HI

I was Wondering if anyone has experienced noise pick up on an RTD sensor. The sensor application is for measuring the winding temperatures of a generator stator.

When the machine is shutdown the readings are steady when the generator gets up to synchronous speed all temps bounce +/- 20 deg F.

I have wired an external sensor in at synchronous speed and readings were steady so i know its not noise generated in other parts of the loop and is internal to the stator winding temp sensors.

sensors are wired to a Woodward linknet 6 channel RTD module.

Any suggestions would be most welcome.

 

My guess is all 3 leads of the RTD are following the potential of the adjacent coil winding.

If the RTD has a shield wire I would try grounding that to the frame of the motor.

If it doesn't have shielding try inserting a fairly large capacitor between one of the leads and ground.
Perhaps isolate the shield at panel end and ground it in the field.

Let us know what happens.

Roy

 

Does the Woodward Linknet module have any kind of filtering, such as a 50 or 60 Hz filter which you could enable?

 

Liam... is this a first-time commissioning operation?

Regards,
Phil Corso

 

Hi CSA

iv been over the manual a few times and there's no mention of filtering for the RTD module.

I have some capacitors of various values which i will experiment with today to see if we can get any improvement.

I ll post back with results.

 

HI Phil

Our units have been in Operation for over a year now.

I have also seen this problem with other units at other sites but it was never a problem for previous customers.

 

dear,

we are also facing similar problem. when machine is not synchronised, temperature values are steady. however, if machine is synchronised, values are fluctuating like a sine wave (+/- 3 deg c)

one experiment we have done. we have connected one Temperature convertor ( rtd/4-20ma). output current of this converter does not vary. but if we connect to i/o module, readings are fluctuating.

we have tried to enable 50/60hz filter available, but of no use.

can the readers help us with their valuable suggestions/solutions to overcome this problem.

thanks in advance
chiru

 

In the drawing below, I've attempted to draw the scheme that GE typically uses for connecting generator RTDs to the monitor/control system. They typically use two-conductor, twisted, shielded cables to connect two of the three RTD leads to the monitor/control system. They ground the third RTD lead in the junction box where the terminal screws are located, and then use one or two conductors of a separate twisted, shielded pair as the third leg of the RTD by grounding the wire(s) in the junction box very close to the grounded RTD lead, and then connecting the wire(s) to the monitor/control system and jumper the third monitor/control system terminals together at the monitor/control system. (So, all the RETurn terminals in the Speedtronic would be jumpered together, and there would be one or two conductors connecting the jumpered terminals to ground in the JB at the generator.)<pre>
Screw
Terminal
-------X-------------------------- EXCitation
| ^
----- |
| | |
| | These conductors must be twisted and
| | shielded, and the shield drain wire must
----- be grounded at ONE place only (the shield
|\ drain wire is not shown in this drawing)
| \ |
| \ v
| ---X-------------------------- SIGnal
|
----- --------------------- RETurn
| |
----- -----
--- ---
- -
These two grounded conductors must be in the same
Junction Box in close proximity to each other.</pre>
I've long had the feeling that there was more to this wiring scheme than just saving money by using two-conductor, twisted, shielded cables (twisted, shielded pair cables) instead of three-conductor, twisted, shielded cables. And, providing some kind of "filtering" by grounding one of the RTD leads to help reduce induced electrical noise may just be that reason. I asked many people in GE about this over many years, but could never find anyone who could provide any answer other than, "That's the way we've always done it!" I may be wrong, but I suspect there may have been multiple reasons for using this wiring scheme--and this just might be one of them. I'm sure some electrical guru reading this will have some input; I hope so. If I'm wrong, it won't be the first time.

This wiring scheme confuses the HECK out of most GE field service people, GE engineers, and Customer technicians and -engineers. GE still uses 50- and 60 Hz filters in the Mark V and Mark VI and Mark VIe Speedtronic control systems to further reduce induced electrical noise, but the earlier control systems had no such filtering and worked reasonably well--and I believe the grounding scheme was one reason for that. In the RTD monitors/control systems used by GE, the SIGnal and RETurn leads are used for the lead compensation method. The fact that one of the RTD compensation leads is grounded in close proximity to a grounded conductor used as the second compensation lead does little to affect the reading/compensation, if at all. The thing that most people fail to do or to recognize is that at the monitor/control system, the third lead terminals must be jumpered together for the scheme to work properly.

I've attempted to draw the GE monitor/control system end of their scheme below, The RETurn terminal of the Speedtronic in the upper-most RTD input is physically wired to ground in the JB at the generator. The second and third RTD inputs below that do not have a wire running to the generator JB, but do have jumpers connecting them to the wire of the first RTD input, which completes the compensation circuit for the second and third RTD inputs.<pre>
--> >--------EXCiation

--> >--------SIGnal

--> >--------RETurn
/
/
|
--> >--------EXCitation
|
--> >--------SIGnal
\
RETurn
/
/
|
--> >--------EXCitation
|
--> >--------SIGnal
\
RETurn
/
/
|
v</pre>
So, if your RTD monitor can accommodate grounded RTDs, I would suggest trying this with one RTD to see if it works. If your system is using three-wire, twisted shielded interconnecting cables, you can try just adding a ground to one of the common RTD leads as a test.

Please write back to let us know how this works for you--AFTER you confirm the RTD modules can accommodate grounded RTDs!

 

That was worth a star just for the ASCII artwork.

The wiring confuses me, too. Thanks for posting what it is, though.

 

Liam... can you provide information regarding the interfering signal, that is, when it occurs (during start, under load, etc), what is its appearance (transient, sinusoidal), and how long does it last (continuous, intermittent, random)?

Another question, does the interference occur on one, all six or 12 sets (if 2 per location)?

Phil

 

You might also consider contacting Woodward at their Fort Collins, CO, offices. Their technical support is pretty good. They might have a solution for you, because it's pretty certain this isn't the first time they've heard of this problem.

 

it's a bit after the fact, but you require non-inductively wound rtd's, and all leads properly routed and twisted, a number of motor manufacturers have specific details of the rtd construction.

once you have induced noise, filtering is about all you can do.

 

Hi CSA

Mega post ! much gratitude for that.

So as for Grounding the RTD, I'm not sure if the input card can accommodate that arrangement. i know in some measuring circuits (voltage dividers and some constant current types the current flows through the RTD element to earth). I've checked across all 3 RTD input terminals on the card and there is no continuity to earth so perhaps not?

On further investigation I found there is no shielding for the Stator RTD's (but there was for all the bearing RTD's?). so its no surprise to be seeing such a noisy signal.

I believe I had some limited success installing 2 filtering capacitors, one between the excitation line and Ground and one between the Signal/Return line and ground.

There was definitely and Amplitude reduction with the capacitors added with maximum swings of 12 Deg F rather than 20+ which we were seeing previously .

I Have posted some pics pre and post Capacitor install for those that are interested the link below should take you there .

https://flic.kr/p/czcUUW
Capacitor values were 10 Microfarads (its all i could get my hands on at short notice) but i now have some 100 MicroFarad's which i can try when we run the machines up again next week which im hoping will do the trick!

off course i shall keep you posted.

Many thanks.

 

Hi Phil

We've got limited test equipment on site so i haven't been able to scope the signal.

What I can say is that the noise is common to all 12 RTD's and is continuous and is small in amplitude at synch idle and then gets much larger as The generator is loaded.

 

Liam... unfortunately without knowing the noise “type” all you can do is trial and error. I strongly urge you to borrow, beg, or in some other(?) way procure a scope.

Phil

 

Liam... do you have a Multi-Amp (or similar) Relay Tester on-site. Something that can deliver, say, 100-300 Amperes of current?

Phil

 

Liam Heslop,

Thanks for the info.

I have seen similar unshielded wiring used for RTDs (and T/Cs) in generator applications, and have always wondered what the people who spec'ed that were thinking.

I have no experience with the Woodward linknet modules, but I would presume the instruction manual would say if it could accommodate grounded RTDs.

Again, I have had good experiences contacting Woodward technical support in Ft. Collins, CO, USA, for help.

I have a couple of last questions: Is this a new application of the Woodward linknet module on an existing generator? If so, what was used previously?

Again, thanks for the feedback and info! It's what really makes the posts valuable for a very long time.

 

A very good diagram.

I think you are probably correct, ground out the noise at the source.

I suspect at the panel end you would want to leave everything floating i.e. might want a separate power supply just for the card and leave the shield off at the panel also.

Roy

 

Hi Liam,
The problem is in the Brush generator.

RTD wiring inside generator housing is unshielded single wire brought out into a junction box on the generator. From there signals are wired with twisted pair shielded wire to the Linknet module's (as per Woodward spec.)

Solution would be to replace wiring all the way with shielded twisted pair, unfortunately winding RTDs aren't accessible so software filtering is the only solution until you open up the generator all together (and even then I'm not sure if you can get up to the RTD's).

We have also seen the same problem on the bearing temperatures were wiring was routed through the generator housing in the same fashion, we replaced this section by twisted pair and connected shields, problem solved)

I have mentioned (and documented) this to both Brush and GE in vain but apparently GE has recently acknowledged the problem and is pushing Brush to do something about their wiring.

 

Hello,

I had the same issue on my LM6k units with woodward micronet controllers.

What I found was that the wiring between the terminal box on the ends of the generator, and the MGTB we not shielded from the factory.

As a quick fix, we ran temporary shielded wiring until we could take an outage and replace the original cabling the right way.

Problem solved.

I hope this helps,

William L. Warnke
[email protected]