Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

60W offline flyback PCB suffers burning

Status
Not open for further replies.
T

treez

Guest
Hi,
Our contractor has done a prototype 60W offline single stage, PFC’d Flyback (240VAC).
The primary FET is a metal tabbed TO220 and is directly screwed to an aluminium heatsink (no insulating pad used) . The heatsink is soldered into the PCB via two pins. The heatsink is obviously at drain node potential.

We sent the PCB for testing where it was put in a chamber at 70degC and run at max power and 190VAC input. After some time, the unit failed.

The unit was returned to me and I inspected it, to see what had gone wrong. The 3.15A input fuse had not blown, but a 3W, 2R7 inrush resistor just downstream of the fuse had blown open. There was a strange tarnishing on the PCB beneath the primary FET…however, the FET itself looked spotless. This tarnishing did not seem to be associated with any component, but was disappearing under the heatsink. Also, the primary FET G,D,S did not measure short to each other.

Anyway, I then replaced the 2R7 inrush resistor, and then powered the unit up , starting gingerly at zero VAC , and slowly raising it……at around 60VAC input, there was a fizzling sound, and so I immediately turned off the power source. I then found that the primary FET was short circuit between its pins.

Therefore, I replaced the FET…and decided to see what the tarnishing was that was going under the heatsink. As can be seen in the attached picture of the PCB layer 2, the PCB was burned almost right through at the heatsink pin. (it’s a 4 layer 1.6mm thick PCB). Anyway, the burn crater had black burned carbonisation in it. Therefore, I scraped out the black carbonisation from the burn crater, and then blew away the carbon dust, to clean the crater. ..

Then, with the new PRIMARY FET in, I again powered the unit up on the AC source. This time it worked fine…no problem at 240VAC.

When at 190VAC and max power, the FET drain node goes up to 580Vpk. As you can see, the primary ground is only 0.25mm away from the plated through hole that contains the heatsink pin where the burn crater is. However, there are 3 other PTH’s where this also applys, and they didn’t suffer burning.

So anyway, do you suppose that the burning may be anything to do with flashover of the primary ground to the drain node copper at the heatsink pin? We are all agreed that 0.25mm is not enough clearance for 580Vpk on an internal PCB layer. However, there was no flashover at the other pins with similar clearances.

At our own premises, we have run this unit in the chamber on many occasions for 8 hours at 190VAC input, internal ambient of 70degc in the PSU enclosure, and max power and did not get any of this PCB burning, or any unit failure. We recorded FET case temp of 100degC.
 

Attachments

  • PCB layer 2 _.jpg
    PCB layer 2 _.jpg
    287.9 KB · Views: 122

blowing the resistor open sounds like a joule stress issue, not a regular heating issue
(my estimates indicate the resistor should be okay under average conditions)

joule stress is a result of passing too much energy through the resistor
it happens very fast - too fast for the resistor to heat up, it just explodes
(this is not uncommon - trees struck by lighting sometimes explode)
i did a lot of calculations of this for resistors designed to current limit capacitor discharges

i had a table from Dale Resistor that listed standard resistor types (mostly US military through hole styles)
and how much energy (joule stress) they could handle
Since Vishay now owns Dale, I looked at their website

this link is a pdf on pulse handling
https://www.vishay.com/docs/28870/pulseloadsmdlimit.pdf

this is a limited usefulness video
it compares a metal film resistor and a wire would under a less than dramatic demo
the wire wound, being bigger, heavier and with more metal to carry the current and power survives

https://www.vishay.com/videos/resis...ulse-handling-capabilities-demonstration.html

even better is a non-inductive version with an Aytron-Perry winding,
such as a non-inductive RWR type - wire wound, right helix and left helix, welded at each crossover
twice the wire for the same resistance, much better at pulse handling

this may be useful:
https://www.google.com/url?sa=t&rct...396_1604.pdf&usg=AOvVaw2RZoCcn7E0WcOjejHgLUiL

as for the burning and possible flashover, I think you want to increase the spacing
i dealt with a problem like this - a diode bridge rectifier in a transformer oil bath
would sometimes blow up - the problem was the a spacing issue and a wetting issue
the arc apparently went along the surface of the circuit board, between the board and the oil
(hence the wetting issue)
we eventually solved it by cutting a small slot in the board, so the arc had to go through the oil


https://www.google.com/url?sa=t&rct...spacing.html&usg=AOvVaw3dPzsMeKkxSQ7CoXdBl1Ig

https://www.google.com/url?sa=t&rct...earance.html&usg=AOvVaw0pKbi28cCHLQ0bryGjQbDS

my apologies if the links aren't quite where they ought to be, per the text
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top