Filter Choke Overheating

[chandos_r]

Has anyone had the problem of the output filter choke in a switching regulated power supply overheating? This is a computer supply with multiple output voltages (+5, +12, and -5 V), and all of these voltages are correct. However, the toroidal output filter, which has three windings on the same core--one for each of the output voltages--gets so hot it would burn up if left running for more than a couple of minutes! Each of the three windings of this inductor is wired between the output diodes (on the secondary of the switching transformer) and a big filter capacitor. All of the filter capacitors have been replaced, but that hasn't helped. A 3842 8-pin IC controls this power supply. Any help greatly appreciated!

[Edd]

RAY:
Is it worth the trouble any more on those supplys ? Anyhow, sort of got me on the
"toroidal output filter" the ones I see are just wound on ferrite rod cores or ferrite dumbell cores. Of the +5 and +12..I think we can disregard the -5 lightweight. No 3.6Vdc ??? Have you scoped those outputs for ripple to see who the real current hog/culprit is? Is there any chance that that donut might in reality be a
current sensor/detector for power consumpion feedback into the IC control circuitry ? My findings are usually just hi-ESR 'lect caps on the sec filtering side. How about the core temp of the other main 2(3) conventional ferrite core SM x-fmrs ? Last thing, you might keep the unit loaded but drop the loads off on the +5/+12 circuitry lines, only one at a time, and see if consumption is decreased appreciably and then that might clue you in. On your sec filter caps U might try this tip I use. Power up and warm up ....limited by the time restraint you previously mentioned.......then shut down and quickly place 4 finger tips on 4 diff caps and you will be aware if you still have any ESR problem units overlooked,if you hadn't already detected it by a ruptured rubber bottom seal, split top vent seam or just a bloated case. BTW whats the uf/dcv specs on that <big filter cap> ?
73's de Edd

[Recycled Electron]

The most likely cause of overheating is either a design or manufacturing defect. The ferrite produces some heat each time it's input voltage cycles. The amount of heat is strongly dependent on the loss factor of the ferrite. Higher switching frequencies also increase the heating. Not all ferrites are the same. Even those of the same size and shape can be made of various composition materials each optimized for a specific application. I once had a batch of ferrites overheat because they were not the material I ordered. We were manufacturing power supplies. Roughtly speaking, ferrites of the same size and shape can be optimized for use over a single band of frequencies. Of course, the problem might be the original design or the wrong specification of the magnetic properties for the ferrite that you have.<p>If you are interested in learing more, look for published specifications of cores from the several manufactures of such devices.

[Peter38]

Just thinking - could it be that the switching supply is not running at the designed frequency? The flux density in the ferrite increases with decreasing frequency, i.e. if your switching frequency is too low the ferrite may become saturated and heats up. I would have to see an actual schematic to make a more definite judgement call.

[haklesup]

Check the output voltage with a resistive load and you may find that it no longer looks so good (see "PC power supply problems" in the computer part of this forum)