Repairing EMC Power Supply Display

[Aragorn_84]

Hi folks.

Work was recently having a clearout of broken/old electrical equipment, and one of the items was a 2U rackmounted 2.5KW power supply made by EMC.

The power supply itself appears to work fine, the reason it was going out was the voltage display on the front panel is broken. The power supply otherwise works as it should.

I'd like to fix the display, so i've set about doing some testing to try and see whats going on.

There are two almost identical displays on the front of the unit, one for current, one for voltage. I first off swapped the liquid crystal display between the two units, confirming the display itself is good. On close inspection i see the voltage display does come on, its just very dim (you need to REALLY squint at it to see it) and is displaying "-1" rather than a reading.

I've then spent some time looking at the inputs to the display board, which it turns out is pretty simple, a 30vDC supply (labelled as +15v and -15v on the mainboard, but they dont seem referenced to chassis ground) then a voltage input. The voltage input is 1:1 proportional with the main output voltage.

The board appears to have a TC7126 ADC which has all the driver components build in to run the LCD segment display. The TC7126 has a maximum supply voltage of 15v, so i figured the board must be doing something to reduce the 30v input voltage down to a lower level. I measured the voltage at the chips V+ and V- terminals on the working display and got just under 9v. However on the broken display, i measure only 2.8v.

Theres no fancy switching components on the board, its all just passives. From what i can tell, both input rails are fed directly into a 10k resistor then go from there directly to the microchip ADC. Measuring the output of these 10k resistors gives the same 2.8v reading that the chip is seeing.

I'm struggling to see what exactly could be wrong with this thing, so i've taken some photos of the board and i'm hoping someone can give me some pointers (i've removed the LCD for extra clarity).

https://www.dropbox.com/s/9cggcn9ub0rl2uo/2015-06-13%2014.47.10.jpg?dl=0
https://www.dropbox.com/s/61uafvtho5igkh8/2015-06-13%2014.47.23.jpg?dl=0

The positive rail comes in on Pin 3, and goes directly to R5. The negative rail comes in on Pin 1 and goes directly to R9.

The Capacitor C3 is missing in the photos becuase it was reading funny while installed, so i removed it for testing. Capacitance checked out OK once removed so i've now refitted it. The other electrolytic tested OK for capacitance in situ. Oddly, C3 on the other display tested fine in situ. A pointer perhaps?

I realise i could simply get a cheapo voltage display from ebay and clag it in, but i'd quite like to actually fix it properly.

Cheers
Kevin

[Aragorn_84]

Oh and a Quick followup, i tested the two Zeners CR1 and CR2 on each board (still in circuit).

With the meter in diode test mode, on the good board, CR1 gave no result one way and 0.6v the other. CR2 gave 1.9V one way, and 0.6V the other.

On the bad board CR1 measured 0.26v in both directions, and CR2 gave nothing one way and 0.6v the other.

I cant make out the readings on all of them as some are positioned with their numbers hidden, but CR1 on the bad board is a 1N47, and CR1 on the good board is 1N33. Clearly the boards are configured slightly differently as ones measuring voltage and the other is measuring current, but could CR1 be shorted? or could something else be giving an erroneous 0.26v reading?

edit: Ok so desoldered CR1, its actually a 1N4733A, presumeably meaning they're both actually the same between good and bad. Its defo shorted showing 180 ohms across it in either direction.

[Aragorn_84]

Ok, so i replaced the Zener with another 5.1v zener powered it all back up and bingo, display comes on and voltage measurements are displayed.

However theres something wrong with the voltage divider network on the display panel, as its showing 6.2v at 40v output, and 3.1v at 20v output, scaling linearly down to 0.

Need to study the data sheet and figure out what its doing.

[Aragorn_84]

Quite confused now!

The voltage divider is perfect, its taking in 40v, and dropping it to 82mv, exactly as the resistor values suggest it should (100k and 205)

After much poking around, i've found that the "VRef" input into the chip is sitting at 4V. I removed the resistor that feeds the VRef from the V+ rail and the Vref remained at 4v, ie 4v coming OUT of the chip! Theres also a fairly low resistance measurable between the V+ pin, and analog (input) ground and V+ and Vref.

The good board measures 240k between v+ and analog ground, and all of that comes from the Vref potential divider which sits across V+ and Analog ground (220k resistor and a 20k pot).

On the bad board, with the potential divider disconnected by removing the resisitor, i'm measuring 18k between VRef and analog ground. I'm also getting 260 ohms between V+ and Vref on the bad board.

I'm now suspecting that the main IC is faulty. I've gone over all the passives and they all appear to check out fine. Nothing else could cause these odd readings that i can see.

I can only imagine there has been some sort of voltage spike which has blown out something inside the chip, as well as blowing the zener on the board.

Anyone have any thaughts? Otherwise i'm going to try ordering a replacement IC, they're only 5 quid at RS.

[jjohannes]

How did it turn out?

[Aragorn_84]

Replaced the IC and that sorted it.

No idea what caused the fault though!

[jjohannes]

Great!