I recently asked the manufacturer of the All-Test Pro 31 low-voltage tester for a spec sheet on the tool's accuracy, and was told that none exists.

The tool measures impedance, phase angle, and current/frequency ratio, among other things. A representative for the manufacturer told me that the lack of specs were not considered to be much of an issue, due to the low (25-800 Hz) test frequency of the tool.

Comments?

Don't know about the spec sheet, but if you do a search on "all test pro" or "Low voltage testing", you will find a bunch of threads on the subject.

D

Thanks, Dave. I have researched those threads. I'm simply curious to know what folks' thoughts are on the lack of a spec sheet of any kind. I did ask for one.

I have never used All Test Pro and don't have any opinion on the overall equipment other than specific claims that have been discussed here on the forum.

I think it would be rare to find a manufacturer of any kind of measuring/test equipment that would not provide some type of description of the accuracy of the measurement output. It would be like Best Buy putting up a computer for sale and they won't tell you how much RAM it has.

quote:

I think it would be rare to find a manufacturer of any kind of measuring/test equipment that would not provide some type of description of the accuracy of the measurement output.

I agree, and I can't see any reason why the tool's ability to perform low-frequency testing should render accuracy specifications unnecessary.

I was able to obtain a spec sheet from All-Test Pro for their companion tool, the All-Test Pro IV 2000. However, that spec sheet does not include general specifications that set the parameters of the accuracy claims: warm-up time, working temperature ranges, working altitude ranges, and so on.

Accuracy for the first spec I looked at - resistance - is listed at plus or minus 1.5% for resistances between 1 and 999 ohms. No LSD or deviation is noted. However, the accuracy drops to plus or minus 5% for resistances between 250 and 999 milliohms. No spec is given for resistances below 250 milliohms; a representative of the company recommended that I simply compare phase readings to one another, or use a milliohmmeter for such measurements. In fact, that's what I already do.

The problem I have with the first piece of advice is that I see no point in comparing readings produced by a meter with no assurance of accuracy at a given range.

Please note that I am not wondering how to take readings. I am simply pointing out my concerns about the specifications - or lack of - for these tools.

Those of you who use the All-Test Pro 31 and/or IV 2000: did you know about the specs prior to using them? If not, does this affect your regard of the tool in any way?

This message has been edited. Last edited by: Jack Rosebro,

I have been using both the AT-31 & the AT-IV for some years now. There are issues that I have with both; which you can find in previous threads on low-voltage testing.

I use them for two things only: check the high frequency impedance values to look for interturn shorts; & the rotor test to confirm rotor bar damage if detected by electrical signature analysis. The AT-31 is much better in this regard.

As most of my work is with generators & large motors, I would anyways do tan delta, PD, surge, etc. So, I do not rely on the All-Test for prediction of connection problems, contamination, ground insulation deterioration, etc.

I actually did not get into the specs while buying them, it was just a choice between PdMA & All-Test. Initially, I thought this might replace tan delta, PD, etc. but that is not so. Guess all have to be used together.

I wonder why is Jank silent so far. This is his pet peeve & I wonder what he has to say.

Regards,

Aditya

This message has been edited. Last edited by: Aditya,

Hi Aditya,

quote:

Originally posted by Aditya:
I use them for two things only: check the high frequency impedance values to look for interturn shorts; & the rotor test to confirm rotor bar damage if detected by electrical signature analysis. The AT-31 is much better in this regard.

By "check high frequency impedance values", are you referring to the current/frequency response test? That seems to be more commonly referred to as a low frequency test (25-800 Hz for the All-Test Pro 31). Is this what you are referring to?

Have you tested the ATP IV 2000 to determine the range of test frequencies that it can put out during the current/frequency response test?

Have you ever failed a motor with either tool, only to find that it had no issue?

Thanks for your input, I appreciate it.

Hi Jack,

Yes, I do refer to the current/frequency response test. However, I prefer the impedance values to the phase angle & I/F numbers, as it is easier to relate to impedance.

I have seen cases where impedance is balanced but I/F & phase angle are not; these machines are running well even today. On the other hand; whenever impedance unbalance showed up, definite problems always existed. The key was that impedance was almost balanced at 50 Hz, but showed significant variation at 400 & 800 Hz.

We always check with both AT-IV & AT-31, because AT-IV will not tell you the test frequency; & impedance without frequency does not make sense to me. The AT-IV tries to make a measurement at 800 Hz, if that does not happen (for whatever reason, I don't know), it will take it at the next highest range (400 Hz) & so on.

At times, the AT-IV will take readings for one phase at a different frequency than the other two, making it seem like a fault exists. No idea why that happens. This is the second reason why I always ask for the AT-31 readings. The AT-IV is good for databases, but for troubleshooting we always prefer the AT-31.

Regards,

Aditya

Hi Aditya,

Thank you for your reply. As you are testing for turn-to-turn shorts with impedance, rather than I/F and phase angle, do you think you would be able to do so with a LCR meter that had a sufficient selection of frequencies for your purposes?

I realize that you would still need the AT31 for rotor bar tests; however, I test PM motors, and have no need for those tests.

Thanks again,

Jack

This message has been edited. Last edited by: Jack Rosebro,

Hi Aditya,
good to see you again after long time ( with the same thread )
Mr Jack
just go through with all previous threads on ATP and you will sure get require info before buying low voltage tester.
well I have used both ATP as PDMA but pdma is much better ( specially when you are giving any official recommendation on prediction of motor failure )
There is lot to know about low voltage tester and Mr Aditya has even done some experiment for prove reliability of low voltage tester.

Njoy

Jignesh Khatri

Hello Aditya, et.al.

I was the GM at ATPro when the AT31 was designed. Not my favorite tool, the original ALL-TEST III was more stable. My preference was still, and I still use, the ATIV along with other tools, as appropriate.

The specification is +/- 0.1% with 1% repeatability. It is a comparitive tester. I only ever used it for rotor positioning if had an ATIV available.

Aditya, be careful. The actual frequencies are not actually the displayed frequencies, but will be within a few hertz. You can get a measurement of the test frequency from the ATIV by taking your impedance reading in manual mode.

So far as usage, we were using an AT31 in the GM (Government Motors) service lab to check the windings and rotor magnets on assembled hybrid transmissions and the ATIV for confirmation of any winding conditions. These were used to compare on the design side and to the results we would see in ESA/MCSA testing (Ref. Uptime Article on ESA testing of hybrids).

Howard

Hi Howard,

quote:

I was the GM at ATPro when the AT31 was designed. Not my favorite tool, the original ALL-TEST III was more stable.

In what way?

quote:

The specification [of the AT31] is +/- 0.1% with 1% repeatability.

Is that for one of the AT31's measurements, or all of them?

According to All-Test Pro, there are no accuracy specifications for the AT31.

What do you make of that?

Thanks for posting, I appreciate it.

This message has been edited. Last edited by: Jack Rosebro,

Hi Khatri, and thanks for posting,

quote:

Mr Jack
just go through with all previous threads on ATP and you will sure get require info before buying low voltage tester.

Thanks. I have done that. I will take a closer look at the individual tests, as well, but first I want to sort out my concerns with the manufacturer's specifications.

quote:

well I have used both ATP as PDMA but pdma is much better ( specially when you are giving any official recommendation on prediction of motor failure )

In what way?

Hi Jack,

It may well be, try it out & let me know. I do have some inductance & capacitance meters, but they are measuring at a fixed frequency only.

The other useful thing in the AT-31 is the IR value measurement upto 500 MOhms, at 500 V & 1000 V. For the larger motors & generators that I usually cover, 500 MOhms is insufficient; but would probably cover the smaller PM motors that you need to test.

Hi Jignesh,

Good to see you are still around. Why the change from "Thermographer"? Could you post some readings of PdMA vs. ATP for the same motor, so that your point is more clear?

Regards,

Aditya

Hi Aditya,

quote:

The other useful thing in the AT-31 is the IR value measurement upto 500 MOhms, at 500 V & 1000 V. For the larger motors & generators that I usually cover, 500 MOhms is insufficient; but would probably cover the smaller PM motors that you need to test.

You are correct; 500 megohms is indeed sufficient for the motors that I test. However, most of the motors come with IR test voltage specifications, some of which are lower than 500V. For such cases, I use a megohmmeter with a range of test voltages below 500V.

Thanks again for continuing to add to this thread. I'm looking forward to responses from the other participants, as well.

quote:

I was able to obtain a spec sheet from All-Test Pro for their companion tool, the All-Test Pro IV 2000. However, that spec sheet does not include general specifications that set the parameters of the accuracy claims: warm-up time, working temperature ranges, working altitude ranges, and so on.

Accuracy for the first spec I looked at - resistance - is listed at plus or minus 1.5% for resistances between 1 and 999 ohms. No LSD or deviation is noted. However, the accuracy drops to plus or minus 5% for resistances between 250 and 999 milliohms. No spec is given for resistances below 250 milliohms; a representative of the company recommended that I simply compare phase readings to one another, or use a milliohmmeter for such measurements. In fact, that's what I already do.

On one hand, the rep may be correct. As you know there is a difference betweeen accuracy and precision. If you are shooting arrows, accuracy is how close you get to the center of the bullseye, precision or repeatability is how closely the arrows are clustered to each other. If for example the 5% accuracy includes only a 1% allowance for repeatability errors, then that 1% would be the more relevant number to look at for a comparison test between phases of a three phase machine.

On the other hand, the specification does not address precision (most test intstruments do not address it). Since accuracy is always greater or equal to precision, it is "conservative" (with respect to evaluating wortst case variability between test introduced by nonrepeatability of the test equipment) to assume precision = accuracy when we have only accuracy specified. A repeatability of +/-5% would certainly not be good enough to detect variations in phase when often the limit for variation is variously specified as 1%, 2%, 3%, 5% (I have documents that mention all those numbers – depends for one thing on whether you are testing from switchgear or motor terminals, and also on how accurate is your test equipment as we have been discussing... also the wording of how we apply those numbers is varies...max difference between top and bottom or mas difference to the average).

Let's look at the low end of the specturm specified test equipment accuracy. IEEE 62.2-2004 (IEEE Guide for Diagnostic Field Testing of Electric Power Apparatus— Electrical Machinery) states

quote:

Vibration of a joint can apply high cyclic forces at a point of high mechanical stress concentration or to an area where the conductor has abrupt changes in hardness. These are conditions which may lead to conductor fracture. Insulating blocks can also vibrate and wear away part of the cross section of copper conductors. Only the precision of resistance measurement offered by a Kelvin bridge or microhmmeter will allow the detection of the subtle changes in conductor resistance caused by these types of damage. These instruments incorporate a measurement range which span the resistance values offered by most copper windings and conductors used in electric power apparatus. Kelvin bridges and microhmmeters measure to 5 significant digits and are accurate to within 0.25%.

I'm not sure I agree with the logic that conductor fracture (which I read as open circuit) requires good precision to detect. And I'm not sure that reduction in conductor cross section is a common detectable fault. But certainly shorted turn requires a precise resistance measurement to detect if we try to detect it with resistance comparison test. Note at our plant for our large critical motors we use a bridge device and I believe the accuracy is stated as 1 milliohm and I'm not sure it would meet this stringent 0.25% limit. Clearly the +/-5% you mentioned does not meet it either.

Now let's look at the high end of the spectrum of allowable variabiility in measurements 5%. Look for example at IEEE1415. It states "The three values are compared—all readings should be within 3% to 5% from the average of the three readings." With your stated "plus or minus 5%" you could have one at 5% above actual, one at 1% below actual, and one at 5% below actual and the one at 5% above actual would be more than 5% above the the average and therefore fail the highest (5%) limit mentioned in IEEEP1415, even though the actual winding could be perfectly balanaced. To be really technical/picky, some of the errors are statistical in nature and the tolerance should be accompanied by a confidence interval (i.e. something like 95% probability that the reading will be within 2% of ___). But I don't ever remember seeing that language in any test equipment specifications.

So even when we look at the high end of the resistance limit range within the references I know of (5% of average), the test equipment specification of +/-5% below 1 ohm that you mentioned might be a problem which could fail a good motor. (note again we made a "conservative" assumption about relationship between known accuracy and unknown precision).

On the third hand (whoops, I ran out of hands!), we could say that the ability to detect shorted turns by dc resistance test during PDM is not critical because:

• for ac windings, we usually expect a shorted turn will progress very quickly into a complete winding failure which is easily detected by open circuit or ground fault. This is due to the autotransformer effect. Such an effect is not present for dc windings which can operate for long periods of time with shorted turns.
  
• There are more effective ways to find shorted turns than a dc resistance test. Power frequency Ac inductance test can be better than dc for this purpose (assuming there are not significant variability in magnetic characteristics at the measurement frequency) because dc resistance varies with N while ac inductance varies with N^2 (where N is number of turns squared). Aditya has also shown the two-frequency ac test used by BJM is even better (and it makes sense from the theory).

This message has been edited. Last edited by: electricpete,

Here is some more info regarding various limits for variation between measurements.
Two terms I usen related to comparing three meausurements for a 3-phase machine:
"Range" = (max – min)/average
"Imbalance" = (Max Deviation from average)/Average

As a ROUGH rule to translate between these two terms, the range is twice the imbalance.

I have a tabulation of winding resistance variation limits from various EPRI specifications and documents. The 6 digit number list is the EPRI report number:

• Within 5% of each other (5% range, ~2.5% imbalance) per EPRI 108773v1.
  
• Within 5% original value and 1% range (~0.5% imbalance) per EPRI 107524
  
• EPRI Rewind spec 1000897 recommends a maximum 1% winding resistance "imbalance" among the three readings (which they define as imbalance = (Max-Min)/average). Corresponds to approx 2% range.
  
• EPRI 111195 (MOV’s) – 5% from original factory data and 2% range (~1% imbalance).
  
• EPRI 111196 (form wound) – 1/2% range (~1/4% imbalance).
  
• IEEE 1290 (MOV) PAGE E41 – 5% range (~2.5% imbalance)

Note you can see the highest of any of these is 5% range and 2.5% unbalance. The +/-5% test equipment variation cited above would roughly correspond to 10% range or 5% imbalance and could fail a good motor by any of these EPRI specs. (again this is based on "conservative" assumption that the +/-5% is error due to non-reproduceability rather than bias type errors).

Also I should have mentioned earlier that the IEEE62-2 document quoted above scope applies to equipment rated 4kvac and above.

This message has been edited. Last edited by: electricpete,

Thank you for taking the time to post. You raise some good points.

quote:

As you know there is a difference betweeen accuracy and precision. If you are shooting arrows, accuracy is how close you get to the center of the bullseye, precision or repeatability is how closely the arrows are clustered to each other.

Quite true. However, the way in which the specs are written continues to raise red flags for me:

No general specifications have been given. Such specifications can be important modifiers - usually, in a limiting capability - to an existing spec. For example, a megger I own is rated CAT III 1000V/CAT IV 600V. But that rating falls if the meter is operated at altitudes above 3000 meters. While that is of no concern to me, it may be important to someone in Leadville, Colorado. Even cheap, sub-$100 meters almost always come with general specifications. Why not these?

Specifications for accuracy are given with no mention of counts or least significant digits (for example, plus or minus 1.5% + 2). Again, almost all meter specs list counts or LSDs. Why the specifications for a tool that costs ten thousand dollars would not do the same, I cannot say.

quote:

Look for example at IEEE 1415. It states "The three values are compared—all readings should be within 3% to 5% from the average of the three readings." With your stated "plus or minus 5%" you could have one at 5% above actual, one at 1% below actual, and one at 5% below actual and the one at 5% above actual would be more than 5% above the the average and therefore fail the highest (5%) limit mentioned in IEEE 1415, even though the actual winding could be perfectly balanced.

I agree. And a motor manufacturer may, for example, issue a spec of no more than 2% variation between phases. So the automated tests would probably be, to me, of limited value; more useful, perhaps, for data collection than for actual diagnosis.

quote:

So even when we look at the high end of the resistance limit range within the references I know of (5% of average), the test equipment specification of +/-5% below 1 ohm that you mentioned might be a problem which could fail a good motor.

Exactly, and this is a concern of mine.

quote:

On the third hand (whoops, I ran out of hands!), we could say that the ability to detect shorted turns by dc resistance test during PDM is not critical because:

for ac windings, we usually expect a shorted turn will progress very quickly into a complete winding failure which is easily detected by open circuit or ground fault.

Open circuit? Not always, in my experience. Some VFDs will limit current to a winding upon detecting a fault. This may lengthen the amount of time it takes for the winding to deteriorate. But I am not making an argument for using DC resistance to identify turn-to-turn shorts.

quote:

There are more effective ways to find shorted turns than a dc resistance test.

I agree, and I want to make sure that my intent is clear. I brought up the tool's DC resistance spec as an example of my concerns with the manner in which the specifications are presented. But I had not discussed DC resistance in the context of the diagnosis of turn-to-turn shorts. To be honest, I really hadn't thought about it.

In all fairness, the company sensibly recommended that I use a DC milliohmmeter to measure sub-ohm winding resistances. Even a cursory glance at the All-Test Pro tools will show that they do not have four-terminal (Kelvin) leads, and thus cannot be expected to yield precise results.

However, I intend to go through the tool's capabilities one by one, and evaluate the relative usefulness of each feature. I need to find something that makes this tool stand out from others in a positive way, and makes it a good value. For those who are measuring windings of more than one ohm resistance, and who - in the absence of OEM specs - are solely concerned with balanced resistances - the ability of the All-Test Pro IV 2000 to measure resistance may be adequate. For me, it is not, and I have to look elsewhere to make the tool useful.

Setting DC resistance concerns aside for the moment, I am also interested in the tool's ability to detect turn-to-turn shorts. Its current/frequency response test is, as we know, one of the features that separates it, and the All-Test Pro 31, apart from many of their competitors.

But we have seen from Aditya's experiences that this feature does not always work as designed. Granted, he has figured out a way to make the tool work for him, but it appears to have taken a lot of trial and error. And that's not something that a user should have to do.

Thanks again for contributing to this thread. I appreciate it.

This message has been edited. Last edited by: Jack Rosebro,

Thanks Jack. I understand where you're coming from better now. Not just a concern with resistance, but the whole thing.

I can't add much except to agree with you again that if it is actually true that the manufacturer will not publish any accuracy specs, it's a very unusual thing and limits the user's ability to understand/evaluate the instrument. Although I suspect many users are not as meticulous as you are in looking at those specs.

As an aside – The VFD current limit feature will not do anything to stop the progression of turn/turn fault toward failure (to ground, to other phase, or open circuit) in an ac motor. The fault current created by a the turn short (due to "autotransformer action") flows in a small loop that includes only the shorted turns and the short itself – that fault current does not flow outside the motor. That's the precise reason why traditional protective relaying cannot sense turn to turn faults until they turn into phase-phase or phase-ground or open-circuit faults.

By the way - welcome to the forum!

Hello again, and thanks for responding,

quote:

Thanks Jack. I understand where you're coming from better now. Not just a concern with resistance, but the whole thing.

You're right . The specs - and DC resistance in particular - are just starting points to encourage observations and conversations as I work through various aspects of the tools.

quote:

I can't add much except to agree with you again that if it is actually true that the manufacturer will not publish any accuracy specs, it's a very unusual thing and limits the user's ability to understand/evaluate the instrument.

I started by searching the company website for product spec sheets. Not having any luck with that, I ran some web searches. It turns out that specs are available for one product, the All-Test Pro OL II (just search "atpol" and "specifications"), and they are fairly conventional: "plus or minus X percent, plus X digits". So we know that All-Test Pro is aware of such conventions.

After failing to find a spec sheet for the company's flagship product, the All-Test Pro IV, I simply emailed the company and asked for one. A very helpful gentleman emailed it to me. He also let me know that no spec sheet existed for the All-Test Pro 31 (and I did ask "are you sure?").

To be fair, Howard Penrose has stated in this thread that the ATP 31 has a spec of plus or minus 0.1%. That piece of information, however, raises new questions. Does that spec apply to all measurements, or just some? Why isn't the company aware of the specs? Most importantly, and given that the ATP IV 2000 specs typically range from plus or minus 1.5 to 5 percent, why would the ATP 31 - a tool made by the same company, measuring many of the same qualities, and costing one fifth as much as the ATP IV 2000 - be more accurate than the more expensive tool by more than an order of magnitude?

quote:

Although I suspect many users are not as meticulous as you are in looking at those specs.

I have to admit that I didn't really consider them at first. But when I started to run into things that didn't make sense to me, I backed up and started from scratch.

Speaking of backing up, something else caught my eye recently. I was reading the case history "Dynamometer at Automotive Plant", which can be found under "Case History" on the All-Test Pro website. Table 4, which can be found on Page 2, provides an example of "a good new 200 HP dynamometer motor." Phase resistance readings, as taken by the All-Test Pro IV are 12, 11, and 10 millohms, respectively. Again, these are represented as good readings on a new motor.

While I realize that low-resistance readings are not this tool's strong suit, the company has said that for such readings, balanced measurements are more important than absolute values, and that the ATP IV 2000 is capable of such measurements. However, the above readings represent a 20% variation between the highest and lowest phase resistance readings. To me, that's a lot.

Assuming that the new motor is indeed balanced, it would seem to me from those readings that the comparative ability of the tool, at least in this case, hasn't been all that great, either. Now if All-Test Pro were to advise "this tool really shouldn't be used to measure or compare resistances below X millohms", then they would be fine. But they don't.

quote:

As an aside – The VFD current limit feature will not do anything to stop the progression of turn/turn fault toward failure (to ground, to other phase, or open circuit) in an ac motor. The fault current created by a the turn short (due to "autotransformer action") flows in a small loop that includes only the shorted turns and the short itself – that fault current does not flow outside the motor. That's the precise reason why traditional protective relaying cannot sense turn to turn faults until they turn into phase-phase or phase-ground or open-circuit faults.

That is interesting; thanks for sharing that. I probably should have been more specific: many of the drives that I work with can operate at reduced power and artificially limited speeds when a fault is detected; the application is flexible enough to put up with such operation. I believe that some of the drives can limit voltage as well as current and speed in response to a performance problem. But I would have to investigate further to be sure.

Question: if a turn-to-turn short manifested itself through a performance anomaly that could be detected by the motor's self-diagnostic capabilites, and as a result, all three of these factors (current, voltage, motor speed) could be limited, could that mitigate in any way the auto-transformer effect of a turn-to-turn short?

quote:

By the way - welcome to the forum!

Thank you. I have seen some interesting conversations here, and I hope to add to them.

This message has been edited. Last edited by: Jack Rosebro,