SICK Gas Analyser failures

[Skogsgurra]

Hello,

I normally work with EMC and transients etcetera on the grid and when I recently was asked to find out why a flue gas analyser failed very often - it was assumed that the mains voltage was the culprit.

After having checked the supply, I can see that everything has been done to avoid overvoltage/undervoltage/transients but the equipment still fails and needs to be repaired quite often.

The device is a GMA 30 - 1 - 110 with 220 V 50 Hz 400 VA.

Anyone had this before? What did you do to avoid future problems?

 

[hacksaw]

How can you assume the details of failure?

Don't you have to piece together the facts first?

 

[Skogsgurra]

I do not understand what you mean, hacksaw. Did you read my question?

 

[hacksaw]

read it, please explain the basis for your assumption that mains interference causes failure.

 

[Skogsgurra]

I did not assume that. IT was assumed by the people that asked me to check the installation. It is a fairly normal assumption if an electronic device fails and one of the first steps in trouble-shooting is to check for transients on the mains voltage when the power supply fails without any obvious reason.

I am surprised. Are you the same hacksaw that I have learned to appreciate from earlier postings in other fora?

 

[analogkid2digitalman]

Have you done continous long-term monitoring/recording of the power to capture sags/rises/surges/spikes?

Is the repaired unit always returned to service or are new units always failing?

Not familiar with the equipment, just looking at the bigger picture.

Good luck!

Wheels within wheels / In a spiral array
A pattern so grand / And complex
Time after time / We lose sight of the way
Our causes can't see / Their effects.

 

[hacksaw]

the question still remains: what specific evidence of failure are you dealing with.

for example, what components are being replaced?

 

[Skogsgurra]

hacksaw,

I repeat what I said above: "It is a fairly normal assumption if an electronic device fails and one of the first steps in trouble-shooting is to check for transients on the mains voltage when the power supply fails without any obvious reason."

There is a line conditioner (MGE magnetic stabilizer) between mains and device. All grounding is OK. So there are no transients on the input. I admit that I have not checked this particular installation (about 20 yards up a smoke-stack) regarding transients. But I think that it is safe to say that they do not exist after this type of conditioner. (Done some recordings with fairly fast recorder - 500 kSa/s - on output during extreme input voltage transients and load step changes. Nothing there.)

The power supply failed. That we know. It has done so several times. I am not asking for general trouble-shooting advice. I am trying to get input on this particular device from someone that might have had the same problem with it.

 

[hacksaw]

the particulars of you failures can be useful in getting manufacturer cooperation, good luck,

 

[Skogsgurra]

I climbed the smoke stack once again. This time I brought a recorder and a pair of current clamps.

It turns out that line conditioner (magnetic stabilizer) output stays very constant and there are no spikes during turn on or off.

The stabilizer output decays from 230 V to zero after switch off. It takes about ten cycles to reach zero.

The analyzer input current goes up heavily during power down. The reason is that the switching power supply tries to maintain a constant output (delivers constant wattage) all the way down from nominal 230 V to about 20 V. It is during these last cycles that the input current increases and reaches about ten times nominal value to keep power constant (UxI=constant).

Considering that the switch is a power Mosfet, with a resistive channel, the losses in the switch is proportional to current squared. So the losses are about one hundred times normal losses for about ten milliseconds. No wonder it fails now and then. In my view, it is a miracle that it doesn't fail on every disconnect.

A classic example of something supposedly good (line conditioner) turning out to be bad. I have told the customer to replace that stabilizer with a simple filter and some varistors.

About contacting the manufacturer: In this case, they are all too happy to sell very expensive replacement PSUs and not at all interested in solving the problem. I guess that 2000+ Euros (abt 3000 USD) for a 300 W PSU is good business for someone.


Thanks for listening, folks. Sorry to have upset some of you with my dumb question. I really should have known better.