B&K 9110 Power Supply Repair

[PwrElectronics]

Some months ago I found this B&K 9110 power supply in the recycling at work.  The cover was removed and placed in the steel scrap and the rest was in the mixed PCB scrap.

https://www.bkprecision.com/products/power-supplies/9110-100w-multi-range-60v-5a-dc-power-supply.html

It was small enough to bother taking home for a look.

Not a surprise, it was dead.

Some rooting around inside determined that there were several linear supplies on the main board and all were OK.  These were off a transformer and rectifier from the AC mains switch.  5V was coming up a ribbon cable to the display board behind the front panel.  Panel was dead.

No activity on the crystal.

This uses a TI MSP430F4250 processor:

http://www.ti.com/product/MSP430F4250

Some investigation on the display board found 5V coming into U101, a SOT-23 package device.  The SOT-23 had a electrolytic cap attached to 2 of the pins, 3rd pin to the ground plane.

I traced the power pins for the micro back to U101 and the cap not having 5V on it.  Given the specs on the micro, a good guess this is a 3.3V linear regulator.  I was suprised it would be a little SOT-23 but this micro is low power.  No 3.3V power though.

Ohmmeter found about 3-ohm across the supposed 3.3V power rail.

Step 1:  Removed electrolytic cap on the 3.3V regulator output to see if it was the short.  Short is not cleared.

Step 2 had to wait until I got a chance to borrow a IR camera from work one weekend.  I've successfully used IR cameras several times to locate shorted parts by injecting a current limited voltage into the shorted rail and looking for what is heating up.  That was the plan here too.

With the thermal camera in hand, I set my bench supply to 3.3V with current limit maxed and watched for the hot spot while slowly increasing the allowed current. 

Surprisingly, the hot spot was U101 regulator.

Step 3:  Removed U101 regulator and measured the resistance to ground again.  It was now about 27 ohms.  Still seemed low given the small regulator size.

Step 4:  Thermal camera again and applied the 3.3V with current limit looking for the next hot spot.  This time it was U103.  And, with 3.3V present on the micro power, there was still no oscillation on the crystal.

What is U103?  I was able to trace the /RST pin for the micro to a resistor connected to U103 and also that U103 had 1 pin going to the 3.3V power and the other to ground.  So, U103 probably the voltage supervisor aka power on reset timer.  The line going to the micro through a resistor was low and also voltage at micro /RST pin was low.

Step 5:  Removed U103.  27 ohm "short" is cleared.  I found the output of U103 had a 100k pull up to 3.3V and at least 2 other resistors heading off elsewhere.  One going to the /RST pin on the micro.

Step 6:  Powered up the board again with the bench supply on 3.3V.  /RST pin on the micro still low so this seemed like U103 supervisor was not a open drain output as the 100k pull up was not pulling up.  So, U103 maybe a push-pull output.

Step 7:  I soldered a wire to the output pin pad of U103 and another wire to 3.3V and positioned them so I could short them together.  I figured this may work but switch bounce or timing concerns it may not.  I had previously determined that other stuff on the display board was getting 5V and only maybe the micro was on the 3.3V. 

So, the LED power would be 5V.

I turned on the unit so the 5V would be there, then switched on the 3.3v bench supply, and finally shorted my reset wire to 3.3V.  Success!!!  The display lit up and the buttons, encoder knob, and beeper were all working.

Step 8:  What to do to replace U103?  U101 or some source of 3.3V no problem.  I got into checking my stash of old junk and found a left over prototype of something I did in 2005 that used a TI TMS320LF2406 DSP.  The supervisor/reset used was a TI TPS3809K33DBV.  This is a 3.3V supervisor in a SOT-23 package.  Pinout matched what I needed and it also has a push-pull output.

So, I took the '3809 off that board and put in on the display board.  Powered up everything again and success!  This time I also hooked up some load to the B&K output posts and checked that it actually worked, etc.  Voltage and current meters working; I could adjust the values.  It all looked good.

Step 9:  Get a 3.3V regulator to replace U101.  The proper repair here would be to get a 3.3V regulator in a SOT-23 package.  I didn't have anything like that in my parts stashes and as it turned out after turning over a lot of bins and bags of parts, no 3.3V fixed regulators at all.  What to do?  I don't want to make a order for a < $1 part and don't have anything else I need right now.  I have a lot of parts so I should be able to cobble something.  So far I have $0 invested.

So, I built up a little hack of a 3.3V regulator on a perf board and attached it to the display board.  I had noticed the accuracy of the meters on the display depended on the 3.3V supply so I put a trimmer on my regulator to set it exactly.

Everything all together and here it is putting out 20V into a 8 Ohm power resistor.