I have a nuisance problem in one of the test rigs that is now umder commissioing. This rig has a multitude of automated functions and is fully controlled by a 89c51rd2. No problems here and even the interface to the 15kW AC drive and other high power heaters like 10kW are extremely reliable - not a single problem in a 50,000 cycle endurance and still counting.

The operator interface has some 9 displays all with AT89c52 MCUs that are housed in the same panel above the AC drive. No issues here.

I am mentioning the above to give some idea as to the general effectiveness of the provided shielding and opto-isolation methods.

But the problem lies with three signal conditioners that handle strain guage sensors ( for pressure, torque and load ). These are housed very next to the relevant sensors and the +/- 10V dc output alone is brought to the DPMs that are housed in the panel ( where the AC`drive is mounted ) The cable is a two core shielded one with the shield connected to instrument ground inside the panel.

The problem :

As soon as the AC drive is started, a fixed value of the order of 0.6V appears on the signal conditioner outputs. All further readings are additive to this and relatively linear !!

I am going nuts trying to figure out how this loading happens. The following have been tried :

1. The 0.6 v loading happens only when there is a wire connected to the signal conditioner output AND the AC drive is ON. If we remove the wire the loading vanishes. And the loading is very consistent when it happens. The loading happens even when the output terminals are touched with a 1m long length of loose wire - not connected to anything.

2. The signal conditioners are fed by a 24VDC SMPS (MeanWell)that feeds other control relays. Changing this supply to a separate one did not help.

3. All means of shielding the output cable - one end; both ends. no end - all have been tried in vain.

4. Ferrite beads in 24V supply input and / or at signal outputs have not helped.

I am giving the schematic of the signal conditioner. It has three port isolation as can be seen :



This is an ADAM 3016 conditioner from Advantech / USA. I have used many such modules in the past and no problems.

When all the MCU based units are fine why should these modules get into this mess ? I am willing to try any trick - including switching on power after saying "Abracadabra" three-and-half times.

Thanks for your time.

Raghu