Power supply shuts off from time to time.
I have a design using a TOP248Y. It's got a 24 volt output at 2.5 amps. When running continuously at a little below the maximum current (still in regulation), the switcher shuts off every few minutes, for perhaps 40-60 seconds. It then turns back on, and functions normally.
It has an oversize heat sink, and I only measure a temperature of 50C or less on the TOP249's metal tab. If I increase the temperature (place a paper towel over the heatsink), the off time goes up a little.
It looks like a thermal cut off, but nothing is close to getting hot.
Comments
That's what it looks like to me, but I measure only about 40 degrees C on the chip's metal tab. (Using on IR thermometer). The supply was originally designed to work at higher temperatures, so the heat sink is fairly large, we use a 249 rather than a 246, etc.
The schematics are attached, with the power stuff on pages 1 & 2. I'm using a Li Shin International Enterprise LSPA10704 transformer ( http://www.lse.com.tw/upfiles/spec01121139842.pdf?PHPSESSID=905cd6c9b64f9f8d528e381617eb5313 ).
Also note that the current limit resistor is not shown on the schematic. However, my test unit has a 14K from X to S. (I can change that easily).
I observe this problem with a load as low as 2 amps, at room temperature.
-Hershel
Are you sure about the IR thermometer? Do you have a thermocouple instead? At 48 Watts only 40 C tab temperature is a little difficult to believe...
Can you try to calibrate the thermometer and make sure its displaying the correct temperature
I did some more testing this afternoon. The temperature was closer to 50C, at 27.5V and 1.98 A. I doubt my IR thermometer is off by much. The thermal limit is supposed to kick in at 140C, and my heat sink is not even hot to touch.
According to the data sheet, the control pin should have a 4.8-5.8V sawtooth during thermal shutdown. I've got a scope connected, and there is no sawtooth while the thing is off. (I see about 150 mV of noise, when it's running, but didn't take any "low noise" measures to watch it.)
-Hershel
I've done much more testing since the last post.
The situation is that I have a fresh batch of 100 power supplies, that all seem to go into thermal shutdown, even though the temperature at the TOP249 is not very high.
I was able to find a supply, from the previous batch (of perhaps a year ago). It's from an identical PCB and BOM, and it does not shut down. (We have been building this unit for 5 years, without any problems).
I had thought that perhaps my contractor used the wrong thermal grease, or something like that, but have since, re-greased it, checked for burrs on the heatsink, torqued the screw, etc., with no difference.
I've thoroughly measured and/or tested every part related to the TOP249, and can see no difference between an old board that works, and a new one that doesn't.
However, all the new boards have a 249 identified as J719/TOP249YN/E46260C ('/' for line). I took the 249 off the old board (J626/TOP249YN/39197C, put it on the new board, and now the new board is OK.
Ironically, the 249 on the "fixed" board runs hotter, about 65C with a 60 watt load. Of course, the power supply no longer shuts down every couple of minutes, which is probably why it was running so cool.
I've got new 249's coming in this afternoon, and I'll try them, but at this time, it looks like there may be a issue that I only see with the J719/TOP249YN/E46260C part.
Are you aware of any changes to the part?
-Hershel
Update:
I received a new batch of TOP249's yesterday afternoon. They are identified as "J931/TOP249YN/59458A.
I pulled 4 boards from my new batch. All had a part with ID of "J719/TOP249YN/E46260C", and all 4 would shut down after a few minutes of operation, then restart after about a minute.
I replaced the TOP249 with the new parts, on all four boards, and the shutdown problem went away. There are no mechanical issues, with grease, burrs on the heat-sink, etc.
Again, the parts that cause a problem, appear to enter a thermal shut down state, but appear to do so at a very low temperature (about 45C on the tab). The replacement parts operate fine, with a temperature of 65C (at room temp), and full load.
-Hershel
For anyone that's interested, the "J719/TOP249YN/E46260C" chips were the problem.
Thanks to PI for sending out replacements.
-Hershel
I had the same problem, but with the LOT Number J508 TOP247Y 32952A. From 40 parts I´ve bought, 18 of them had thermal shutdown problem around 42 degrees.
How can I proceed?
Thanks
Eduardo
Your description certainly suggests thermal shutdown.
Beyond approximately what load do you see this behavior?
Sometimes if the device is not connected "flush" with the heatsink this results in improper thermal resistance and therefore higher temperature.
Can you send me your schematic?