AC input stage blows-up
Hi, I designed a SMPS by using PI experts (TOP244YN) with next characteristics: Vin 85-265VAC, Vout: 8V - 2.5A. At Vout = 8V, 2.25A with Vin = 220VAC, input current was 90mA RMS. The input stage has two 15ohms NTC (one on each phase), a WM10 bridge, and a 1.5Amps EMI chocke, schematics is just like PI recommends.
I built around 80 SPMS whichs runs well fine but I found in a few the next problem.
1) Two of them blowed-up their input stage. The TOP244, WM10, EMI chocke and both of two NTC were destroyed. I think this happened because I unpluged the SMPS and inmediatly pluged-in it again in another outlet (it could be the phase and common 220VAC cables were interchanged in this outlet). I do not remember if I did this with the SPMS loaded or not and I not sure if this is the reason of the blowing.
2) Because of this, I changed one of the NTC by a 1 Amp fuse (ten times the RMS current at full load). All worked fine until this week when another two SPMSs blowed-up their fuses. Replacing them, the SMPSs worked again. This means the rest of the components weren't destroyed (the fuse protected them all), but the fuses still blowed-up and I don't want this happen.
3) In order to check it, I take randomly a few SPMSs and I pluged-in and un-pluged them several times in different outlets, at full load and without any load in order to repeat the blowing but nothing happened!!! the SPMSs still worked fine, but suddenly I taked one SMPS which runed well fine, and I used it regulary and then I unplug and pluged-in it again and the SMPS blowed-up it input stage.
I don't know exactly why this happen, why the input stage blows-up in some of the SPMS which are running well fine at the instance of plug-in, when I'm using all the protection recommended. I´m worry about my clients who can use regulary their SMPSs and suddenly became destroyed. In all the cases, the blowing happened at the plug-in instance in the AC outlet.
Could any give me some orientation, please ?
Comments
Please, see the pdf schematics attached on this topic. Yes, I think you could by right about the NTC heating but I wondering if the RMS input current is about 90mA at full load (in AC line), this means, by looking the V-I curve for SCK-154 NTC, that at this current the NTC resistance still remains in 15omhs so, there is no heating termistor at all. By unpluging and pluging-in again the NTC only reduce the inrush and then it cool down again.
Hi, I tested again the schematics in the PI Experts and it passed (no optimization) by using one NTC and one 1Amp FUSE. By looking deeply, I can see that there are two kind of fuses, fast and time-lag fuses. I'm using an 1Amp axial 4x11mm glass type fuse model ZH207-010. I think this is a fast blowing fuse and this is the reason of the blowing-up. I think this fuse, at 220V/15omhs(NTC) = 15Amp blows at 10msec when I plug the SMPS. May be I must use a greater fast fuse or to use a time-lag 1Amp fuse. What do you think about it ?
Yes, you're right. I see in the oscilloscope peak current when I plug and unplug several times the mail cord. The peak rises !!! This is because the NTC is heat and so its resistance is low. By looking for the input stage datasheets I observed the next:
WM10 bridge: max current 1.5A, peak current 50A
EMI chocke: max current 1.5A
NTC: 15E 4A
input capacitor 2x47uF 400V
at 220VAC the peak current is (considering 10% main voltage) 265/15 = 18A. Datasheets recommend a peak current below the half of the minimun peak current (25A for the bridge, EMI has not reference of peak current in the datasheet). That means that 15E NTC is well suitable for this input design, isn't it ????
By the other hand, I think to put a 1.5A slow fuse at the input (NTC in a 220VAC input fase and fuse in neutral fase for example) in order to only protect all the input in case of several plugins and unplugins. If NTC reduce it resistance because heat, the fuse will be blow and the bridge and emi filter will be protected. What do you think about it ???
My principal concern is to protect the input stage avoidinf the destroy of all the components but at the same time avoid fuse blowing if not necessary, remember the full load input current is about 100mA RMS.
When I put a 1A fuse it blowed almost all the times at the pluging time (may be you're right and I use a fast fuse. Icould to see the peak current durantion and choose a fuse slower than this time. What do you think ???
best regards, Esteban
What is the value of input capacitor?
When the NTC heats up its resistance dramatically drops. So when you unplug the supply its resistance is low. The next time you plug it in (before allowing the NTC to cool down) it presents low resistance at start-up and does not limit the inrush current. This inrush current probably destroys the input rectifier and or the EMI choke.