Cooler recifier diodes?

[mole42uk]

ISTR a comment on here some time ago about a type of diode, suitable for use in a rectifier, that would run cooler than a typical 1N5401. I don't remember who, what, when or where. Can anyone help?

I have a power supply in a keyboard synthesiser that has been running very hot!

[toshiba tony]

Can't you just put some meatier diodes in, current wise. What are the current ones?

[Craig Sawyers]

Either go for TO220 packaged devices and heatsink them, or use a lower voltage drop sort such as a power Schottky.

[jjl]

You could use schottky diodes that have a typical forward voltage drop of around 0.2V rather than around 0.6V for a silicon junction diode like the 1N5401. This would divide the watts lost in each diode by 3. The only disadvantage is that schottky diodes are only available with reverse voltage ratings up to 100V or so, but this is unlikely to matter for your application.
If the existing rectifier diodes are running hot it sounds as though under-rated diodes have been designed in or there may be a fault.

John

[snowman_al]

Schottky diodes?
Lower forward voltage drop so less heating for same current?

[Keith956]

You could replace the existing power supply with a SMPS module, probably lighter, smaller and cooler.

[Ed_Dinning]

A dodge that was used on some 70's TV's was to leave the diode legs long and to fit a small piece of copper foil on them.

Ed

[Craig Sawyers]

Originally Posted by Craig Sawyers

Either go for TO220 packaged devices and heatsink them, or use a lower voltage drop sort such as a power Schottky.

I've just been looking at a little more detail about power Schottkys. It is true that signal level devices have a small voltage drop, power Schottkys are very similar to conventional junction diodes, with a drop of about 0.7V to 0.8V at a forward current of 10A or so, and 0.55V at 1A.

That is for 150V repetitive PIV.

You can get 650V plus, but only is Silicon Carbide Schottkys - and they have a forward drop of about 1.4V. The advantage is very high operating temperature.

The real advantage of a power Schottky is no, or exceptionally low reverse recovery, which is a big advantage when used in a switched mode power supply.

But, in spite of my comment of "use a power schottky", there is no real advantage in a regular linear power supply.

Craig

[PJL]

A 25% reduction in forward voltage will halve the power dissipation. The 5401 is a 3A diode, what is the application?

[mole42uk]

Unfortunately it's a rather badly designed power supply - lots of heat in a very small space....

[GrimJosef]

I've used these

https://cpc.farnell.com/stmicroelect...5v/dp/SC07414#

when making bridges to rectify 6.3VAC. At 3A Vf is 0.3-0.4V depending on how hot they are, but I'm usually running them at lower currents. They are only rated for 25V VRRM but again that's fine for my application. Combined with a very low forward voltage drop regulator I can run small-signal valves' heater wiring with DC rather than 50Hz AC and in a badly designed modern audio amp (of which there are rather too many) that can allow quiet operation with sensitive speakers.

Cheers,

GJ

[mole42uk]

The rectifier is on the output of a 7V 12VA transformer so really it shouldn't be stressed at all.

[Radio Wrangler]

1N5401 is 100v 3A well within what Schottkys do

Look at the circuit design. If they rectify and feed a capacitive load, then they will charge it with narrow spikes of high peak current and things like 1N5401 are spec'd at 1.2v drop which is a lot more than standard silicon Vbe.

Schottkys also exhibit increased drop at high current, but their starting point advantage still leaves them better at high currents. With the right parts, you could still halve the dissipated wattage.

David

[mole42uk]

Bridge rectifier, 3300µ capacitor, TO220 5V regulator. Nothing surprising or dodgoise there. No signs of transformer heating.

[Neutrino]

The photograph shows the following regulators:
• 7812CT
• 7912CT
• 7905CT
• LM2940CT -5.0
• MIC29150 3.3WT (not sure: hard to read)

These are linear voltage regulators. Using rectifiers with a smaller voltage drop would result in more voltage drop in the regulators so that the total heat dissipated in the power supply would be the same.

A data sheet for the rectifiers is here:
https://www.vishay.com/docs/88516/1n5400.pdf

The rectifiers are rated for a maximum junction temperature of 150°C and rely on conduction through the leads to dissipate heat into the heat sink (printed circuit board). The blue electrolytic capacitors are marked 85°C. If the power supply is running very hot it might be worth changing the electrolytic capacitors to higher temperature types rated at 105°C.

David

[Chrispy57]

As Ed mentions in post #7 long leads may be the answer? That would also leave the diodes further away from the pcb and allow for air circulation all around the diode bodies.

Cheers
Chris

[turretslug]

That PCB shouts "hot, hot, hot" and "quart in a pint pot" to me- I get the feeling that some layout designers see rectifiers as wattless devices, others get a bit further and have "0.7V" in their minds without appreciating (as mentioned above) that 1.1V or more is likely to be nearer the mark once peak currents and bulk resistance are taken into account. It's a shame that the diodes weren't fitted with a kink in their leads to hold them a few mm above what is a pretty good thermal insulator, fitting them flush with a PCB isn't helpful to either party. There's a balance to be struck- rectifiers have lead thickness quite disproportionate to current rating in order to get the watts away from the speck of silicon as effectively as possible and hopefully into an area of copper- somewhat limited opportunity here. Supports like ceramic beads are temptingly neat and convenient, but act as insulators, so shaped bare leads would be the better stand-off solution.

[mole42uk]

Yes, you are right, the design has far too much heat generation per unit of dissipation. A fan would help but this is a music synthesiser so quiet is better than cooler.

I have ordered 105º capacitors and several new 1N5401 which will be installed 20mm above the PCB - that's all the space that is available!

Of course, the next problem will be that all the power is delivered to the processor board through a 10-way ribbon cable, the connector is visible on the photo. Perfectly adequate for logic-level signals but this one carries +12V, +5V, +3.3V, -5V, -12V as well as the ground connections....

[snowman_al]

Any space for a small computer fan in there?

There is a swings/roundabout thing here, reduce the diodes dissipation and that will up the regulators by the same amount.