Posted this on the technical/applications thread also:
Within the last 60 days our cooling tower Lakewood TDS meter's
operation became erratic . We have linked it to the operation of the
Danfoss V.F.D.s. When the drives are bypassed, meter operation is
normal and TDS is right on the money when checked against a handheld.
As the drives ramp down TDS rises, and vice versa. Both drives/towers
exhibit the same effects on the meter. Power was checked with a true
RMS meter and no voltage spikes or harmonics are evident. We also ran
the meter on UPS clean power with no change in results. The meter and
sensor are wall mounted outside in the tower yard approx. 40 ft. from
the drives which are in the chiller mechanical room. The copper
sampling pipe is grounded immediately upstream from the meter to the
tower earth ground arrestor system. At the meter location the TDS can
drift by as much as 1000 in a three hour period. Using a second
Lakewood and sensor placed in the tower basin we have found that the
the effect is still present but to a lesser degree. The closer you
move the second meter toward the first the greater the deviation. No
new equipment has been added and the system has been running trouble
free for over six years. Lakewood pronounced the meter to be in
perfect working order. Has anyone had any similar experiences? Any
input is appreciated.Thought of the RF issue but consider it unlikely
to be radiating from the cabinets. Why would both units develop the
same problem at the same time? Both units individually create the
same issue when the other is off line. What about a high impedance
ground connection? Wouldn't that have the effect of allowing noise
from the drive to radiate from the power leads instead of draining it
away? If so would the connection be local or might it extend back
toward the service entrance through the various MCC's? I'll contact
our electrical contractor for input, but what tests or procedures
should be implimented to track down this problem?

I think you are on the right path.

Assumptions:
1) TDS = Total Dissolved Solids Probably milliVolt level signal, possibly unbalanced.

2) Installed TDS system is OK. Similar results to handheld unit. Also similar unit in sump reads fine.

3) Problem is related to VFD’s. Problem goes away when VFD’s are bypassed. Problems tracks either VFD operation.

4) Amount of interference is location dependent. As sump TDS unit is moved closer to installed unit, the amount of interference increases.

5) The sampling pipe is grounded right by the sensor.

My guess is that you are getting noise being coupled into the conduits around the motor leads, and it is seeking a path to ground thru the sampling pipe. I would bet that you will find a high impedance path to ground from the motor conduits. Check for any current flowing in the sampling pipe just before the ground connection.

Was any maintenance done recently? Any painting, inspections, replacements of tiles, overflowing of the system, etc? Have there been any recent lightning strikes in the area?

What is “the tower earth ground _arrestor_ system” Is that a grounding grid?

LarryC

Larry, The arrestor system is am aluminum braided loop around the top of the enclosure with lightning rods which connect to a copper braid running down each side of the brick enclosure through the slab to earth ground. It is tied to the conduit at two points prior to entering earth. I don't have an RFI meter but I can pick up the "hash" around the conduit and in the genral area of the meter with an AM radio. It is the same hash that can be detected next to the drive enclosures themselves. There has been no significant maintenance or changes to the system othe than the usual monthly cleaning and we haven't seen rain here in over a month. I have inspected the conduit which runs from the motors through liquid tight to IMC with threaded connectors to the diconnects through IMC back to the drives and can see no visual signs of failure. Since most drains are open at one end and grounded at chassis at the source would temporarily disconnecting the liquid tight from the IMC prove or disprove a loop anntenna theory?

“I can pick up the "hash" around the conduit and in the general area of the meter with an AM radio. It is the same hash that can be detected next to the drive enclosures themselves.”

How is the grounding at the Drives?

____ASSUMING THE ARRESTOR SYSTEM AND THE PROTECTIVE EARTH ARE SUPPOSED TO BONDED TOGETHER____

How is the bonding between the arrestor system and the service entrance? What is the voltage between the neutral feeding the VFD’s and the arrestor system?

Using the radio, check for hash along the motor conduits especially in the area of the ground system bonding. If the hash does not change where the bonding occurs, I would suspect that the bonding connection is not secure.

Try a temporary jumper from the conduit bare metal to the copper braid. See if the hash changes. When you are dealing with RF noise, the long run of pipe looks like a high impedance path.


“Power was checked with a true RMS meter and no voltage spikes or harmonics are evident.”

Where? The Service entrance, the TDS equipment, the input of the VFD’s, or the output of the VFD’s?

LarryC

This is just a possibility that I would like you to consider. Many analog to digital converters have jumper selected filtering for 50 or 60 Hz. You would naturally select the filtering "notch" based on your line frequency. A VFD, ramping down would be steering the frequency of your potential line noise source farther below the center notch frequency. Of course, if you really had a phase controller, instead of a VFD, noise would be through the roof. You might ask them if they have different model #s, jumper settings, programming, Et,c. for 50 and 60 Hz. If so, there might be quite a challenge, like an active tracking filter, required.
Good luck,
Joe

I don't know the answer here, but I do have some knowledge of how TDS meters work, and this might be useful.
To measure TDS, the meter actually measures the conductivity of the water, as this is directly proportional to the TDS. Alas, it is not possible to simply apply a DC voltage to a couple of probes because of polarisation effects, so the measurement is made using an AC voltage and some kind of synchronous rectifier.
Unsurprisingly, this measurement is easily upset by stray voltages and currents in the water. Drive harmonics seem a likely villain here.
Now, as to why our poster is having problems, it seems to me that we need to identify just what it is that has changed since the TDS readings started to play up. And this is not clear from the original post.
I hope the above is of some help.
Mark Monson

Thanks to all for the input. The problem appears to be RFI radiating from the conduit. After locating the radio noise along the conduit run, I used a pair of jumper cables and grounded the conduit to the lighting arrestor earth ground in the tower yard as far from the affected area as I could manage, which was about 15 feet off the ground and 10 feet from the tower and first meter. Not great surface area contact but enough to have an effect. Both meters now read approx the same and though there is still some drift it has been reduced substantially. The original meter which is the furthest down stream from the grounding clamps is now more accurate than the meter in the tower. We are having our grounds checked by our contractor on Fri. There have been no changes to the system, so it will be interesting to see what has happened. Since both drive paths are affected, and since there are two additional drives fed by the same MCC which are not affected, I’m anticipating that the problem will be located between the drives and the MCC.

I'm glad to hear that is getting under control.

LarryC

jkochan:

Its been bugging me why the problem all of a sudden showed up. Something triggered this!

“Within the last 60 days our cooling tower Lakewood TDS meter's operation became erratic.” “There has been no significant maintenance or changes to the system other than the usual monthly cleaning”

“No new equipment has been added and the system has been running trouble free for over six years.”

SOMETHING HAPPENED WITH IN THE LAST 60 DAYS.

I would suggest examining EVERYTHING that was done in the last 60 days. My guess is that you will find that the grounding of the conduits was disturbed. Check for structural modifications, cutting and welding, addition of _anything_ to the brackets that the conduit is mounted to, somebody cleaning up the area, new paint, apparently unrelated actions like paving, new siding, additional lighting, new signs, etc.

It could be that the grounding path was unintentional but it worked. Now that somebody ‘fixed it’ to the way it should be, and your problem showed up.

The other thing to check is the RF grounding of the motors themselves. Are the bodies of the motors isolated from the grounding system? Are the motors on vibration mounts? How about the output shaft couplings? Has there been any mechanical alignments, checks, or repairs with in the last 60 days?

For the system to work fine for six years and all of a sudden develop a problem in the last 60 days is a bit unusual. Poke around and try to figure out what changed.

LarryC

All grounds tested .2 ohms or less between the fan, all the way back to the service entrance.In addition, our BAS went berserk every time we opened or disturbed a ground anywhere. It apparantly relies heavily on grounding for communication and is very touchy. Even loosening a ground screw connection and teaking the wire was detected and we received a communication alarm. This would seem to indicate that the grounding system is fine or we would have had other problems. more infor will follow... thanks for all the input...it's helping.

No luck, no change, calling in the big guns. We called the electric testing Co. that tested the building originally and they are coming out to check the grounds between MCC's and the Mains and the building ground. They should have the original data and we can then compare if there has been any degradation.
I kept coming back to ground issues and an idea that has been percolating in my mind. A grounding issue if that's what it is, may not be as localized as we have been thinking. I've begun to suspect the building service entrance grounding resistance has been slowly increasing to the point where it is just now becoming apparent in it's effects on our equipment. The areas outside and parallel to the main electrical room were, for the longest time so heavily irrigated and plagued with leaks that the ground was always saturated to the point that we had developed a sink hole in the center area right outside the main electrical room and we had lime leaching out of the concrete walkways along the edges. Over the last year changes to, and failures of the irrigation system has allowed these areas to dry out. This has also been the driest year weather wise for quite a while.
Since I started typing this a field engineer stopped in to look over the situation. He said that this area of Auwatukee/Phoenix has some of the highest ground resistance readings in the city and a super market about 5 miles away had to add earth grounding conductors for the same reasons. This was before I mentioned the irrigation theory. So it's plausible that this may put us on the right track.

“Within the last 60 days our cooling tower Lakewood TDS meter's operation became erratic.” “There has been no significant maintenance or changes to the system other than the usual monthly cleaning”

“No new equipment has been added and the system has been running trouble free for over six years.”

SOMETHING HAPPENED WITH IN THE LAST 60 DAYS!


The drying out of the soil and the high ground resistance are issues, but SOMETHING HAPPENED WITH IN THE LAST 60 DAYS!

"I used a pair of jumper cables and grounded the conduit to the lighting arrestor earth ground in the tower yard as far from the affected area as I could manage, which was about 15 feet off the ground and 10 feet from the tower and first meter. Not great surface area contact but enough to have an effect. Both meters now read approx the same and though there is still some drift it has been reduced substantially."

The results of the jumper cable test shows that the TDS meter EMI issue is an internal problem, independent of the changing ground resistance. SOMETHING HAPPENED WITH IN THE LAST 60 DAYS!


"All grounds tested .2 ohms or less between the fan, all the way back to the service entrance."
Was this resistance check done with a multimeter? When dealing with RF, low voltage DC measurements don't always mean anything. When you are dealing with RF noise, the long run of pipe looks like a high impedance path. Low DC resistance doesn't mean that the RF impeadance is equally low.


_______________SEPARATE ISSUE________________

"In addition, our BAS went berserk every time we opened or disturbed a ground anywhere. It apparantly relies heavily on grounding for communication and is very touchy. Even loosening a ground screw connection and teaking the wire was detected and we received a communication alarm. This would seem to indicate that the grounding system is fine or we would have had other problems."

I disagree. I think that means that the BAS signals are running thru your grounding system. Should the BAS signals be referenced to ground? That doesn't sound like a reliable design for a multi-structure system. That means whenever there is a voltage difference between the 'ground' of structure A and the 'ground' of structure B, the BAS signals change too.

LarryC

Here's an update: All ground resistances as measured with a megger 4 Wire bridge were .2 ohms or less including the main grounding conductor to earth.We haven't recieved the report yet but as I recall THD at the drive inputs was 1.26% or lower wave for was clean. The ouput of the drives had mutiple harmonics in varying percentages. Figures will follow when we get the final report.