Power supply design review

Can you please clarify if this is only a thermal problem or an electrical problem?

If it is purely a thermal issue, you should be able to apply full load (144W) to the power supply for a short period of time. If this is the case it sounds like you need more effective heatsinking.

What are you using for a heatsink currently? Selecting a heatsink with a lower Rth (thermal impedance) will lower the device temperature.

Good luck!

-Wesley

Hello,

Yes, it is strictly thermal issue. I believe it can deliver the full load. Thus far, I have only added upto 100W and it operates for ~10min.
I do have a heatsink on it, but I thought the device should operate in openframe without heatsink (i.e. per datasheet), is this assumption incorrect?

BTW, for your information, when I blow cold air on it, then it can run continuously.

Thanks.

John

Hi John,

Yes the datasheet specifies that this device can deliver the power in an "Open Frame" enclosure. What this means is that air has the ability to move, to some degree, freely through your power supply and thus the heatsink attached to the device. This is as opposed to an adapter thermal environment where there is no air flow within the enclosure.

A cell phone case or a laptop power supply are good examples of adapter applications where there is a plastic case completely surrounding the power supply and thus does not allow much air to enter or exit the power supply. Basically this adds another thermal impedance from the heatsink to the ambient temperature, that being the case.

I hope this clarifies how we specify our parts and the necessary steps that you need to take to keep the device temperature down.

-Wesley

Thank you Wesely.
We are currently revisiting our heatsinking at the moment as well. We do have issues with real estate in this regard.
Is there a way to improve the efficiency of the design or do you think this is optimized as it is?

BTW, TOP259EN (adaptor) can deliver upto 80W (according to datasheet). Our unit overheats with even 80W load and this is with TOP259YN (open frame). This is why we think there might be something wrong with the efficiency of the design.

Any thoughts?

Thanks.

John

Hi John,

If you have a small, or no (!) heatsink on the device I am not surprised it cannot deliver 80 W. If you are worried about low efficiency, measure it and see what you get. That should tell you if you really have an efficiency problem. However just having an efficiency number will not tell you everything since the losses are distributed across the major components like, input diode/EMI filter, transformer, device and output diode.

You can try to go up a device size, possibly two sizes to decrease the conduction losses and thus the power dissipated in the device. This should also increase your efficiency. However as you increase your device size, so that the internal MOSFET has a lower RDSon, it is true you reduce your conduction losses but you also increase your switching losses. This is why I suggest only trying a device size or two larger since going too large can increase your switching losses to a point where you start dissipating more power in the device and thus lowering the overall efficiency.

Also if you do try larger device sizes make sure to use the current limit feature of the device to make sure the maximum current limit will not exceed a peak flux density of around 4200 Gauss since that may saturate the transformer and cause the device to exceed its SOA.

Good luck!

-Wesley

Hi Wesley,

Thanks again for the quick reply.
Just one more question. Is there a typical heatsink that is used with TOP259YN devices? If so, can you please pass the mfg partnumber. For your information, this is the heatsink we are currently using, as you can see the heatsink is shared between TOP259YN and the output diode. Is this a bad idea to share heatsink between these two devices? Also, do you see any obvious flaws with this heatsink?

Thanks.

John

John,

The heatsink looks fine- if not a little small perhaps. It's not necessarily a bad idea to share the heatsink. I noticed that you're using insulators on both devices. Our device's pad is tied to source so to heatsink would be at the primary ground potential. I'm not sure if the output diode being shared on the same heatsink meets safety isolation standards though.

That all said if you can eliminate the insulator pad and use thermal grease from case to heatsink it will lower the total thermal impedance and help the heatsink be more effective. In your case though I don't think that alone will solve your problem.

Also, obviously, sharing the heatsink means that the heat generated by both the diode and device must be effectively carried away by the single heatsink.

I have no specific suggestions for a heatsink. I would try anything with a lower Rth and then go from there. Unfortunately it's a bit of an iterative process.

Good luck!

Thanks Wesely,

Your suggestion of removing insulation on TOP259 device is a good one.
We only need the insulation on the diode, this should keep it the two device from shorting each other. We will give this a try, also we are planning to go with copper heat sink instead of aluminum.

Thanks.

John

Good luck! Glad I could help.