Possible SMPSU fault.

[ke5frf]

Hello.

I have an issue with a piece of equipment I was asked to troubleshoot today at work. I was preoccupied with another issue I was working on but that is another story (which I might discuss too haha). Long story short, I gave it a quick look over and noted the fault symptoms.

Anyhow, the equipment is an Agilent Technologies Micro GC (gas chromatograph)
https://www.electro-tech-online.com/custompdfs/2009/11/3000microGC_especi.pdf

GCs are electromechanical, temperature controlled lab equipment with analog sensors amplified and converted to digital with signal voltage integrated vs. time axis. Acquisition software on a computer retrieves the real-time data and creates files.

It is a portable apparatus, so for flexibility the power supply is external giving the option of battery power in certain applications. My facility has only ever used the power supply, no batteries available.

The supply is a switcher of good quality, black box type with AC cable and a 4 pin DC cable with a keyed connector. 15 VDC, 6.6 amps. Small cooling fan built in with vents.

The fault condition was no communication with peripheral PC. A secondary symptom was the power LED on the main switch, the operator mentioned that it seemed dimmer than usual.

I removed the cover and found both motherboards (identical boards for two-in-one testing) were powered up, board mounted LEDs indicating operation.
I depowered the instrument to visually inspect it, and unplugged the power supply. I found evidence of excessive heat, the external label was discolored, brown and gray like it had been cooked. I reconnected it and applied power, and felt vibration. Inspection confirmed the cooling fan was binding, turning slowly, stalled.

I checked voltages with my multimeter:

Found 15 volts as expected. Checked ripple...intermittent evidence of .5 volts of AC. Checked frequency, 24 Hz???? Huh?

I know switchers use high frequency to reduce transformer size. I wouldn't necessarily expect 60 Hz. Wasn't sure why 24 Hz would show up.

I ordered a new supply from Newark, but this kind of puzzled me.
I might expect ripple to cause noise in the communication to the PC, so I fel I might have found the issue, but I was hoping someone with more knowledge could confirm that 500 mV of ripple would be out of the ordinary for a SMPSU, and explain why 24 Hz might show up and that this would be disruptive on a serial port cable.

Thanks in advance.

 

[mneary]

24 Hz could be the rate at which the overload/over temperature circuit checks back to see if it's safe to turn back on.

 

[Blueteeth]

With switching power supplies...who knows lol I've had an old TFT power supply slowly ramp voltage from 0 to 12V in a perfect sawtooth over a period of 3 seconds. The main transformer had a short, as I could hear it hiss/whine...'why' it still worked but in such a bizare manner I don't know.

Just out of interest, what did you check the frequency with? many 'basic' frequency counters or the ones in multimeters have a lower limit, which it seems to idle at when thers any input noise. My cheapy one dances around 43Hz. As mneary said, could just be the over-temp circuit kicking in, and holding it in power down for < 42ms. I would scope it

 

[ke5frf]

Well, I'm using my Fluke, not a cheapy meter. I know what you mean though, all meters will sometimes pick up noise, but usually I see 60 Hz whenever I see noise. Seems like I remember troubleshooting a switching supply years back and I saw 120 Hz when a capacitor had gone bad. Yeah, they can be tricky boogers.

Unfortunately, my scope isn't battery powered, its an old Tektronics scope and I don't have an isolation transformer for it. Generally speaking I don't often enough have reason to scope anything because I usually get things with the meter or else we have priority on getting the equipment running and just replace boards.

 

[ke5frf]

I just checked the supply out again this morning under the load of the equipment, after having cooled off overnight, and the equipment was able to communicate with the computer. The fault is obviously intermittent and triggered when it overheats. Perhaps it is the overtemp circuit polling at 24 Hz, that sounds like a reasonable explanation. I'm also considering that the stalled fan (12 V) is pulling the supply down and causing just that much more heat. My check this morning was without the fan connected so the load was less. Without the cover on the audible whine of the supply was noticeably loud, so I'm thinking something has become damaged that only shows up when the supply overheats.

I'm not worried about it now. This supply is garbage and another one will be here by tomorrow. My main concern was identifying the source of the equipment malfunction and having tested the supply this morning I'm confident in the diagnosis.