TOP262 Soft Start
Hi!
The soft start block of TOP262EN seems not to work properly in my circuit, resulting in output voltage overshoot at high line voltage and light load condition. I tried the soft finish solution presented in AN43 but it didn't work. I attached diagram and waveforms to facilitate analysis.
In attached figure, CH1 shows output voltage (set at 5.0 V) with respective overshoot and CH2 shows CONTROL pin voltage. CH2 waveform suggests that, by overshoot time, although pin voltage exceeds internal shunt regulator value (5.8 V), the excess current is not shunted through internal current mirror (as would be expected from the page 8 of TopSwitch-HX Family datasheet description). Furthermore, power MOSFET switching seems to be controlled during this time. Why does it happen? Once this time is a bit greater than soft start time, can I suppose that TopSwitch internal soft start circuit disables current mirror and PWM control?
Please return ASAP as we are very close to finish prototype tests and start production.
Regards.
Comments
Thank you, Spock.
I attached CONTROL PIN current triggered with output voltage. CH2 shows CONTROL PIN current measured across a 10R/1% resistor. The overshoot time was extended if compared with first waveform sent because of this resistor, but you can extract the information you want (the point at witch loop is closed and soft start is decoupled).
The capacitor value I used in soft finish circuit is 10 uF but I think that this is not the problem, because of the waveforms sent. D6 and TZ4 interchange seems not to be a problem once they are in series connection and I did not invert anyone's position.
I'm waiting for your comments.
Regards.
Christos
ok It looks like you should try to speed up your loop. A couple of suggestions
1) what is the power level you are operating at? It may turn out that TOP262 is too large a size of device... do let me know.
2) Consider doubling the value of the voltage divider to LMV431.
3) Change C51 to a lower value to make sure it does not interfere with the bandwidth of your power supply. I'd recommend 100 - 220 uF for this cap. Are you operating in the full frequency mode? Do you know which mode of the TOP-HX you are operating in?
Spock
It doesn’t seem to me that overshoot is caused by control loop speed. Instead, in my understanding, waveforms show that control circuit doesn’t operate for the first ~20 ms, during which overshoot occurs. Maybe soft start feature disables control circuit and, during the respective time, power transfer depends only on device current limits (which are reduced in the beginning and gradually increased by soft start circuit). In this case, I would have to change my TopSwitch by a smaller one.
Power level is about 12 W and input voltage range is 38 to 370 Vdc. I’ve selected TOP262EN because of its low RDSon value, since I don’t have space for heatsink, and because we use the same p/n in other product. Once its standard Ilimit is too large for my power level, I used external resistor to reduce it to 25 %.
Please verify if my analysis is ok, specifically if during soft start time the control circuit is disabled.
Thank you.
Christos
ok I think I understand now what may be happening
1) You are using too big a TOPSwitch for this application. You can get awaywith the smallest TOPSwitch available and still be ok.
2) If you dont want to use a heatsink at 12 W you can use the P package or M package device which does not require an external heatsink.
3) With such a large TOPSwitch you will need to depress the current limit so low that the accuracy of the current limit is not that good.
4) Because you use TOP262 more likely that the TOPSwitch operates in the low frequency mode or multi-cycle modulation mode which are both slow in terms switching frequency (30 kHz)
5) Its true for the 17 ms initially control loop is out of the picture. However when the C pin receives a signal from the opto it will take over control of the feedback system.
6) The argument about low RDSon therefore better efficiency is not 100% correct. As your MOSFET size increases the Coss increases thereby increasing switching loss. So beyond a certain point (we say 1 or 2 sizes larger than required) there is no benefit of using a larger device
Ok, I’ll use a smaller TopSwitch. TOP262EN was good for me because, besides it is already used in another product its package occupies small area on board.
Thank you very much!
Christos
I see your point about area on PCB but honestly the E package parts are intended to be used with an external heatsink...
Anyhow pls try the smaller part in P or M package and that should work for you.
Can you also attach a waveform of the control pin current? Can do this by monitoring the voltage across a resistor in the feedback path. I want to see at what point the control pin current reaches a level that the loop is closed and soft-start is decoupled.
Its odd that the soft finish circuit did not remove this overshoot.. what values of capacitor did you use? I would recommend 22 uF.
Finally you have an error you your schematic on the clamp section. Diode D6 and TVS TZ4 are interchanged... I assume you have these components propoerly connected on your prototype.