View Full Version : Recreating a Selenium AFC dual diode from 2 diodes?


Jon1967us
08-27-2020, 11:31 AM
I've discovered a faulty twin diode of the Selenium variety in my '56 GE portable and wish to replace with 2 Silicons. I'm aware of the voltage drop difference, but was wondering if there was any problem in simply soldering 2 diodes together at the cathodes (original was a common cathode), thus effectively making a 3 leg component?

I used 2 1n4007s because of the current requirements ( i don't have Germaniums that can take that amount of voltage)

BTW, should I have used 2 Schottkeys instead?

Electronic M
08-27-2020, 01:05 PM
I've discovered a faulty twin diode of the Selenium variety in my '56 GE portable and wish to replace with 2 Silicons. I'm aware of the voltage drop difference, but was wondering if there was any problem in simply soldering 2 diodes together at the cathodes (original was a common cathode), thus effectively making a 3 leg component?

I used 2 1n4007s because of the current requirements ( i don't have Germaniums that can take that amount of voltage)

BTW, should I have used 2 Schottkeys instead?

It will work but shotkey or other high speed diode are preferable... there's a good thread on ARF that recommends a specific part.

Jon1967us
08-27-2020, 02:39 PM
OK, I'll search for it. Thanks.

Jon1967us
08-27-2020, 02:52 PM
Reposting some info from ARF here to save anyone else the trip:

A fast recovery silicon diode such as the 1N4087 that he used should work. I believe the 1N914 would also work.

The 1N4087 has a recovery time of a couple of nanoseconds which is plenty fast.

The selenium rectifiers had a fast recovery time because they had negligible charge storage. However, the capacitance should also be kept small which means a small contact area which in turns limits the current capability.

The 1N4007 has a recovery time of 30 microseconds. That's almost half a horizontal scan so that obviously wouldn't work. The 1N914 has a recovery time of 4 nsecs, so it should work.

Jon1967us
08-27-2020, 08:07 PM
https://ibb.co/kJGy30Q

Actually, two 1n4007s tied at their butts works just fine apparently.
At least, in my little GE...

old_tv_nut
08-27-2020, 10:24 PM
Reposting some info from ARF here to save anyone else the trip:

A fast recovery silicon diode such as the 1N4087 that he used should work. I believe the 1N914 would also work.

The 1N4087 has a recovery time of a couple of nanoseconds which is plenty fast.

The selenium rectifiers had a fast recovery time because they had negligible charge storage. However, the capacitance should also be kept small which means a small contact area which in turns limits the current capability.

The 1N4007 has a recovery time of 30 microseconds. That's almost half a horizontal scan so that obviously wouldn't work. The 1N914 has a recovery time of 4 nsecs, so it should work.

1N4007 spec sheet says 2 microseconds recovery time, so should work.
https://www.rhydolabz.com/components-diodes-leds-c-172_194/1n4007-1a-power-diode-p-2214.html
I thought 30 microseconds sounded suspect.

Jon1967us
08-28-2020, 12:54 AM
Good catch old TV nut!