Selenium rectifiers

[Ty_Bower]

Originally posted by Eli Duttman
Add an inrush current limiting thermistor and, possibly, some fixed resistance to protect the power trafo and to bring the O/P voltage into line.

Assuming a FWCT power supply topology, where is the recommended location for adding an NTC device? I've used them on one lead of the primary winding but I'd imagine they could also be added to the center tap of the high voltage winding, or one on each side of the secondary (between the transformer and the rectifiers).

Does it make a difference? Is one method preferred over others?

[Eli Duttman]

I replaced a selenium rectifier by a selenium rectifier
and had no problems.

To paraphrase Yoda, you will. Those blasted things are ticking time bombs.

[kevinkr]

Originally posted by Ty_Bower


Assuming a FWCT power supply topology, where is the recommended location for adding an NTC device? I've used them on one lead of the primary winding but I'd imagine they could also be added to the center tap of the high voltage winding, or one on each side of the secondary (between the transformer and the rectifiers).

Does it make a difference? Is one method preferred over others?

I just stick them in the primary...

[Eli Duttman]

Assuming a FWCT power supply topology, where is the recommended location for adding an NTC device?

Ty,

I dislike placing inrush limiters on the primary side, because bias supply start can/will be delayed. IMO, the best sequence is bias, filaments, and (finally) B+.

For FWCT topology, a thermistor at each end of the rectifier winding or a single part between the diodes and the 1st filter element seems best. Which of the 2 is settled on depends on the resistances (both hot and cold), along with the current handling capability, of the part(s).

[atmars]

the current capability of selenium diodes/rectifiers is rated according to the area of the 'cooling' fins. It's not 20 Watt a fin but each plate can widthstand approx. 20 Volt reverse voltage (could it be that you've mixed it up?).

yes, i believe I read 20V, but I started typing instead of thinking and "watts" came out.

I know that the easiest way to cook a power transformer is to replace selenium with silicon. Been there done that. The advise I got at the time was to use 2) 1N4007's in series to repalce the selenium rectifier. And it didn't work.

This is an old transformerless amp (I have an isolation transformer in front of it) so I don't need to worry about wrecking the PT and the heaters are fed off the ac. Is there any reason to put a dropping resistor in this situation? If I do need to, is 2ohm 10w the consensus?

Thanks for the help

[DigitalJunkie]

IMO: No excuses,they get replaced ASAP on my bench.

They *will* fail,eventually. And they will release toxic fumes,and probably cook your equipment when they do it.

Use a silicon diode,and a resistor to compensate for the lower forward-drop of silicon.

[Wavebourn]

Consider 3 silicone diodes in series for 1 selenium plate if you want the same voltage drop. It is waste of diodes. If you gear don't afraid more voltage it is Ok to replace one by one, otherwise lower taps from transformers are needed.

[atmars]

http://s294.photobucket.com/albums/m...mosoundL-2.jpg

Here's a link to the amp I am working on (conversion to guitar amp) (filter caps replaced). Does anyone have a suggestion on the value of the resistors required ( if required) or a short tutorial on how to calculate the value for a theory challenged modder like myself?

Thanks

[analog_sa]

A short tutorial? About applying Ohm's law?

Can't see a lot from your circuit but seems to me it'll do just fine with diodes. The variation seems less than 2% - quite unnecessary to lose any sleep over that.

[kevinkr]

Originally posted by Eli Duttman



Ty,

I dislike placing inrush limiters on the primary side, because bias supply start can/will be delayed. IMO, the best sequence is bias, filaments, and (finally) B+.

For FWCT topology, a thermistor at each end of the rectifier winding or a single part between the diodes and the 1st filter element seems best. Which of the 2 is settled on depends on the resistances (both hot and cold), along with the current handling capability, of the part(s).

I use them primarily to suppress the first few cycles of primary inrush current, the ones that trip breakers and blow fuses. I do not use them primarily to slow warming of filaments and the like. I have noted that in such instances the bias supply usually charges almost as quickly as without. I don't use SS rectifiers for high voltage in most instances, so the tube warm up delay is usually much longer than the NTC imposed time delay. If you use SS rectification you should make sure the bias is in range before power tube filaments have warmed to the point where plate current can flow uncontrollably.