Hi all, I am working at a electr

[hhhansen]

Hi all,

I am working at a electric utility company. We currently have complaints from a client claiming that the grid is subject to electrical disturbances.
The problem of the client is that a number of industrial PCs suddenly shut down (in some case only 3 in other cases all 16 PCs at the factory) apparently due to some sort of transients in the supply voltage. The client does not have any UPS installed, and he has never experienced any problems before with shut downs.
We have installed measurements on the substation transformer supplying the client and the results from the measurements indicate that the voltage quality at the PCC is fulfilling the EN50160 std.

Rather then just telling the client that the problem problably is related to EMC matters we have decided to help him solving this case. Thus we have carried out a number of measurements of transient incidents leading to total shut down of the computers. One of such measurment is shown attached. The voltage is seen to drop very fast from 272,5 V to 257,4 V within only 70 mikroseconds as indicated by cursor values on the legend.

I am not able to explain why this limited but fast voltage drop lead to shut down of the PCs. The supplier of the PCs has stated in the specification sheet that the PC fulfill the emsssion standard. This was EN50082-1 at the time of the delivery of the PCs. Now this generic std. has been superseeded by EN61000-6-1. According to EN61000-4-11 referenced in EN61000-6-1 the voltage dip is allowed to be 30% within 10 ms. Also regarding fast transients the allowance is +/- 0,5 kV, 5 kHz. However I am not sure if the values are the right one for complience evaluation.

My question is: Which requirements in the standards are relevant for comparison of the fast voltage drop observed in order to tell whether the PCs are behaving strange or not.

Best Regards

Hans-Henrik

 

[Skogsgurra]

This doesn't look like a power problem at all.

I have had similar cases in rolling mills, paper mills and other such industries where the bus between the computers was heavily disturbed. Those customers had other problems as well, not just computers shutting down.

It is very commendable that you decided to help the customer. And not so common. Have you had a look at the bus and transducer connections? A scope at 1 us and a high-pass filter that blocks frequencies below around 100 kHz is very useful for such investigations. Transients on the N conductor can also cause trouble and such glitches usually don't show up at the PCC.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.