Flir IM75 Insulation Tester / DMM

Introduction

This blog continues my series of comparing some insulation testers to the Keysight  U1461AU1461A that I have had issues with and the Megger  MIT420/2MIT420/2 that I carried a road test review out on for element 14 last year

Keysight U1461A meter fault blog

Megger Insulation Tester - Review

Chauvin Arnoux CA6526 Review

The next insulation tester on my list is from Flir. The IM75 is their only insulation tester, that they also classify as a multi-meter. I mostly know Flir from thermal imaging cameras, but they have had a multimeter range out now for quite some time, but I can't say that I have ever seen anyone using any of these instruments. This does surprise me, knowing the quality of Flir instruments in general, I would have thought they would be of interest to industrial electricians.

Package Contents

The unit also carries the high price tag often associated with Flir instrumentation. The instrument comes packaged in a presentation style case and without a protective case included. This is extra, and I found that the price for it varied between £50 and £90 here in the UK. That is a substantial increase for an instrument that already costs £667 RRP and is not comparable to the competition, specifically the Megger and Keysight units, both which come with a hard case.

The case for the Flir is very good quality and fits the meter well. However, the pocket to the side of the meter is a little small to contain both leads and crocodile clips. Luckily enough the lid of the case contains a separate compartment where the crocodile clips can be stored. The case I have is the TA10-F and the extra compartment in the lid is moulded for the TA72 or TA74 current probe that Flir manufacture to go with their multimeters. The also make the TA10, which, in hindsight would have been a better option if you don't have the current probe, as the lid in this just has a centre divider.

Along with the instrument, the kit contains a set of leads, two crocodile clips, a magnetic hanging strap, instruction book in paper and electronic format. Unlike the other two units, the instrument does not have remote probe operation capability. It also doesn't come with a factory calibration certificate, again given the cost and the competition, I would have expected this to be included. I have also been unable to find a calibration procedure for the unit, so if it ever needed adjustment, I presume it would have to go back to the manufacturer or an approved dealer.

The crocodile clips screw onto the probe tips, so are a unique fitting for the Flir. They have reasonably large jaw capacity, slightly bigger than the Megger clips, but smaller than both the Keysight and Chauvin Arnoux clips.

The probes themselves are formed directly onto the leads so there are no options to use alternative 4mm shrouded connection accessories without changing the whole lead The leads are silicone insulated and good quality they read 22GOhm on the IM75 1000V insulation test range I also took the liberty of testing them with the  MIT420/2MIT420/2 and the insulation goes up to 200GOhms at 1000V after around 5 seconds

Comparing the Flir instrument to the Megger and Chauvin Arnoux Insulation Testers, shows a similar range of specifications. It does offer a higher capacitance range than both of those units along with a diode test function. For electricians dealing with power electronics, both of these would improve the flexibility of the tester. However, in terms of an insulation tester, although it has a full range of test voltages, the maximum resistance capability is seriously low in comparison to the other units, especially at the 500V level. Although the Flir offers DAR and PI tests, which would be restricted by the lower resistance capability, it does not offer a ramp test or a variable test voltage function.

In comparison to the Keysight unit, that is a fully specced insulation multi-meter, the Flir falls quite short in terms of functions and ranges. For me the Flir isn't really an insulation multi-meter and is much more in line with the insulation testers on offer.

Function Tests

The Flir instrument is a good looking quality unit, finished completely in black, it has a slightly rubberised feel to it and is comparable in size and weight to the other two units.

I do not find the screen particularly clear to read, especially the small function icons towards the bottom of the screen. The display however, is enhanced by the backlight.

The instrument is reasonably intuitive to use, with a function switch to select the voltage, continuity, ohms and insulation test voltage ranges. A mode button is then used to scroll through the functions available for that specific switch position. Voltage, capacitance, diode and continuity functions are all automatic. Zero ohms and the insulation test functions are all manually activated using the centre test button.

The meter has 3 input jacks that are colour coded to the function switch positions. The voltage and continuity positions are on the blue coloured jack and the zero ohms and insulation ranges are on the white coloured jack. This jack is protected with a 400mA HRC fuse but this only provides protection for the zero ohms function. I did also note that there was no fuse blown alarm on the instrument, the meter merely maxed out at 40 kohms on the zero ohms function with the fuse removed, and the insulation test range continued to work as normal. A bit disappointing given the cost of the unit and less expensive offerings from competitors that will indicate a blown fuse.

The majority of testing was based around the insulation test capabilities, as this is the main aspect of my role. As expected the Flir performed reasonably well with the basic function tests of the insulation ranges. Short circuit and 1mA load current tests all produced results within specifications for the IEC standard.

The 1mA tests were comparable to the three other units I have tested, the short circuit current is towards the upper end of the expected values. There also seemed to be more variation in the short circuit current across the different test voltage ranges in comparison to the other units as shown in the table above.

The open circuit tests showed a poor level of voltage accuracy in comparison to the Keysight and the Megger units. The results were more in line with those from the Chain Arnoux unit.

The voltage output regulation at 500V was worse than the other units, showing a higher value of voltage as the insulation resistance value increased as depicted in the plot. However, this is a response typical of the vast majority of insulation testers, with the Keysight and the Megger units showing the more unusual response of a highly regulated output.

Tested against the motor winding simulator configured in a delta winding mode showed comparable plot data to the other units tested, with the final DAR ratio value showing a +0.9% deviation from the average of the DAR from the Keysight and Megger units, and the PI value showing a deviation of 1.8% from the average value. The full PI plot data for all of the units can be seen below. It was noted that following a DAR or PI test, only the ratio value was displayed and the actual resistance value could not be retrieved. This is quite a disappointment, when carrying out timed insulation tests both the resistance values and the ration values need to be recorded. Both the competitor units allow these values to be displayed once a timed test is completed, so again the Flir unit isn't matching the competition.

Its was further noted that once a DAR or PI test function is selected, all other menu settings are lost. Therefore the test result cannot be saved to memory. Again offerings from competitors allow all of the significant data surrounding a DAR or PI test to be saved and recalled to the screen as required.

Tested on the winding simulator, the 500V insulation test function was adequate. However, out in the field on a real motor the 550MOhm limit of the 500V range soon proved to be inadequate. Insulation systems on modern motors will very quickly rise above 550MOhms taking the meter over-range. The result of this during a DAR or PI test is to produce an 'Err' message on the screen as seen on the second photo below.

On the 1000V test range, the instrument fairs a little better with a 20GOhm reading capability. The insulation tested recorded a value of 3.3GOhms and the DAR test recorded a value of 1.0. Both readings are typical for this specific motor.

Insulation testing of rotating machines within the UK is guided predominantly by IEC 60034-27-4, last updated in 2018. Further information can also be obtained by the IEEE 43 standard and Nema publications. The IEC standard allow for a new 400V piece of apparatus to be insulation tested at 1000V, but after being put into service, the recommendation is to drop the test voltage down to 500V. This means I can test a new motor with the Flir instrument, but subsequent tests at the lower voltage may not produce values, unless the insulation has dropped below 550MOhms. Insulation testing is just as much about trending values as it is about spot readings, so not being able to see the trend above 550MOhms limits the value of the testing.

The same standards provide guidance on DAR and PI test methods. From both the IEC and IEEE, guidance is given that a piece of apparatus with a 1 minute reading above 5GOhm may not give a PI value indicative of the condition of the insulation and therefore must be used with caution. The maximum PI I have come across in my experience has been around 8, although this is rare. To get a PI value of 8, with the 1 minute at 5GOhm needs a 10 minute value of 40GOhm. This puts the test well outside the capability of the 500V range and is probably pushing towards the limitations of the 1000V range. This is quite a dissapointment for me with my specific job requirements but also given the brand name, quality and price of the product in question.

The resistance of the winding simulator was also measured with the zero ohms function.

In general the Flir unit produced values higher than the nominal values and typically higher than any of the other three units. The Flir measured around 3.5% higher than the expected values against the manufacturer's specification of 1.5%. Whilst measuring such low resistances, the quality of the connections will drastically affect the results, but the simulator is built around a standard motor terminal block, so this represents a real world test method. Despite the questionable accuracy, the Flir unit did produce consistent results across the three phases for each test.

The test current was measured as being 278mA. This is significantly higher than any of the other units tested, that have a test current in the region of 205mA. The IET recommendation is for a minimum test current of 200mA for an earth bond test. Generally when performing this test, a higher test current is more likely to show up high resistance joints.

This is a significant current drain for a battery powered instrument and introduces one of the major misgivings I have with the unit. Flir have chosen to utilise AAA sized batteries to power the unit. All other instruments of this type I have come across, have opted for AA sized batteries.

The meter has a 3 bar battery indicator that drops down to 7.50V at 0 bars. The unit will however, function down to 7.00V before it cuts out with a low battery warning. This only uses 22% of available battery voltage where as the Megger unit used up to 36% of the battery voltage capacity before cutting out.

The table above shows the typical current drain on the battery for the various functions. Voltage, diode, continuity and capacitance test functions all show quite light loading. Use of the backlight shows a 0.021A increase on the voltage range, again not to much of a load. The load goes up somewhat when using the built in torch, this was again tested on the voltage range and shows a 0.149A current increase.

The big load is the zero ohms function, showing a substantial 0.435A battery drain, this would go up to 0.456A if you used the back light and then take it up to 0.605A if you decided to turn the torch on at the same time. A data sheet from Varta gives the typical discharge rate of 1.220Ah for a AAA battery, discharge type data gives 0.86Ah for a discharge down to 1.0V with a 5.1Ohm load. With all 6 cells at 1.0V, the Flir IM75 would already have cut out, requiring a minimum voltage of 1.17V per battery.

It seems fair to say that battery life may be a big concern for this unit, if the zero ohms function is used extensively. Indeed, after carrying out my own set of function tests, which will be the same for all units, the battery indicator on the Flir unit now drops down to 1 bar when the zero ohms function is used. Both the Keysight and the Megger units, showed full battery capacity after the tests, the Chauvin Arnoux had dropped to 95% battery capacity remaining. It should also be noted that the Flir actually conducted fewer tests than the other units due to its lower maximum resistance value limits. The Keysight, Megger and Chauvin Arnoux units carried out 220 tests each, the Flir conducted 120 tests.

Accuracy of the unit on the insulation resistance ranges were all within specification as seen in the data collected below.

The plot shows the deviation from the expected value being measured and it can be seen that the ranges generally follow one another. The 100V range does divert away from the other test voltages between 60 and 90 MOhms. The average accuracy of the instrument on insulation testing over its full range, came out at -0.21%. Compared to the other instruments, this is quite high but interesting, when the results range of the other instruments is restricted to that of the IM75, the average accuracy becomes almost the same. Therefore when testing out in the field, the results from the IM75 will be just as good as either of the other instruments.

The unit also features a built in torch at the top. As the case has the probe grips moulded into the back, this will adequately light up a work area whilst probing around. With the probes removed from the case grips and the meter stood up with its tilt stand, the torch is less useful.

Memory Function and Software

The IM75 has a built in memory function and software connection via bluetooth to either an IOS device running the Flir Tools Mobile app or another Flir device with the MeterLink option installed. The onboard memory holds 99 measurements. However, I have found that these are only, voltage, diode, capacitance, resistance or insulation resistance values. A DAR or a PI test cannot be saved to local memory. The reading stored can be recalled to the screen, however it cannot be downloaded to a computer. There is only the option to delete all saved readings. In my opinion, this gives the onboard memory function limited functionality and use.

Connecting the device to an iPad did not offer any significant improvement. The IM75 did not need to be paired through the iPad setting panel, but was found automatically by the Flir Tools app. The meter display is very basic, although it does appear to work for all the test functions on the meter including the DAR and PI functions. However, during these tests, only the resistance value was displayed and not the voltage value, also only the live value is displayed and not the trend plot. Once the test is complete, there is only the option to take a screenshot of the value.

The three screenshots above show the connection screen, a live insulation test and capacitance test from the iPad.

Flir also offer Windows based PC software package, Flir Tools. However, this did not offer direct connection to the insulation tester. To utilise this software, the insulation tester has to be paired with a Flir infrared camera and then the data from the IM75 can be transferred and saved to an infrared screenshot. Flir Tools itself is free to download, they also offer Flir Tools+ with extended capabilities that is licensed software and comes at a further cost of around £300+vat for the UK.

Teardown

The Flir IM75 looks like a high quality unit, and in that respect the dismantling process did not disappoint. The battery compartment lid was held in place by four captive machine screws, two of the screw are underneath the tilt stand when it is closed and Flir have marked screw heads on the corners of the tilt stand to indicate that the screws are underneath, very nice attention to detail. Removing the battery compartment reveals the six AAA batteries and the 400mA fuse.

The back case is held in place with six machine screws. These days it is rare to find a multi-meter case held together with machine screws, even on other high end brand multi-meters, a sign of the quality that Flir represent. Inside, the back case has its upper half shielded and there are two fly leads with connectors on them to the main board, one for the torch LEDs and the other to the battery pack. This allows the rear case to be separated completely from the front section.

To continue with the board removal, there are three machine screws securing a PCB board to the input jacks and a fourth securing the function switch contacts to the rotary switch itself. It was whilst I was looking to remove this, that I noticed a foreign body trapped underneath the switch contacts. Closer inspection revealed that this is likely a bristle from a brush or perhaps hair, either way I am sure it shouldn't be in there and it is quite a disappointment to find it on such a high end instrument from a reputable manufacturer. I doubt that it would ever have created a longterm issue within the instrument.

With the switch contact removed, seven more screws hold the circuit board in place within the front case, there are four more screws that hold the display shroud to the board, but these can stay in place.

With the circuit board out, it can be found that the majority of the components are on the rear side of the board. On the front side a small board is connect via a header block to the main board that houses the function buttons below the display. These function buttons seemed to be a sealed metal spring plate, something I have not seen before. I do not know if this type of switch would be more reliable than the usual carbon pad onto copper track switch found in other meters. It certainly adds a nice feel to the buttons when using them.

The rest of the front board contains some PTCs and a high wattage resistor for the input circuit. The isolation slots cut into the PCB can also be clearly seen. Further elements of the input circuit can be seen on the rear of the board, some varistors and more PTCs along with the input fuse and resistor based voltage dropper circuits.

The larger chip on the board is a Holtek HT1622 Display driver and memory mapping device. The main MCU appears to be a Fujitsu MB95F718E.

An unused header connector can also be seen just to the left of the main chipset and I am wondering if this is used for calibration, which would make the process and open case methodology, which would be quite unusual these days.

The input jacks are moulded into the front case and secured to the input PCB via machine screws. It looks like the force on the jacks is taken up by the case when inserting the leads, so there should be no reliability issues with the connections.

Overall, the construction is of a high quality and does go towards justifying the high price tag of the unit.

Conclusions

The Flir IM75 is certainly a high quality unit and comes with a price tag to suit. I would like to think that the foreign object found within the unit, is a one off and should distract form the overall quality of the unit. There are however a number of misgivings regarding the operation of the unit, that lead me to being quite disappointed with the overall performance of the unit.

• The use of AAA batteries seems to be a poor choice for an instrument that can have quite high power demands.
• The implementation of the DAR and PI functions is poor compared to competitor offerings.
• The maximum insulation value of 550MOhm for the 500V range is much too low for modern insulation systems.
• On board memory functions are very limited and also not up to competitors offerings.
• Bluetooth software connectivity has limited functionality unless a Flir Infrared camera is also available.

The findings mean that the Flir IM75 just doesn't have the test capabilities as an insulation tester for the role that I do and I find that such a shame as I do like the build quality and style of the unit. In fact, the implementation of the insulation test DAR and PI functions is so poor, it just isn't worth me taking the unit up to the generator rotor to see how it performs, I already know it won't be up to it. This does not mean that the unit is totally useless. I do like the Meterlink functionality to link multiple instruments together and I do have a couple of applications where the recording of voltage and current to a thermography picture would enhance maintenance activities.

I need to verify that I have the correct type if infrared camera available to me to investigate these applications and if I do, it won't be the last you will see of the IM75. But for the time being it will have to be assigned to a storage shelf as its use as an insulation tester is just too limited by the design decisions for my liking.