Selenium Rectifier replacement

You can replace the Se unit with any silicon bridge, or make one up from individual diodes - as long as you keep the 80 amp number in mind. Go to digi-key and search.

You can replace the amp meter with any small milliameter by using a shunt formed from a short piece of smaller gage copper wire. Some experimentation required for that. It will be easier to find a small milliameter like that then a surplus 80 amp meter.

Jim

Brad

Do you have access to broken alternators. Their diodes can be used if cost is a consideration. The amp meter from an old truck might be affordable too.

Jerry

Whatever you do it won't be cheap unless you get incredibly lucky. And you will also need luck to find a drop-in panel meter.

Other than that, it should be easy to find a rectifier and meter for your ratings, just watch ebay for awhile. Not too many buyers need those specs. When you get your rectifier, take the time to find the actual datasheet, download it, and read it for the manufacturer's suggestions on mounting especially heat sinking, and use new heat sink grease. If you have to retrofit a new panel meter with a slightly different footprint it isn't too hard to cut out panels with a jeweler's saw, or else make a bunch of punches along your profile and then use a small sanding drum in a Dremel tool (or carbide burr in a die grinder, or...) to clean up. You can also chain drill, although I hate drilling sheet metal.

GWE

I've done a fair amount of selenium replacements in old TVs back in the stone-age(tubes). You can replace the rectifier part of the function with any suitably spec'ed silicon replacement, full-wave bridge units are cheap. Check the usual electronic surplus joints on the web like All Electronics, Digi-Key and Mouser are two retail outfits that don't mind onesie orders. Surplus Center out of Lincoln, NE and C&H Sales also have both rectifiers and amp meters from time to time.

The other thing about seleniums is that they've got quite a voltage drop, part of the replacement procedure was adding a power resistor in series to get the proper voltage to match the original selenium setup. An 80 amp power resistor might be a little spendy. You really don't want too high a voltage applied to your batteries or they'll boil and overheat. The selenium rectifier added something resembling voltage regulation to the circuit, too, with that voltage drop.

Stan

How about an extra diode in series with the output after the bridge? This will drop the voltage about 1v under heavy load. Better be a big diode with a heat sink though. This also protects the bridge if someone hooks up the leads backwards. Without this diode, connecting the leads to the battery backwards will provide a short circuit across the transformer secondary. Been there, done that, replace the bridge and added a fuse for the next time. A diode would have been better than a fuse, but the extra voltage drop was too much for my transformer.

The resistance of the selenium rectifier would limit the current of a charger; using another silicon diode would not. I would add the extra diode if I could, but before I did that I would make sure that I had enough resistance is the circuit to equal the equivalent series resistance of the selenium rectifier.

It may be cheaper, in fact, to get your hands on a bunch of 5A rectifiers and load them with lower power resistors rather than finding a single 80A bridge (at the cost of more work, of course). Then connect them like so:

___ .---->|----|___|--. | | | ___ | ------o-o-->|----|___|--o-o------ | | | ___ | '-->|----|___|----' created by Andy´s ASCII-Circuit v1.24.140803 Beta

Just make sure that:

• the parallel equivalent resistance adds up to your desired resistance,

• the current capacity of the diodes adds up to your desired current capacity,

• the voltage drop across the resistors is at least as much as that across the diodes,

• the resistor power ratings are observed,

• and that if one of the diodes fails open, the others will follow in short order.

So I can parallel either diodes alone, or rectifier blocks to get increased current capability? I understand about the cascade failure if one should go. Since the charger has about 6 taps on the primary, for increasing rate of charge (simply higher voltages) can't one just select a lower setting and ignore the voltage drop the old selenium rectifier had? I'm not quite sure I understand how the old unit acted as a voltage regulator. Is that just because it had resistance/voltage drop that modern diodes don't have? The old amp meter has a shunt of about 1/2" wide copper about

1/16" thick, so it was just a milliamp meter with a calibrated scale. I suppose I could hook up something like an old toaster with a known dc resistance to calibrate a surplus one if I wished. Alternator diodes are a good idea. They might be rated high enough to do the job individually, without paralleling them. I looked at some online places and didn't see much over 50A with a quick peek, hence the first post here. The old unit has the selenium stack under a fan, so it should be easy to mount the new rectifier on a surplus heat sink and place it in the air flow. This was a curb find, so at a minimum its a good high current/low voltage ac transformer.

A tiny amount of resistance in series with each diode will help equalize currents among them.

But this is getting crazy. Bridges with 500 Amp peak capacity are inexpensive enough not to have to cludge something together. I don't think your transformer will deliver that.

LLoyd

You cannot parallel diodes directly because they will have a small voltage mismatch and one diode will hog all the current. You can only parallel diodes if you force them to share current by putting a small resister in series with each diode.

Seems very reasonable and worth a try.

That's what I hoped, but so far digi-key, Mouser, and all electronics are coming up with 50A as the max. They are only ~$ 2.50 though so using two in parallel would be cheap.

Grant, Next time you need to drill holes in sheet metal go buy a "Bullet" Drill. These are (were?) made by Balck and Decker. Anyway, they have a grind on them like an endmill except they have a small split point that extends away from the drill face. When drilling you end up with a little washer. They are great for drilling sheet metal. ERS

Chuck, I built a three phase rectifier that used 4 1/2 bridges. Actually, I had 9 winding to contend with. This rectifier assembly is being used to power the servos in a CNC conversion. If one diode hogged all the current it would have blown already. Am I just lucky? The rectifier bridge assembly recieves 90 VAC from the xmfr. The servos draw 15 amps max. The output from the rectifiers goes into an electrolitic cap for filtering before going to the servo amps. Should I put resistors in between the diodes and the xmfr? If so, what values do I need? Thanks, Eric R Snow

I don't understand your circuit; If you have diodes connected directly in parallel good engineering practice is to put resisters in series.

However, since you say its a 3 phase rectifier, I suspect you have the outputs connected to the same cap but the inputs are connected to different windings. In this case your diodes are not connected in parallel and you have nothing to worry about. In fact diodes in a 3 phase rectifier are not working nearly as hard as diodes with single phase input because the input phases overlap.

chuck

Surplus sources for rectifiers are phenomenally cheaper than new, as well as ebay. A diode is just as good new or used, providing it hasn't been cooked excessively. I highly recomend avoiding parallelling diodes. One will always pass more than the other, and that one will get hotter than the other, allowing it to pass more current, thus leading to thermal overload. Spend the money you'd spend on two parallel diodes on one with the right rating and save yourself the hassle of trying to pack it all in. Just ain't worth it.

Check this from ebay out:

You can find more on ebay and elsewhere by patient searching. I'm keeping an eye out for a set of four for a welder DC conversion at a good price. They're out there, you just gotta look.

These are cheap but can be made to work with one extra wire fed back to switch them on. Great starting price.

And you can do a search on ebay for "shunt meter" for your high current application.

I sold a bunch of Fuji 150A rectifiers on eBay early this year. They sold for about $25 apiece.

i

I disagree with the note below - get a 100 miliamp meter and make a shunt out of 12 gauge wire - start with about a foot - apply a known load and check calibration - make the shunt longer or shorter as required. Note - if "shunt" doesn't mean anything to you, I mean to take a length of

12 gauge wire and connect it across the meter terminals (like a short circuit) - you want to make a 1000 to 1 divider so you don't need much resistance.

LOL. I did this by accident on the battery charger I made from surplus stuff. I had a nice 5 amp meter, and a 50 amp panel meter as well.

I wanted to be able to put them in series to read them both, but didn't want to burn out the smaller one or bend the needle. So I put a toggle switch across the 5 amp one, to short it out when I went above 5 amps.

The odd thing is, once I close the switch, both meter needles deflect the exact same amount - the wire and switch make a perfect shut to make the 5 amp meter read 10x less senstive!

Jim

Greetings Chuck, Boy, this diode thing has been bugging me for a long time and you just answered the question. Yes, the outputs are indeed paralleled and connected to the filter cap. And the inputs are each connected to one of 9 windings from the three phase xmfr. Now I understand that when people speak of parallel diodes they mean the inputs and outputs are connected in parallel. This explains why the diodes haven't blown. Thank You, Eric R Snow

You'll need to determine what type of rectification is used if you intend to use the unit as a battery charger. Charging lead-acid batteries with too high of a voltage will shorten their life. The transformer output voltage (before rectifiers) could possibly be 18VAC or more.

The added resistance that Stan and the others have mentioned is what typically complicates replacing selenium rectifiers with diodes. If the transformer's secondary windings are on the outer layers, it might be possible to carefully remove winding turns to achieve the proper/desired output voltage, when utilizing diodes.

If the transformer isn't a constant voltage-output type design, output regulation would be a good idea, but maybe more involved than you'd care to get into. Disregarding regulation, I'd try to obtain a 13.5V to 14V maximum output voltage for charging 12V batteries. I personally wouldn't want to charge 12V L/A batteries at more than 14V.

Changing the rectifier configuration will also change the output capabilities.

1 rectifier is half-wave rectification 2 is full-wave 4 are the equivalent of a full-wave bridge rectifier There are also other configurations for voltage doublers. Battery charging doesn't necessarily require full-wave bridge rectification (4 each or 2 pairs).

If you want to *not* use it as a battery charger, you can choose any diode configuration you want to.

If you want a bench-use power supply, a jeweled moving-coil type meter would be suitable, but for use as a portable battery charger, a ruggedized moving vane type would be long lasting (typically less accurate, more of a monitoring indicator).

Stud-mounted diodes are relatively easy to use in a variety of heatsink designs. There are both case-anode and case-cathode for many stud-mount diode types/ratings, so that using them on a single heat sink doesn't require insulating washers/bushings, etc.

Diodes are destroyed in the blink of an eye, well really fast, and normal fuses usually aren't fast enough to protect them (common circuit breakers certainly won't). Fast acting rectifier protection diodes are what's used to protect the output semiconductors in DC motor drives. Many battery charger owners have learned not to lend them to anyone, as it often results in getting back a dead charger.

Choosing diodes with a high peak surge current rating (exceeding maximum x3 or more) will often add to their longevity. I wouldn't even consider using under-rated diodes (connected in any manner, or marginally useable diodes) to save money. The common failure mode for PN junctions is shorted, so you can end up with an AC output.

Meters, rectifiers and lots of other great stuff can also be found at Fair Radio in OH.

WB .............