Very high bias voltage
Hi,
The bias voltage of my design with TOP267E 14V 2.5A single output, is 15V at no load. It increases to 27V at half load which sends the power supply into auto-restart mode.
PI Expert file is as attached.
The power supply manage to supply full load 2.5A with V pin grounded.
Read in the forums that added a small series resistor could help. But for my case it's too much voltage to drop. I'd rather find out the cause of the problem and fix it there.
Is it just down to the coupling of the transformer?
Please give suggestion to solve this problem and to improve design.
Thanks,
bbmilo
Comments
Hi Traveller, thanks for your reply. Unfortunately we don't have the facillity to make transformer in-house. However we do have an LCR meter if that's good enough to measure transformer characteristic. According to RTSS Sample report, the leakage inductance is 11uH which doesn't sound too bad.
The V-pin has already been earthed which, as you mentioned, disables all the voltage protection features. The power supply is delivering full load at the correct voltage.
I have tried adding a resistor between the rectifier diode and the smoothing capacitor at the bias winding but it didn't help.
The LCR meter isn't going to be too helpful unless you have the ability to set the frequency at which it makes measurements...then you're going to be stuck with making a long series of measurements to verify the transformer design. That being said, I would look at the bias winding inductance with all other windings open and then again with only the secondary winding shorted. This should give you an idea of what the bias to secondary leakage inductance is.
It's really a shame you don't have an impedance analyzer. They're really handy for checking/evaluating transformer designs. It's best not to leave the measurements up to the magnetics vendor (at least during the prototype phase).
A couple things you might try:
-With the Vpin tied to ground and VR1 removed, bring the supply up at a very low input voltage and very light load. Using your scope, I'd double check the voltage waveforms on D2 and D3. The snubber circuitry might need to be modified.
-If it's only a light-load problem, you might be able to temporarily work-around it by adding a small pre-load to the bias output. This might help keep the bias voltage in check.
-It probably wouldn't be a bad idea to look at the bias, secondary and primary winding voltages using HV differential probes. Large leakage inductance values will show up as large voltage transients during turn-off.
Barring some measurements to evaluated the initial transformer design, I don't have a whole lot of suggestions. If this was a design I was working on, the first thing I would do is verify the transformer design using an impedance analyzer. If the transformer didn't look right, I'd restart the transformer design with a very simple winding layout with the bias located away from the primary and get the supply up and running in that configuration.
-The Traveler
bbmilo -
From what you're describing, I think the transformer is probably suspect. Couple of questions:
Do you have the ability to make transformer prototypes in-house?
Do you have the ability to measure transformer characteristics?
If so, I'd experiment with some different winding layouts. I'd probably start by eliminating the shield windings and do a simple transformer layout (pri, sec, bias). I'd also be curious to see what the leakage inductance on the bias winding looks like on your current transformer.
From a debugging/troubleshooting standpoint, I'd recommend disabling some of the extra features in the design such as UVLO, OVP, etc. Get the bias voltage issue solved first and make sure the power supply is functioning correctly before you start adding the other features on. Then you can test them as you go.
-The Traveler