Appropriate heat-sinking for bridge rectifiers

[mzy12]

Hello!

I am designing a board to mount a bridge rectifier for a 100V - 6A linear power supply for an existing audio power amp. One part I have chosen is the TBSA60M. There's nothing in the the datasheet to suggest that there's any need to provide heat sinking, but I decided to do so anyways. I would like some feedback to see if the footprint design I have is appropriate/if it will cause any issues.

Another bridge rectifier I am looking at using is the RDXK810, which has a higher forward current and an actually published recovery time. Per its datasheet, it suggests using a, "10 cm * 9.4 cm * 2.6 cm Fin type heat sink." I instead would be looking to mount it to the side of a 2U 19 inch rack chassis, and I was wondering if that would provide an equivalent heat sinking capability. The 2U chassis is made of ~2mm thick steel.

Thanks for reading.

[Konkedout]

If you are rectifying 50/60 Hz mains, then Trr will not be an issue.

I think it may be tough to keep either one of those cool enough at 6A of current.  You are likely to get 4-5 watts of heat due to forward voltage drop.

Of the two, I think I would prefer the vertical block.  You can bolt on some sheet copper or aluminum, perhaps with some Thermal Interface Material or heat sink grease.

There are some bridge rectifiers which have an aluminum mounting base plate.  I think I would prefer something like that.

It is not cheap, but how about something like this?  https://www.digikey.com/en/products/detail/vishay-general-semiconductor-diodes-division/BU2010-M3-45/4832722

[PCB.Wiz]

There's nothing in the the datasheet to suggest that there's any need to provide heat sinking, but I decided to do so anyways. I would like some feedback to see if the footprint design I have is appropriate/if it will cause any issues.

There is info if you decode the thermal resistance numbers.
They spec the I-T curves at case temperatures, so are very optimistic, but they also give 55'C/W on PCB, which means something just over 1A RMS on the PCB area they give.
you also need to worry about how that heats other parts on the PCB.

I instead would be looking to mount it to the side of a 2U 19 inch rack chassis, and I was wondering if that would provide an equivalent heat sinking capability. The 2U chassis is made of ~2mm thick steel.

The bolt-down types are far superior thermally, and the more metal you can give them the better. The vertical side of a rack is a good choice.

[ajb]

The central pad is of dubious benefit if the package doesn't have a thermal pad that can be soldered to it.  You will get some thermal coupling from the package into that copper, but not a lot, since there will be an airgap that impedes conduction and the pad will have low emissivity that impedes radiation into the board.  It also reduces your clearance substantially, which isn't ideal.  You will get better coupling into the PCB via the leads/pads, and connecting them to chunky tracks/pours where possible will help with dissipation. 

If you intend to fit a heatsink, then the second vertical part will definitely be easier, and more effective -- note that it has much lower thermal resistance figures to begin with, plus a much larger package area specifically designed for coupling to a heatsink.  It will also run at a lower temperature even without a heatsink, thanks to its much larger surface area, and vertical orientation that exposes more of that area to convection.

[mzy12]

Sorry, I realised I left some information out on my original post.

I'm doing this to replace four individual diodes on the main board, as they were located physically close to the reservoir caps and the AC power line runs too near to the signal line for my liking. The diodes I'm replacing are rated for 6A and the voltage its rectifying is 80V, hence me looking for diodes rated for at least 100V, 6A.

As for the switching speed:

If you are rectifying 50/60 Hz mains, then Trr will not be an issue.

The faster the reverse recovery time, the less switching noise. I would like to create a robust, quiet solution.

There is info if you decode the thermal resistance numbers.
They spec the I-T curves at case temperatures, so are very optimistic, but they also give 55'C/W on PCB, which means something just over 1A RMS on the PCB area they give.
you also need to worry about how that heats other parts on the PCB.

Thank you. I will take that into consideration.

[mzy12]

... they also give 55'C/W on PCB, which means something just over 1A RMS on the PCB area they give.
you also need to worry about how that heats other parts on the PCB.

Sorry to come back to this again, are you saying the datasheet asking for 1487.5mm³ of copper on the PCB to get a consistent 1A of RMS current?

[drvtech]

Sorry, I realised I left some information out on my original post.

I'm doing this to replace four individual diodes on the main board, as they were located physically close to the reservoir caps and the AC power line runs too near to the signal line for my liking. The diodes I'm replacing are rated for 6A and the voltage its rectifying is 80V, hence me looking for diodes rated for at least 100V, 6A.

You didn't say whether the 80V was RMS or peak. The PIV rating of the bridge will need to be at least the peak voltage of the supply. If your 80V is RMS then that would be 113V so use a 150V part.

[Konkedout]

Sorry, I realised I left some information out on my original post.

I'm doing this to replace four individual diodes on the main board, as they were located physically close to the reservoir caps and the AC power line runs too near to the signal line for my liking. The diodes I'm replacing are rated for 6A and the voltage its rectifying is 80V, hence me looking for diodes rated for at least 100V, 6A.

You didn't say whether the 80V was RMS or peak. The PIV rating of the bridge will need to be at least the peak voltage of the supply. If your 80V is RMS then that would be 113V so use a 150V part.

If you are using PN junction silicon rectifiers, there is no advantage (maybe even no price advantage) to skimping on peak reverse voltage rating.  I would go with 200V or higher PIV diodes.  Also some faster Trr diodes have higher forward drop which will make them hotter.  60 Hz is quite slow for diode Trr.

For standard recovery silicon diodes, I am not aware of any 150V ratings.   These days you may not see much rated under 600V because silicon processing has progressed since the 1N400X series was introduced.

[mzy12]

Apologies again for not being clear. I am definitely NOT looking for diodes with just a 100V reverse voltage rating. I was saying that I needed one rated for at least 100V peak inverse voltage and 6A.

Thanks