TNY268GN Problem (Help)
Hi!
We are manufacturers of time, temperature and PLC controllers.We are experiencing a problem with one of our teams using the TNY268 for the switched source. The project was scaled with the help of PI Expert in the version that still supported TinySchwitch II. It is happening that surroundings of 10% of the equipments are returning with problem in the source, that is to say, the equipment does not connect more and in some cases the equipment turns on and after some time it turns off. In cases that do not connect the TNY268 is literally exploded. And in cases that turn off after some time the TNY268 overheats. Follow the schematic and the layout for analysis. I do not know what to do!
评论
Thanks for the answer! And sorry, my English is from google translator!
We marked in red in the esquematic the modifications we are thinking of implementing, we have some doubts before passing this to the PCB.
The main doubt is in P1, P2 and P3, are they correctly applied?
P1: We will not have physical space to implement fuse. But we were able to put MOV and Thermistor NTC 10R.
P2: Is it recommended the inductor in this position? Or change your position for M1?
P3: PIExpert does not support TNY268. Is TVZ 180V suitable?
P4: Let's implement 2M2-2512 (SMD);
P5: Are these values adequate?
P7: For space let's switch to a bridge with smaller encapsulation.
We know it is suggested not to use this TNY268 for new projects, but due to the stock we have we will have to improve this project and continue to use it.
| Attachment | 大小 |
|---|---|
| Esquematico (130.72 KB) | 130.72 KB |
Hi Visus,
Nothing to worry for your English. The translation is good.
Regarding your questions, below are my comments on each item.
P1: In any design, fuse is a mandatory component since it protects from excessive current in case of failure that can lead to fire. If you are limited with space, you can remove the MOV and then use the space for fuse. You have high voltage input capacitance anyway that will help clamp surge voltage. Just check if it will pass your surge requirement.
P2: Usually you need a PI filter (C-L-C) for EMI requirement so you basically needs additional cap after the bridge. You can use 10uF/400V then you can reduce your C12 to 15uF or just both use 10uF.
P3: 180V is fine. You can fine tune during testing.
P4: Single 2512 resistor is only rated 200V so you needs 2 in series to have 400V rating which is what you need.
P5: Typically good but you needs to test it if you have good regulation or you need to do some adjustment.
If you still have stocks with you, that is fine to deplete or if you needs additional parts, it is still available from distributors.
Thank you for the informations!
In this project we are really limited in relation to space, as you can see by the layout of the PCB. My last doubts.
P1: The fuse will be implemented a track on input AC1;
P2:
-The PI filter, here really is a space problem. Implementing a CLC filter is practically impossible. Since it is not possible to implement the CLC filter, the way we did with LC only helps in anything?
-Is there another way of implementing beyond the CLC, considering the space we have?
P3:
What should I consider to adjust this 180V parameter?
Does the TVZ have to be unidirectional or can it be bidirectional?
Are we predicting a PT6 PE6K180AC, any partnumber tip that is SMD?
P4: The insulation description of the manufacturer of resistors we use for the 2512 says it is 500V and the 0805 is 400V. So I believe 2512 is appropriate.
| Attachment | 大小 |
|---|---|
| Fig2_PCBLayout.PNG (105.85 KB) | 105.85 KB |
| Fig3_PCB3D.PNG (129.22 KB) | 129.22 KB |
| Fig1_2512.PNG (66.25 KB) | 66.25 KB |
Hi Visus,
You really have space constraint with your PCB. Regarding several points, here is my answer.
P1: Good, fuse is really important for safety.
P2: The CLC is basically for EMI performance. If you passed with just LC then it should be fine.
P3: Zener clamp is fine, TVS is more expensive. Your consideration is your VDS voltage that can damage your MOSFET inside and EMI if you are not clamping the peak voltage due to leakage inductance.
P4: If it is rated 500V then it should be ok.
Hi Visus,
Thank you for using Power Integrations' part.
Regarding your question based on available info that you provided me, this are my comments and possible cause of failure.
1. You basically don't have input protections. No input resistor at the inpul LI for inrush protection. It is usually around <10 ohm resistance in series with fuse. You also don't have MOV between L1 and L2 to clamp any surge/spike voltage. You also have big Bulk Cap of 22uF which really needs input resistor to limit stress on your bridge diode.
2. You should have at least two 1206 resistors for R29 to have higher creepage. This is high voltage across the EN pin and Bulk+ which potentially arc and damaged your IC. Looks like R29 is only a 1210 resistor which is just 100V rated.
3. If your VCC is 12V, your Z1 should be at least 1V lower for the optocoulpler diode of PC817 to have some room to operate and provide the feedback. Putting a resistor also in parallel with the optocoupler's diode helps lessen stress on Z1 and optocoupler itself to ensure better reliability.