Help Identifying Selenium Bridge Rectifier Specs

[fuzzoli]

This is a NOS Selenium full wave bridge rectifier, and I'm trying to figure out the ratings.   I've tried all the requisite google searches with no luck. 

Any thoughts?

Thanks!

[floobydust]

Is it surely selenium? I would measure V_F on it. I found the ITT "10B1" is silicon 100PIV 1.3A to 1.8A with heatsink, circa 1961.
Otherwise, current capability is proportional to area and voltage rating achieved by stacking elements in series.

[Andy Chee]

Do you have a curve tracer?

[Terry Bites]

Try dropping a line to thse guys
https://www.gdrectifiers.co.uk/product-category/power-assemblies/selenium-rectifiers/

[wraper]

AFAIK it's about 25V per cell/plate.

[fuzzoli]

@floobydust, I'm pretty sure it's selenium (please see below).  Vf is 0.4 - 0.46v depending on which diode is measured.

@Andy, I wish I had a curve tracer --- maybe there's an app for my scope or DMM...

@Terry, thanks for the link!
 
At the bottom of the pile, I found the top of the box, so that's a win-win!  It's an RCA Full Wave-Bridge (Selenium), part #59504.  Googling the part number, I found it being used in a broadcast audio Limiting Amplifier (RCA BA-6A), which makes sense considering the basement I found this stuff in. 

https://funkwerkes.com/web/wp-content/techdocs/MixedProAudio/RCA-BA-6A-Compressor-2.pdf

Page 13 of the manual shows it as "Rectifier, selenium, bridge, input 18 v ac, output 12.5 v dc at .780 amps", with an RCA part number of 59504.

...so I think I've got it.  Thanks for all of the replies and links!

-Frank

[CaptDon]

Those seleniums sure had a lot of loss!! I remember seeing rectifiers that looked exactly like seleniums with the big plate structures but they had little 'top hat' silicon diodes contained within the plates. I think GE made the ones I ran across. Maybe they were designed as a drop in replacement for seleniums since they had the same footprint?

[TimNJ]

Are you trying to replace it with something modern? Or what is the impetus?

Selenium rectifiers were often (always?) used with some external series resistance. I guess the surge current rating of selenium was not very good? And they have fairly large effective resistance of their own -- I don't know, maybe 1 - 10 ohms per plate for something like that.

Mentioning because circuits designed with selenium rectifiers need some consideration before replacing with silicon. Voltage drop of the rectifier itself is ~10x lower, and if a series resistor was used, there's some influence on ripple (if used for mains rectification), and nominal rail voltages.

[TimNJ]

For a low voltage bridge, voltage drop won't be too crazy I guess, but probably at least a few volts. I think for the high voltage versions, they  stacked a bunch of 20 - 50 volt cells, so the voltage drop of the total series arrangement could be rather high.

[wraper]

Mentioning because circuits designed with selenium rectifiers need some consideration before replacing with silicon. Voltage drop of the rectifier itself is ~10x lower, and if a series resistor was used, there's some influence on ripple (if used for mains rectification), and nominal rail voltages.

You cannot blankly say 10x lower. Low and high voltage silicon rectifiers have barely different forward voltage drop. Selenium rectifiers on other hand were assembled with different number of cells in series depending on voltage required. This bridge rectifier has just 4 elements, so there are 2 elements in series between input/output at any given time, that means about 1V(per element)*2=2V voltage drop.

[TimNJ]

Mentioning because circuits designed with selenium rectifiers need some consideration before replacing with silicon. Voltage drop of the rectifier itself is ~10x lower, and if a series resistor was used, there's some influence on ripple (if used for mains rectification), and nominal rail voltages.

You cannot blankly say 10x lower. Low and high voltage silicon rectifiers have barely different forward voltage drop. Selenium rectifiers on other hand were assembled with different number of cells in series depending on voltage required. This bridge rectifier has just 4 elements, so there are 2 elements in series between input/output at any given time, that means about 1V(per element)*2=2V voltage drop.

That's what I was trying to get at in my follow-up post. Drop can be 10x higher than silicon if you are trying to rectify line voltage. For low voltage rectification applications, the difference is less.

Also, and I am not sure what property explains this, but forward voltage tends to increase much more with increased forward current compared to silicon. There could be a forward voltage increase of 2 - 3x over the usable forward current range, whereas silicon might be more like 1.5x over the same range.

Just saying that there are a few things to keep in mind.

[The Doktor]

Also, and I am not sure what property explains this, but forward voltage tends to increase much more with increased forward current compared to silicon. There could be a forward voltage increase of 2 - 3x over the usable forward current range, whereas silicon might be more like 1.5x over the same range.

That effect is because a selenium rectifier has a fairly large resistance compared to silicon, so with a silicon rectifier the majority of the voltage dropped is just the PN junction, the selenium is more like a junction in series with a resistor.

If you are planning to use this keep in mind that they are very inefficient, lifetime is limited, and I am told they emit a smell when they burn up which is so horrible you will remember for the rest of your life.

[ArdWar]

I am told they emit a smell when they burn up which is so horrible you will remember for the rest of your life.

The fume is also extremely toxic, so yeah you'll definitely remember it one way or another.

[TimNJ]

Also, and I am not sure what property explains this, but forward voltage tends to increase much more with increased forward current compared to silicon. There could be a forward voltage increase of 2 - 3x over the usable forward current range, whereas silicon might be more like 1.5x over the same range.

That effect is because a selenium rectifier has a fairly large resistance compared to silicon, so with a silicon rectifier the majority of the voltage dropped is just the PN junction, the selenium is more like a junction in series with a resistor.

If you are planning to use this keep in mind that they are very inefficient, lifetime is limited, and I am told they emit a smell when they burn up which is so horrible you will remember for the rest of your life.

Indeed, I understand that a selenium rectifier effectively has a series resistance, and can be modeled as such. Just curious whether this is inherent to selenium (and why?) or if it’s something particular to way the old diodes were constructed.

[The Doktor]

I am told they emit a smell when they burn up which is so horrible you will remember for the rest of your life.

The fume is also extremely toxic, so yeah you'll definitely remember it one way or another.

Or possibly you won't? 

[wraper]

Another 30V 600mA bridge rectifier as example.

[ArdWar]

Who decided scaling the X axis like that is a good idea?

[floobydust]

I know of customers that specify using selenium rectifier diodes, to this day. They're still produced.
This is for cathodic protection power supplies. They experience a lot of big lightning surges and selenium diodes in reverse-breakdown are just like a huge MOV. So they can be more reliable than silicon in the abuse they will tolerate. There is a fungus that eats selenium in high humidity and that is one nemesis.

As far as replacing selenium diodes with silicon, it's never gone well for me. You end up with way too much output voltage.
Even in (single stack) car battery chargers, too much output voltage and not usable at say 17VDC. I just toss the charger in the garbage instead of ruining batteries.