Ways to improve efficiency power supply efficiency?

[japlytic]

I had thoughts about increasing the PWM frequency in a power supply containing a TL494 (which is around 50kHz) to 100kHz.
Secondly, I would also replace the two primary side output transistors with MOSFETs, because they generally switch faster.

Will either of these (or those combined) improve efficiency of a switching power supply (the type with two 200/250V electrolytic capacitors)?

[davmax]

Re: Ways to improve efficiency power supply efficiency?

Careful when changing transistors to MOSFETs they have different drive requirements. Yes MOSFETs switch faster reducing switching losses and improving efficiency. Doubling switching frequency is feasible with MOSFETs however it would do little to improve efficiency without changing more components. Certainly the output ripple will reduce.

[PeteS in CA]

Re: Ways to improve efficiency power supply efficiency?

Oy, this could get weird. In principle you would want MOSFETs instead of bipolars if you're switching at 100KHz. The rise and fall times for MOSFETs are much faster; with bipolars, those rise and fall times are a greater percentage of the period. As dm pointed out, the drive requirements are very different. bipolars are current operated devices: typically the base current is somewhere between 1/5 and 1/10 of the collector current; the V(BE) will be on the order of .7V - 1V. Except for charging the gate capacitances, MOSFETs are voltage operated devices - you need 7V-12V to turn the device on solidly.

With a TL494, the topology is likely to be a half-bridge inverter. This means that one transistor will have its emitter connected to the negative rail and its collector to one side of the O/P transformer primary (probably through a DC blocking capacitor). The other transistor's emitter is connected to the same side of the transformer (DC blocking cap), while its collector is connected to the positive rail. The other side of the O/P transformer is connected to the center of the two series-connected I/P lytics. This is a very long way of saying that you have to have a transformer to provide base drive - the base of the top side transistor is either at ~50% or ~100% of the bus voltage, which would be a bit rough on the TL494. A transformer for supplying 1V at .5A - 1A base drive to a transistor will be very different from a transformer designed to provide ~10V gate drive to a MOSFET.

It can get worse. If the base drive scheme is proportional drive, the transformer has a winding that samples the collector current and provides a base current that is (turns ratio)*(collector current). I'm not sure how well that would work with MOSFETs, but it might cause a bright flash, aloud noise and a bad smell. Or it might do nothing (no O/P).

I think I'm saying that you're envisioning an electronic equivalent of transgender surgery. If you worked out a new drive scheme (probably a new gate drive transformer), your drive and switching losses might be lower, but the core losses in the O/P transformer will be higher, and if the core material doesn't do well at 100KHz, the increase in core losses could exceed the decrease in drive and switching losses.

[arneson]

Re: Ways to improve efficiency power supply efficiency?

I think I'm saying that you're envisioning an electronic equivalent of transgender surgery.

May be like an octoganal wheel, it can role but you'de be in for a bumpy ride.

[bogsof]

Re: Ways to improve efficiency power supply efficiency?

In this way the improving will be small. You need to change many components and to make a new transformer for gate drive.

For a big improvement you will need to do something like this:
https://www.badcaps.net/forum/showthread.php?t=2549