TOP261 getting very hot

Hi,

The output voltage at no load should be comparable to the output voltage with load for a well designed power supply. From your schematic it appears that the optocoupler polarities have been reversed. The optocoupler diode anode and cathode connections need to reverse and the optocoupler collector-emitter connections need to reverse as well. Without this correction, your feedback circuit is essentially open.

Because the output voltage is higher, there is excess dissipation in the snubber D2,R2 which could be the cause of your problem.

There can be several reasons why the device temperature is high such as a saturating transformer.

If the above change of the optocoupler does not solve the problem, please send us the design details of the transformer to help you further.

If you have not done so already, I encourage you to download PIExpert. This is a software provided by us free of charge and it will help you with the selection of the necessary components and their ratings.

Please use the link below to download PIExpert:
http://www.powerint.com/en/design-support/pi-expert-design-software

Please refer to our application note AN-43 which provides specific tips/recommendations for board layout and thermal design:

http://www.powerint.com/sites/default/files/product-docs/an43.pdf

Thank you for considering Power Integrations for your application.

Regards

PI-Sarek

it seems in fact that the overheating seems to be specific to only one of our 6 prototype boards, having switched to another one, the overheating problem seems to go away, but there are problems too

I had spotted the incorrect opto polarity and have changed that, I have also modified R4 to 3.4K and changed D1 to an MBR20100

The problem now is as follows :

We still see an off load output of around 39V, at this voltage, D5 is reverse conducting and the optocoupler is "on", we see a very low duty cycle on pin 7 of U2 as we would expect.

If I now apply a load across the output of the PSU in the form of a pot and slowly wind it down, we see the output voltage rapidly reduce. Once the output falls below 24V, D3 ceases to reverse conduct and the opto turns off. I would expect the duty cycle to now rapidly increase, but we do not see that, in fact the duty cycle remains low and the voltage on the output collapses. Even if I now ramp up the load, it fails to recover, and I have to remove the load completely for it to recover back to 39V

This seems quite unusual behaviour and I'm not sure where to start looking, I have attached the revised schem, transformer data and the UDS file (ETD39 168W TEXE Design Bosch 271109.pdf rename extension to uds to view, your forum won't let me attach a UDS file for some reason)to illustrate what we are looking at

Thanks for any help you can offer

Hi,

From the transformer drawing it appears that the transformer is connected like a "Forward transformer". For the circuit to operate as a flyback, if the dot end of the secondary is connected to the output diode anode, then the dot end of the primary should be connected to the MOSFET drain node. On your schematic it appears to be reversed. This explains why your power supply is overheating.

Also there is a significant mismatch between the transformer drawing and the schematic in terms of pin numbers of the transformer. The drawing shows pin 9,10 and 11 as secondary start end but the schematic shows pin 10,11 and 12 as the start.

Also from the transformer drawing it appears that the primary is split in two halves consisting of windings however your schematic shows pin numbers that show pins 5 and 6 connected together which would short the primary half.

It appears that your schematic and the transformer have a significant mismatch. You have probably corrected some of those connections so please send us the updated schematic.

If indeed your transformer has got connected in a forward configuration, please have it corrected and let us know if you still continue to experience any difficulties. If the dot end of the secondary is connected to the output diode anode then the dot end of the bias winding must connect to the bias winding diode anode as well and the dot end of the primary winding must connct to the drain of the MOSFET.

We could not open the transformer design file so we could not determine the primary inductance. If you designed the circuit using PIExpert, please send us the PIExpert files if you need further assistance.

PIExpert is a software wthat we supply free of charge and can be used in designing power supplies using our parts.

Hope this helps.

Regards

PI-Sarek

Hi, yes the transformer spec is for an older desing, apologies for that, the transformer pinout is correct on the schematic and is wired as you suggest, the file I attached is a PI expert file, but this forum will not allow me to attach UDS files, so I had to rename it as pdf, just change the extension back to uds and it will open fine

Hi,

Please try the following:
1. Currently you have the resistor R8 that is routed to a connector. Please disconnect the resistor and connect a resistor of the same value from the X-pin to the "source" pin directly.

2. Disconnect diode D3 from the circuit and apply 12V regulated DC voltage from an external isolated bench supply across C14 which is the filter capacitor for the bias winding rectified output. Turn ON the 12V bench supply before turning ON your circuit. This guarantees that the bias supply is stable. If under these conditions, your circuit maintains a tightly regulated output from no-load to full load, then it suggests that the output of the bias winding is not remaining in regulation and leads to circuit malfunction. The bias winding turns may then need to be adjusted or the bias winding filter capacitor size will have to be increased.

When you test the above suggestions please try them together and not one at a time.

Please let us know the result of this experiment. If you continue to experience difficulty, please provide us a copy of the layout of your circuit board.

Thank you.

Regards

PI-Sarek

Under normal circumstances, pins 1 & 2 of J2 are connected together anyway, so R8 is already connected to ground, however I have done what you said

I have removed D3 and and attached a bench supply, I have supplied 15V infact because that i smore representative of what we see from the transformer.

With no load, we see around 25mA drawn from the 15V supply, this drops to around 17mA just before the opto trips at 24V and from this point on, the current drain is as close to zero as makes no difference. In this mode, pin 3 of the TOP261 is effectively only connected to ground via C16 and R9/C17. In this configuration I would expect the see a high duty cycle on the switcher output of the TOP261 but we don't - can you think of any reason why this would not be the case ?

Hi,

The 25mA current draw from the bench supply is abnormal. During normal operation when the output is unloaded, the power supply requires minimum duty cycle for operation and at this time the control pin current is less than 7mA. Please see the datasheet page 7 for explanation.

As the load on output increases, the dutycycle required to support that load level increases and the control pin current drops. The duty cycle will be full once the control pin current drops below approximately 2mA.

As such there is no "tripping". This being a linear PWM control system , the control pin current changes as a function of load and input voltage. The only time when you will see burst mode of operation called the MCM mode in our datasheet is when the power supply is unloaded completely.

With the use of a bench supply in place of the bias winding output you should never be seeing a voltage higher than the setpoint(in this case 24V).

The exact cause is not obvious from the description provided by you. It may be best for you to communicate directly with one of our field applications engineers.

Please send me your complete contact information and I will have one of our FAE's contact you.

Please send the information by email to Application.Engineer12@powerint.com

Regards

PI-Sarek