TNY280 Transformer problem

ali_asadzadeh -


As you're discovering, transformer design can be an artform :-)


A couple things I'm noticing from what you've told me so far:
-You are measuring your transformer inductance and leakage inductance at 10kHz. You'll want to remeasure this at the switching frequency of the device, in this case 132kHz.
-Before you start applying full input voltage and full load to your power supply, it's a good idea to power up the design slowly. You'll want to look at your drain voltage waveform and your output diode waveform to make sure you have your snubber/clamp circuitry setup correctly before really pushing the power supply to full-load.
-If you're still having problems reaching your full output current, use an oscilloscope and watch important circuit nodes (drain voltage, output diode, enable pin, etc) and see what's happening as you approach the point where output voltage starts dropping.
-TinySwitch devices have a fixed MOSFET current limit and a fixed frequency of operation (neglecting jitter). Output power then becomes proportional to Lpri x Ilim x Ilim x fs. Switching frequency and MOSFET current limit are not parameters you have control over, so your maximum output power is controlled by choosing your primary inductance. If your transformer design isn't right or you're not measuring its inductance correctly/accurately, it's possible you've reached the maximum output power for that transformer. Once the TinySwitch device reaches its maximum current limit, duty cycle, frequency etc, any increased load will cause the output voltage to fall out of regulation.


-The Traveler

Hi
I have still got no lock! But I have noticed that I have used a LTV817C device which has 200-400% CTR can this cause the voltage drop?
Do you have any other recommendations?

The opto does not solve my problem. I have tried to reverse the secondary windings. And used a variac to power the power supply, but when my input reaches 30VAC the TNY280 make some sound and smoke!!!
http://www.novinupload.com/do.php?img=11354

Do you have any idea!?

You can see my PCB files in PDF and altium designer here
http://www.4shared.com/file/xQCZCx-R/5V_3A.html

ali_asadzadeh -


As I said before, you need to go back to the basics.
Did you check your transformer inductance at 132kHz?
Did you check your transformer leakage inductance at the switching frequency as well as the turn-off ringing frequency?
Have you verified the correct winding polarity?
Have you used an oscilloscope to verify the drain voltage/current waveforms and make sure their correct in comparison to the output diode voltage/current waveforms?


If you're blowing up your device at only 30 Vac input, you've got some major problems going on, likely with the transformer construction. You really need to go back to the basics and start verifying everything.

-The Traveler

I have solved the problem. The problem was transformer. Now take a look at the output waveform, the ripple is 40mv and I have spikes in the range of 200mv.
http://0000.4.img98.net/out.php/i50984_imag0102.jpg

I have used the 3.3uH and 100uF output post filter. Do you have any recommendation?

How Can i remove the spikes in the output wave form?

Could you give me a little bit more information? The picture of your scope isn't that great...

Are you having problems with turn-off ringing with your output diode?

-The Traveler

I have an old scope and it’s the best picture that I could get from it! It’s like 40mv ripple and 200mv spikes decaying very soon. I did not get your point about turn-off ringing with my output diode. Would you explain it more?

Do me a favor and get some scope images of the following:
Drain to Source Voltage
Output Rectifier Voltage


Get these plots @ Vac(min) for 25%, 50% and 100% load. Repeat @ Vac(max).


Also, please send me a PDF of your PCB layout.
Remeasure your transformer inductance and leakage inductance at a frequency closer to 132kHz.


-The Traveler

Dear Traveler
This is the Drain to Source Voltage @ 2.7A
http://www.imeezo.com/v/images/83044108129214062677.jpg

And this is Drain to Source Voltage @ 1.47A
http://www.imeezo.com/v/images/05707436604770069683.jpg

And this is Drain to Source Voltage @ .74A
http://www.imeezo.com/v/images/67422588362358241055.jpg

This is Output Rectifier Voltage @2.7A
http://www.imeezo.com/v/images/42978417240388277259.jpg
This is Output Rectifier Voltage @1.47A
http://www.imeezo.com/v/images/75591331767173110672.jpg

This is Output Rectifier Voltage @.74A
http://www.imeezo.com/v/images/40122179501109351649.jpg

I could not measure values on maximum because I do not have a high voltage probe!

And here is the PCB layout file in PDF
http://www.4shared.com/document/p2Xx4bMs/Power_supply_5v_3A.html

The primary inductance @132KHz is 442uH and primary leakage inductance is 14.5uH@132KHz

Thanks for your time.

Ali -

Looking at the layout you sent me...I think it's the cause of probably 99% of your problems. There are some SERIOUS concerns with this layout.


I've attached a copy of your layout that I marked up alongside a markup of one of our existing designs RD-91 (DER-228).


The green loop that I drew shows the path of the primary switching current. This loop contains pulses of current at switching frequencies (fairly high frequency) and the loop area needs to be kept as small as possible.


The loop shown in red is the path of current in the clamp circuitry. This loop contains even higher frequency noise and needs to be kept very very small. The clamp circuitry in your layout is so large that you will likely run into numerous conducted/radiated emissions problems. the long traces in your clamp circuit are adding additional trace inductance which makes it even harder to control mosfet turn-off ringing (the high-frequency ringing that is in all your scope plots).


The blue loop is the path that high-frequency switching current on your output follows. You also want this loop very small.


The area in red is the safety isolation barrier for this design. Even with the opto in your design, you don't really have any isolation barrier which is going to cause you some regulatory certification problems.


The area in orange is PCB copper area used to provide heat sinking for the PI device. You need significantly more copper area to keep your part cool.


At this point...I would recommend going back to the drawing board. Your PCB layout has so many problems that it would be incredibly difficult to fix. We provide many many reference designs to aid our customers in designing their own power supplies. The PDF I attached for DER228 is a design that we've already completed and has an output power pretty close to what you're using. It would be fairly easy work to change the output voltage/current ratings and just reuse 90% of the layout. No need to reinvent the wheel....


Please let me know if I can be of any more assistance.

-The Traveler

Thanks a lot for the PCB tips.
I have redesigned the PCB, Would you take a look at it, and inform me about my routing?

http://www.4shared.com/document/nbCLDKK7/Power_supply_5v_3A_Rev15.html

Regards

Ali -


It definitely looks better than before. You could probably make your drain current loop a little bit smaller if you rotated the device 90deg CCW and then used the clamp diode to hop over the B+ trace from your cap.


For your secondary side components...What's going on with all that ground plane? Ground plans can be useful...but if you just fill every available square inch with copper it can cause more problems than it solves.


Get your output diode closer to your transformer and route the traces in such a way that the high frequency content of the output is limited to a small area comprised of the output diode and C2/C3. The way you have it laid out now, this high frequency noise is circulating through much of your ground plane. You need to move R6 so that it's closer to your bulk caps.


You're definitely moving in the right direction. I'd really suggest going through some of our reference designs and make sure you understand why components are placed where they are and why things are laid out the way they are. Don't just think of things in terms of DC. Think in terms of switching frequencies, loop area, ringing frequencies and parasitic trace inductance.

-The Traveler

Dear PI-Traveler
I have ordered the new PCB, but still I got the spikes in the output. They are about 300mv when the load is about 3A.Do you have any idea?
Can the output capacitors cause this problem?

And this is the new PCB layout.
http://www.4shared.com/document/pfTNJM-C/Power_supply_5v_3A_2.html

Regards