and my next flutter on some lab gear will be one of those atlas component testers...

BTW, for those caps C1 and C2 - would 400V rated ones still be okay or does this effect the 'doubled' voltage?

Your circuit above will work fine, but it's just easier long-term to have a variac on the bench. You just build a standard LA2-A style PSU using that Hammond transformer followed by rect/cap/resistor/cap etc. and you put it in a box.

Cool, well this is the board i made for the la2a psu bits:

P3250017.JPG


So i guess hookup would be pretty straightforward with that eh...

Cheers,

D
 
Cool - caps sorted then...

Just a couple more q's about that ckt, sorry. For R16, it says use a pot to set 500V - have you any ideas on a set resistor value I could use? as if i went with a pot I'd need one with a high power rating...they cost quite a bit dont they.

And for the Neon lamp indicated - will the NE2 from the la2a be ok here? Wasnt sure if there was differences from NE and NE2...
 
Hmm,

That's a bit tricky- it depends on lots of variables. The voltage doubler circuit doesn't actually double your rms AC voltage, and it also depends on the load you are feeding, the transformer secondary, and the value of the caps in the doubler.

Probably the best thing to do is to build the circuit from the mains transformer up to the voltage doubler. Then miss out R16, and just connect your voltage divider- i.e. the four resistors R1-R4. Make sure the circuit is off and the caps are well discharged, and connect up a DMM set to its highest DC range (1000V or greater) across the resistor chain, so that you're measuring the DC voltage at the output of the doubler. The resistors are there as a temporary load, to roughly simulate the actual circuit conditions.

[Tip- when measuring a high voltage (above 200V or so) don't use the standard "prods" supplied with the DMM. Make up a set of test leads with the standard red and black 4mm plugs on one end, and then solder insulated crocodile clips (the ones with plastic "boots") to the other end. This way you can clip both leads onto the points/terminals where you want to measure, and then when they are both secure, switch on the power. This way, there is no chance of you slipping or accidentally touching exposed live terminals, or touching a leaky probe. The absolute worst thing to do is to wade into an HT circuit with the two DMM probes, one held in each hand when the power is on :shock: A half measure (and how I usually work) is to have the 0V lead as a croc clip lead connected to the 0V point, and then have the +Ve lead as a standard probe, so that you can at least work "one handed". This is much safer as there is less likelihood of HT across the heart...]

End of safety lecture.

Switch on the circuit, and measure the voltage. Switch off once you've got a stable reading. You can then work out the size of R16 by the following:

1. Add together the values of the 4 resistors R1 to R4. The voltage across these needs to be 500V.

2. Find the current through these resistors when there is 500V across them. I = V/R, so I = 500/R1+R2+R3+R4

3. This current will also flow through R16. So the voltage drop will be V = I * R. You know the current, and you'll now know the voltage you want to drop: Vacross R16 = Vmeasured - 500V

4. So drop these values into R = V/I to find R16:

R = Vacross R16/Ithrough R1+R2+R3+R4

Find the nearest preferred value, and you're sorted.

5. Before you select the resistor, you'll need to find its power rating: P = I^2 * R

Drop I and R into the above, and multiply it by 1.5 to 2 to be on the safe side.

Neon- I have a feelling you need a "naked" neon- i.e. one without a resistor- I think M@plin sell them. The reason is that it's used as a "blinker"- check this link: neon blinkers Neons trigger and glow at ~90V, and when you strap a cap across them and feed a voltage via a resistor, they act as an oscillator (as the cap charges to 90V, the neon flashes, discharges the cap, and the cycle starts again...)

:thumb:

Mark
 
greatstuff - thanks again ;)

re the neon bulb needed - when you say "you need a "naked" neon- i.e. one without a resistor" do you mean just the bulb on its own? the 'NE2' which i asked for and got for me la2a bits is just a bulb - there is no resistor connected to it - you do that in the application ckt if needed right? All i can find from maplin is this: http://www.maplin.co.uk/Module.aspx?ModuleNo=1952&TabID=1&WorldID=&doy=17m7 which looks the same as the NE2 i have for the la2a parts...I've searched for NE1 but had no luck finding one :roll:

I've found this other reformer ckt too:

http://www.vintage-radio.com/projects/capacitor-reformer.shtml

Connected straight to mains tho :shock:

Cheers :thumb:
 
Yep,

You can tell neons from filament bulbs by looking closely inside- a true neon just has two fat rod electrodes inside the bulb- and you can see they're not connected together. A filament bulb (which can look similar to a neon to some Craplin staff :roll: ) has a visible filament connecting the two wires inside.

The NE2 from the LA2-A will work fine.

Now get reforming!

Mark
 
Nearly set to go then - just 1 last q (I promise this time!) before firing up...

The power tx i'm using is the Hammond 369JX -

369JX_LA2A-PSU.JPG


That's what it'll be like in the la2a, but for use in this reformer I'm just wondering if the red/yel centertap on the 250V red secondaries needs connecting to ground in the reformer ckt or not...is it left floating as D2 does the gnd ref?

Ta
 
Yea- just use the two red leads. The voltage doubler needs a floating AC source- as you noticed, the other side of D2 is the actual "ground" or 0V point of the circuit.

:thumb:

Mark
 
Mark,

Soz, but here I go again... I was goin over your a,b,c a few posts back and just to be sure I'm getting the math right - does this scenario look right?

1. Add together the values of the 4 resistors R1 to R4. The voltage across these needs to be 500V.

We're goin with 10K for R1 to R4 so thats 40K total...

2. Find the current through these resistors when there is 500V across them. I = V/R, so I = 500/R1+R2+R3+R4

500/40000 = 0.0125

3. This current will also flow through R16. So the voltage drop will be V = I * R. You know the current, and you'll now know the voltage you want to drop: Vacross R16 = Vmeasured - 500V

Say for this example we measure 540V so, 540-500 = 40V

4. So drop these values into R = V/I to find R16: R = Vacross R16/Ithrough R1+R2+R3+R4

40/0.0125 = 3200 = 3K2

5. Before you select the resistor, you'll need to find its power rating: P = I^2 * R Drop I and R into the above, and multiply it by 1.5 to 2 to be on the safe side.

1W (give or take)

Cheers..again...
 
groovy ;)

well, i just tried the ckt with 10uF/450v caps and 4x4k7/2W (for r1-r4) stuffed, measured voltage from point M and ground and it was of the scale! like 1088V on the dmm - its max limit is 1000v so I dunno if it was more...it beeped at me anyway... :shock:

EDIT: was connecting the 250/250 windings so the input was seeing ca500VAC - worked out i need just one of the red 250 windings and the red/yel 0v centertap...
 
Back
Top