Rectifier diodes

[akis]

I have read through some old threads on the subject of choosing rectifier diodes for a PSU.

I learned a bit about possible EMI interference and "snubber" circuits.

In this thread I would like to know if I can use a 5A Shottcky to replace the 1N540x rectifiers simplybecause the forward voltage drop is less on the Shottcky and I get more voltage on the output (and they should get less hot too I'd imagine).

[macboy]

Schottky diodes have a fast and soft recovery. They generally work fine for rectification, but you need to take care in selecting appropriate Schottky diodes. If you are not comfortable reading and interpreting datasheets, then maybe you should stay away from Schottkys. With Si recitifiers, it's easy, just stay inside the reverse voltage specs and current specs, and chuck it in there. Schottkys are more difficult. Except for some newer exotic types, Schottkys have a relatively low reverse breakdown voltage. In general, there is a compromise between maximum reverse voltage, and forward conducting voltage. So if you want to minimise voltage drop and heat, pay attention to both specs; don't simply buy the highest reverse voltage you can find. They also must not be allowed to get hot, because they become leaky (in terms of reverse current) when hot. That's why the larger current ones are often heatsink mounted. You can get into a thermal runaway mode where the reverse leakage current contributes to heating, which increases reverse leakage, which makes more heat, etc...

[Sch3mat1c]

They actually make silicon schottkies up to 300V nowadays, which is amazing. The way they do it is pretty neat.

If you're not doing high voltages, then the main problem is really leakage. High conductance, high leakage types can go into thermal runaway quite easily, as mentioned. This, and probably expense, is the main reason junction diodes are used in supplies over 12V.

Anyway, EMI / snubbing is no problem at 60Hz, anyway, even for the crummiest of 1N540x's.

Tim

[akis]

The only reason I am considering Schottky diodes, eg the SB560 (5A 60V) is for their very low forward voltage drop which means I can squeeze an extra volt or more out of my transformer.

However I did not know about the thermal runaway problems. Is there a particular characteristic I am looking for in order to decide if the diodes are suitable? Eg are the SB560s suitable in a bridge configuration for 5A at 33 Volts DC?

[jan.didden]

Originally Posted by akis

The only reason I am considering Schottky diodes, eg the SB560 (5A 60V) is for their very low forward voltage drop which means I can squeeze an extra volt or more out of my transformer.

However I did not know about the thermal runaway problems. Is there a particular characteristic I am looking for in order to decide if the diodes are suitable? Eg are the SB560s suitable in a bridge configuration for 5A at 33 Volts DC?

Be carefull with the voltage rating. For a 33V DC supply, a 60V diode is not enough. If the diode is cutoff with the secondary of the xformer at neg wave, it will probably be at -35V. The other side of the diode will be at the capacitor DC of 33V (in your case). Total diode voltage is 68V so the 60V diode may say poofff...

For rectifiers, the diode reverse voltage should be at least twice the rectified DC voltage plus some safety margin. In your case I would use minimum 100V diodes. Check the rating on the existing diodes and use that.

jan didden

[metalsculptor]

The gains in using shottkey diodes are relatively small at 35V, Keep in mind that supply regulation is +/- 5% for most power companies, if your transformer is toroidal another turn around the core will compensate for diode drop.

Be carefull with the voltage rating. For a 33V DC supply, a 60V diode is not enough. If the diode is cutoff with the secondary of the xformer at neg wave, it will probably be at -35V. The other side of the diode will be at the capacitor DC of 33V (in your case). Total diode voltage is 68V so the 60V diode may say poofff...

That is true for CT full wave rectifiers but with the bridge rectifier the negative half of the cycle is shunted by the other rectifier to the negative rail so it can only swing to 0V - diode drop. If the transformer is 24V then 60V should be fine if it is 33V then I would be looking at a 75V diode just to give a safety margin for mains transients.

[akis]

I bought a 230v/2x25V/6A toroidal at a very good price, and then realised it might not be good enough for providing 30VDC/5A.

mains: 230V
trans reg: 6%
filter: 20mF
diode drop: 0.8V each / 6A

DC output base: 30.76VDC
plus ripple 3V = avg 31.82V DC

So I need just like maybe an extra volt to go over the losses on the PSU (pass transistors etc) so as to get a net 30V, hense asking about the diodes.

But if I can simply remove a few turns of primary winding from the toroidal would that be a better idea?

[jan.didden]

Originally Posted by akis

I bought a 230v/2x25V/6A toroidal at a very good price, and then realised it might not be good enough for providing 30VDC/5A.

mains: 230V
trans reg: 6%
filter: 20mF
diode drop: 0.8V each / 6A

DC output base: 30.76VDC
plus ripple 3V = avg 31.82V DC

So I need just like maybe an extra volt to go over the losses on the PSU (pass transistors etc) so as to get a net 30V, hense asking about the diodes.

But if I can simply remove a few turns of primary winding from the toroidal would that be a better idea?

Hi,

Why is this exact 30V so important? As pointed out, mains fluctuations are quite large. If you are worried about power amp output, that varies even more with speaker impedance which itself varies with frequency. And even with 1V delta in supply, the difference in output power is probably too small to notice anyway. So what's the real issue here?

jan didden

[akis]

I want to use the transformer to make a PSU "30V 5A" - obviously if it can only reach 29V then it is no good

[godfrey]

Originally Posted by akis

But if I can simply remove a few turns of primary winding from the toroidal would that be a better idea?

No, add a few turns to the secondary.