Low temperature problems with TNY267
Hello PI profesionals.
I'm fairly new to the design of the power supplies. And recently I encountered 2 problems.
First and major is that at below -35C we encounter serious noise and lower voltage issues. Please see my scope photo. Our power supply must work in the range of 80-480V, and temperature range must be from -40C to 85C. Our output is 3.3V upto 2A, which works.
Second problem is EMI. Power supply needs to be cheaper but X1 capacitor C81 is very expensive, about $2.5 a piece and it barely does the job. May be you guys can give a suggestion on how to solve these problems?
Thanks.
Comments
I've definitely seen temperature problems with cheap ceramic capacitors in the past. A handy tool (for future problem solving) is an inexpensive can of freeze spray and a heat gun. If you're noticing odd power supply issues that are associated with changes in temperature, cooling/heating different components individually can sometimes point you in the right direction.
In regards to your EMI issues, it really depends on what frequencies you're having trouble with.
- If the trouble frequencies are in the neighborhood of your switching frequency, you might need to add some differential filtering. If your Xcap solution is too expensive, you might look at some inexpensive inductors.
- If the trouble frequencies are fairly high (up in the MHz territory), it's probably a common mode noise issue and I would suspect your clamp design, your output rectifier, or both.
- An easy way to check and see where the noise is coming from is to use a high-voltage scope probe and look at your drain voltage during turn on/off as well as the voltage across your output rectifier. If you see ringing at or around the frequencies you're having problems with in your EMI scan, you've likely found the culprit.
- We have a feature built into PI XLS and PI Expert for the implementation of shield windings in your transformer. Using shield windings can be extremely effective at solving common mode noise problems.
I hope that helps. Please let me know if you have any other questions.
-The Traveler
Traveler.
My noise is in the switching range and little higher up to 300kHz. After that everything is fine. May be you could suggest where should I install them and what size those coils have to be.
Thanks
Definitely isn't the first time I've experienced some cheap ceramic caps that don't work right when they get cold!
For the EMI issues: Before you start adding decoupling caps and differential inductors all over the place, take a few minutes and track down the source of the noise in your circuit. I would try running the power supply with a resistive load on the EMI tester and see how it does. If EMI is fine in this setup, your problem is likely related to the microcontroller and other power switching components.
-The Traveler
Hi Traveler. Thanks for your post.
The problem I have in my circuit is that secondary is not isolated from primary (that is the way electrical meter ICs work). I tested my circtui the way you have suggested with variable resistor. When I isolated secondary, there was a little noise, and if I added load, noise level would go up but not by much. However if I connected back Secondary to primary, noise level would go up significantly, and with load it would go even higher. I keep X1 cap in the circuit wich makes it expensive, it keeps noise just a hair below FCC limit, but if I add extra load it may go over the limit. So noise is coming from secondary, and secondary connected through the choke to mains. I'm thinking about increasing Inductance on the choke also, may be that will help too.
Hi Traveler. Thanks for your post.
The problem I have in my circuit is that secondary is not isolated from primary (that is the way electrical meter ICs work). I tested my circtui the way you have suggested with variable resistor. When I isolated secondary, there was a little noise, and if I added load, noise level would go up but not by much. However if I connected back Secondary to primary, noise level would go up significantly, and with load it would go even higher. I keep X1 cap in the circuit wich makes it expensive, it keeps noise just a hair below FCC limit, but if I add extra load it may go over the limit. So noise is coming from secondary, and secondary connected through the choke to mains. I'm thinking about increasing Inductance on the choke also, may be that will help too.
I was getting your post mixed up with another post....my apologies.
If you can, could you attach any of your EMI plots?
It sounds like you might have a common-mode noise problem. It might be a good idea to take a look at our E-Shield technology. It's a fairly simple way of adding an extra winding or two to help nullify the primary-secondary parasitic capacitance.
This link does a pretty good job of explaining the method:
http://www.powerint.com/en/community/documents-quizzes/puzzler-8
-The Traveler
Turned out to be bad capacitors they installed in China.
As soon as it would hit -35, capacitance would fall 100 times. Installed good caps, and everything is fine. I'm still researching better EMI solution to getrid of expansive X1 cap.