Capacitance Box

The regulars around here know I'm working on the power supply article for the book. I've gotten some help from some of the forum members as to possible power supply capacitor formulas, for which I thank everyone involved. Being a hands on type (and a little weak in some math) I'm going to build a capacitance box that I can use with several transformers to test the prediction ability of these formulas.

Before I start drilling the 7" X 5" box I bought from Radio Shack I'm putting it out to see if any other ideas crop up. Some of the capacitor values were dictated by what I could get, not what I would like.

I decided against using 10µF, 22µF, and 47µF because they were simply too small to be very practical. This is meant to test filtering for a bulk power supply after all.

These caps are all rated for 50VDC.



Any ideas for improvements?

Where is the "discharge" switch? (might want a power resistor or few in series with it)
Where is the fuseholder?

I hadn't even thought of a fuse. Interesting idea, though I probably won't do it. While there will be some surges in this box (as I intend to play with it) most times it will be pretty sedate.

The discharge shouldn't be a problem, since I'll be testing these ideas with loads. I bought a bunch of 160Ω 2W resistors to go with the 12.6V transformers for loads. Since this isn't continous use I am running these parts to their max specs.

I have room, I may add another 33000µF, 3300µ, and 330µF for a wider range.

Bill_Marsden said:

I hadn't even thought of a fuse. Interesting idea, though I probably won't do it. While there will be some surges in this box (as I intend to play with it) most times it will be pretty sedate.

The discharge shouldn't be a problem, since I'll be testing these ideas with loads. I bought a bunch of 160Ω 2W resistors to go with the 12.6V transformers for loads. Since this isn't continous use I am running these parts to their max specs.

I have room, I may add another 33000µF, 3300µ, and 330µF for a wider range.

Click to expand...

Even with a load, when you switch a capacitor out, it will remain charged. If you later switch it in while connected to another circuit, it may damage that circuit.

I've got one of these Heathkit condenser substitution boxes and a resistance box too. Never find much use for the condenser box but I thought you might like a look. Both flea market finds.

jgessling said:

I've got one of these Heathkit condenser substitution boxes and a resistance box too. Never find much use for the condenser box but I thought you might like a look. Both flea market finds.

Click to expand...

I have one of those too, from a similar source.

Ron H said:

Even with a load, when you switch a capacitor out, it will remain charged. If you later switch it in while connected to another circuit, it may damage that circuit.

Click to expand...

I'm aware of this, but the thing is, it will be connected to loads. Plus, I like seeing sparks. This circuit has binding posts, easy enough to add a wire to.

The switches I'm going to use have a 6A spec, so I'll add a 6A slow blow fuse. Thanks for the idea Wookie.

Bill,
You might consider doing something like this for a discharge circuit:



I've shown a TL783 being used, which is basically a high voltage version of an LM317. That would give you a range up to around 125v; with an LM317 you'd be pushing your luck at 40v.

The circuit will sink about 50mA constant from the cap when the cap is de-selected; the LED will remain lit until the cap voltage gets down to around 5v.

[eta]
You'd need to repeat this circuit for each cap used.

Bill_Marsden said:

I have one of those too, from a similar source.



I'm aware of this, but the thing is, it will be connected to loads. Plus, I like seeing sparks. This circuit has binding posts, easy enough to add a wire to.

The switches I'm going to use have a 6A spec, so I'll add a 6A slow blow fuse. Thanks for the idea Wookie.

Click to expand...

It's your project.
I was imagining switching various capacitors in and out of the circuit. The ones switched out will not get discharged when you remove the box, no matter where or what the load is.
Enjoy your sparks.

If I could find a 470000µF I could make 0.11 F, which offers ideas for energy storage. Or I could get another 330000µF and settle for 0.096F. I wonder how hard it will be to blow the fuse with this setup, charged over 20V. I'll let ya know.

OK, if I add discharge capacitors this is how I would do it...



If I used a 2W resistor it would be 1.2KΩ. This would be good for the X10000µF scale.

I like it.

Couple of comments.

1)
You might like to add one or more LEDs to tell when the caps have charge.

2)
Many power supply circuits use one terminal of the capacitor as the earth star point. You might like to add some extra terminals to accomodate this.

3)
I often add extra sockets / terminals as instrument connection points anyway.

I've already exceeded the complexity I was originally thinking about. The LEDs are a bit over the top (and I like LEDs). Currently I'm looking at a 330 second RC time constant for 10KΩ, which is why I'm thinking larger wattages. For most cases a long discharge time is OK though. My current thought is to parallel ¼W resistors for the desired wattage, so no one part gets that hot. I'm figuring worst case, 50V, for wattages.

Binding posts are a star point, they screw up and down and have a hole in the side. This isn't counting the banana jacks of course.

Hard to say without adding it all up, but it looks like it is costing around $30, give or take.

Why don't you use some 7-1/2W Xmas-type 120vac bulbs for discharge? The caps would be discharged down to nothing really quickly. Incandescent lamps aren't exactly constant current sinks, but their non-linear resistance for hot vs cold will help to drain them quickly.

Or, use some other small bulbs rated for 120V.

I'll look into the light bulbs. I lean towards resistors due to reliability, and I suspect 48VDC bulbs are going to be a challenge. I could use a string, though I'd not a fan of the idea.

I currently plan on using 24 gauge solid wire for the innards.

Could you wire a large value resistor across each cap that would not have a significant effect on the circuit but would discharge the cap, although slowly.

Well, if he had permanent bleeder resistors on the caps, they would interfere with the readings if they were large enough, but take forever to discharge the caps if they were small enough to not interfere with the readings.

Bill,
I think AWG 24 is a bit undersized. I'd use at least AWG 22.

I had an epiphany while shopping at Tanner's for more parts. The problem is not wattage, but current, since the switches are rated for 6A max. If the caps were charged with 50VDC a 10Ω 3W resistor would limit the current to 5A. The RC time constant for the 33000µF cap would be .3 seconds, so the 250W surge is too short to make much difference, I hope.

Opinions?

Always go for overkill - http://knifeswitch.com/

To be more serious, I would use heavier wire - 16 ga lamp cord is still reasonably cheap - and a switch rated for more current. You don't have to worry about contact errosion that way.

Any resistor will drain the caps over time. You can have two in there, one to do a slow discharge and another on a pushbutton to do it rapidly if you are doing experiments.

I've already bought the parts. I'm about to start the drilling on the box. I did buy another roll of solid 18 gauge wire to increase the surge capacity per Wooks recommendation. The 6A SB fuse should take care of the rest.

It will be interesting to see if the resistors show any sign of surge, either a glow or smoke.