Di-LOG DL9307 Insulation Tester Review

Introduction

Continuing with my insulation tester series, I have changed the rules for this review slightly. Whilst the earlier reviews showed different offerings from manufacturers, this one will be a review of two very similar meters, detailing them in comparison to one another, rather than a comparison against the previous testers reviewed.

Keysight Insulation Multimeter Failure

Megger MIT420/2 Road Test

Chauvin Arnoux CA6526 Insulation Tester Review

Flir IM75 Insulation Tester Review

Extech MG302 Review

The DL9307 from Di-LOG Instruments looks to be the same design as the MG302 from Extech. Due to the troubles, I was having with the Extech instrument, I decided to get hold of one of these to see if it could aid with trouble shooting and just out of curiosity off the similarities between the instruments. This would also help me finish off testing out the software side of the MG302/DL9307.

Di-LOG Insulation Tester Range

Di-LOG is a UK based company, and as far as I know they do not design and manufacture their own instruments, instead rebranding outsourced instruments. They only offer three standalone insulation testers as seen above.

A bit of research shows that the MG302 / DL9307 are, in fact, manufactured by CEM Instruments, a Chinese based company, as the DT9985. Closer review of the items offered by CEM, shows that the complete insulation tester range offered by Di-LOG appear to come from CEM. There are also numerous similar items in the Extech range and also some Flir look-a-likes.

CEM Insulation Tester Range

Unboxing and Overview

The Di-LOG DL9307 arrives in a cardboard box, packaged in a much more compact way than the Extech MG302. Inside, the meter and accessories are contained within an egg carton style box, no protective case is supplied with this instrument, in stark contrast to the MG302, that came in a robust hard plastic case with foam padding around the contents.

Unfortunately, this becomes my first grumble. In all other aspects, the DL9307 compares directly to the MG302, the same accessories in both quantity and type arrive with both of the instruments. But as seen in the comparison chart below, the DL9307 has a recommended retail price substantially higher than the MG302, but comes without any form of case.

In reality, I picked both of these instruments up at prices well below the RRP, but the DL9307 was still £50 more expensive, meaning that, in my opinion, it is of much poorer value for money in comparison to the MG302.

Technically, the DL9307 offers the same specifications as the MG302, but I have posted the comparison table above to show how it compares against the other insulation testers I have reviewed.

Performance Tests and Comparison

Being the same meter as the MG302, I expected to find that it would perform to a similar standard. Comparing the output voltage at 500V between the units shows a no significant difference over a 10GOhm test resistance range.

Expanding the tests over the voltage range of the instruments and looking at the current control aspects starts to show some differences between the meters.

As with the voltage regulation, the output at the different voltage levels are directly comparable. The DL9307 shows a little more variation with regards to the 1mA load test, but still remains compliant with the IEC standard requirements.

The big surprise is the short circuit current of the DL9307. I had already found that the MG302 does have a high short circuit current, but this still remains compliant to the IEC standard. The DL9307 was found to have an even higher short circuit current that significantly exceeds the 15mA upper limit of the IEC standard. The short circuit current is that high, that it exceeded the full scale range of my insulation tester calibrator and I had to measure it with an external ammeter.

Having such a high current capability, I was interested to take a look at the voltage overshoot and sure enough found it to be much worse than the MG302. The 500V range on the DL9307, saw a 38% overshoot up to 693V in comparison to the 15% overshoot for the MG302.

During the testing, it was found that if the short circuit was maintained for a while, the output current would then get reduced down to 5 to 6mA over a 40 second period.

The behaviour is very unusual, and the first time I have seen it from an insulation tester. I am more used to a constant short circuit current as seen in the plot for the MG302. Even the higher voltage output testers that I use, that sometime have a higher current output, will have a constant output. The decreasing output of the DL9307 doesn't make much sense to me, as when a short is found, the insulation test is stopped, probably before the current reduces to more acceptable levels.

In the plots, the current values are a little higher than measured with a meter, I believe that this is due to a calibration issue as the current was derived in the oscilloscope by measuring the voltage drop across a 1 ohm resistor and creating a custom probe within the Picoscope software to carry out the current conversation. The resistor measured slightly above 1 ohm, therefore giving the slightly higher current readings in the oscilloscope.

The question of course, is the exact reasoning behind the IEC standard for limiting the short circuit current to 15mA. The effect of current on the human body is detailed in IEC60479-1 within the UK and IEEE Std 80 in America. The effects of current are dependent upon many factors, but a generalised table is available. This identifies the current for a specified duration into categories. From the plot 15mA for over 1 second is perceived as entering category AC-3 where a human will start to feel pain, 25mA only requires 0.5 seconds to reach AC-3 category. If exposed for longer than 10 seconds, 25mA gets close to AC-4.1 category, which could be fatal to 5% of humans, I cannot imagine though, someone keeping hold of an insulation tester for 10 seconds, if they were getting a shock from it.

Ultimately, whilst it exceeds the recommended standards, I think it would be hard to believe that someone would be hurt by it. The issue has been raised with Di-LOG and Extech. Di-LOG have never responded. Extech have responded and sent me a calibration procedure, but it does not include adjustment of the short circuit current and they seem to be unable to tell me what the short circuit current should be.

Back to the performance tests and a set of resistance reading tests were carried out using the insulation tester calibrator.

The deviation plot shows a set of results slightly better than the MG302, that deviated between +/-1%, compared to the deviation of the DL9307 between +0.6% and -0.8%.

However, an overall averaging of the test results shows that the MG302 had a better average accuracy than the DL9307. The main culprit for the deviation on the MG302 was the 125V range, with the other ranges showing a much tighter set of results. With the DL9307, the spread of results deviated across all the voltage ranges leading to the higher average accuracy result.

Battery consumption tests also produced some deviations between the two instruments. For the current measurement mode, the DL9307 drew less current from the battery, but drew slightly more during the Diode and Continuity tests. All other ranges produced comparable results.

Battery use between the two units is similar. The MG302 did operate down to 3.30V where as the DL9307 cut out at 3.80V. The backlight also stopped working at a higher voltage of 5.00V for the DL9307 in comparison to 4.10V for the MG302. The low battery warning came in at 6.80V on the MG302, but came in lower on the DL9307 at 6.40V.

Multimeter Functions

These meters are much more than just insulation testers and contain respectable range of standard multimeter functions that I though I would go through and compare to one another. I set up a few bench test scenarios to test the Voltage, current, resistance, capacitance and temperature functions.

Test set up varied depending on the function, my AC sources can only go u to 300V and my DC source 200V, so I then utilise them to drive high voltage boosters. During this, the voltages are also measured by an 8846A, to provide a direct voltage reading and not rely upon the ratio of the boosters.

Micro and milli amps can be derived direct from my sources, the SMU is accurate enough to do this directly, for the AC I use a burden resistor, driven form the AC source and again cross-reference the readings to the 8846A. The SMU was also used to simulate the temperature readings utilising a conversion table for a K-Type thermocouple, readily available off the internet.

For higher currents above 1A, I can use an SMRT1 injection test set from Megger, that will inject both AC and DC. This has an adjustable frequency output from 0Hz to 1kHz and was also used to test the frequency response of the AC voltage and current functions.

Resistance and capacitance functions were mostly checked via the decade boxes, these values were also verified using the 8846A.

A mixed bag of results were seen across the multimeter functions, with MG302 sometimes more accurate than the DL9307 and vice-versa. The MG302 was predominately more accurate with the voltage and current functions and the DL9307 more accurate within the auxiliary meter functions. However, on occasions, it was found that the DL9307 would be more accurate at specific ranges within the voltage and current functions.

With all functions averaged out, there wasn't much to choose between the two units.

Winding Simulator Tests

Back to the insulation testing functions, the phase balance and polarisation index test comparisons on the winding simulator can be seen below.

For the majority of the winding resistance tests, the DL9307 shows results closer to the nominal values than the MG302 does. The DL9307 also appears to show more consistent results across the three phases than the MG302 does.

Polarisation index tests and the insulation plot are comparable between the two units apart from one point in the plot at 45s, where the DL9307 produced a blip in its response. This sometimes happens during insulation testing and is generally not considered to be an issue. For a single deviation, some test engineers will regard the result as an outlier and manually smooth out the curve.

Motor Tests

I took the DL9307 out to the motor store to see how it would perform on a couple of motors. The first picture below is of a 45kW motor on test and the second, the now familiar 90kW motor tested in all of the previous blogs.

The DL9307 tested both motors without issue and returned results similar to the other insulation testers reviewed. For both motors, I elected to save the data to the local memory of the DL9307 and download it to the computer at a later date. The phase balance tests were carried out and saved first, followed by the polarisation test.

The DL9307, had the same operating quirk as the MG302. The phase resistance readings were saved directly using the manual store function on the meter. The polarisation index test was saved using a 15 second interval, that was set up first, then switching back to the insulation testing function to initiate the PI test, to finally return to the memory function to initiate the data save.

As the DL9307, does not have a specific DAR or PI functionality, the test is manually timed and stopped after the 10 minutes.

Software Operation

The main reason to purchase the Di-LOG was to carry out further testing on the software functionality after the Extech unit had failed, so back at the computer, the data from the motor tests was downloaded all at once for both motors and all types of tests. This produces and unusual shaped graph, that doesn't really mean anything as seen below.

To make any sense out of the test results, the only option was to load the data up into Excel and format it within the spreadsheet to allow the tests to be displayed individually. The supplied software is really only useful for displaying an individual test. This limits the functionality of the meter if there is no other way to manipulate the data, as each saved test would need to be downloaded and then deleting from the meter memory before carrying out the next test.

I appreciate that these are budget meters but having a few basic data selection functions in the software would be a great enhancement in my opinion.

The software does automatically save the data in an Excel format as well as its own database file type. However, upon opening the file within Excel, the data units do not translate correctly and must be manually entered. Something that is easily rectified, but adds a further step to the data manipulation. The spreadsheet screen capture below, shows the raw data downloaded from the meter, that I then split up into the different motors and different types of tests.

Reformatting produces an acceptable test result sheet.

Instruction Manuals

A quick mention of the instruction manuals provided. The MG302 manual is included in 7 different languages in an A5 format, the Di-LOG manual is only in English in an A6 format.

I didn't find the manual for the DL9307 to be written that well, it seems highly likely that this is a poor translation into English. The manual for the MG302 has been re-written and makes much more sense.

There was also an error in the Di-LOG manual regarding the clearing of the memory, the manual specifies the use of the 'Exit' key which does not work. The instructions are correct in the MG302 manual, that states the use of the 'Range' key, which works on both units.

Build Quality

I have aimed this section to be a direct comparison between the MG302 and DL9307. They both appear to be a CEM designed device and the board markings show this, so I was expecting them to be of a similar built quality.

The first difference noticed is with the instrument front, the different colour is obvious, but as you look, more subtle differences start to appear. I find that there is more clarity to the text on the buttons of the DL9307, the function switch selections are also easier to read.

Around the jacks, the MG302 uses colour identification for the insulation and current test terminals, where as the DL9307 sticks to the normal text.

A minor gripe, but I feel it is typical of the quality between the two units, is that the MG302 is supplied with Energiser brand batteries, but the DL9307 came with GoldenPower brand, that have already started to show a voltage drop.

I did note that that the MG302, came with a serial number. I could not find a serial number on the DL9307, quite a surprise for an instrument aimed at professional use that these days requires traceable calibration proof, that almost mandates the use of a serial number on the instrument.

The back cover is removed in the same way for both units and the view inside is also pretty much identical. Confirmation that both of the units are in actual fact DT9985 is found at the bottom of the top board.

A difference in the quality of the units was immediately noticed when looking at the transformer for the HV output. The securing clips for the transformer were missing on the DL9307 and there was evidence of solder rework around the transformer and on its connections. Solder spatter was evident around the base of the transformer, there was also misalignment of other components and scratches on the PCB surface and some of the components. It gave me the impression that the transformer had been replaced on the DL9307, although not very carefully. Both units use the same SIBA fuses and connection to the input jacks is made in the same manner, which looks to be strong and reliable.

The current shunt on the DL9307 looks to have been attacked, there is more solder around the shunt than on the MG302 unit and the finish is very jagged. On both units, there is very little clearance between the shunt and the common input jack.

At the top end of the meter, the wireless communications board can be found.

These units look to be different and I could not get the MG302 to work with the DL9307 receiver and vice versa. The board on the DL9307 looks to be a later design with a shorter aerial. The USB receivers also looked to be different versions, with the Di-LOG unit again having a shorter aerial and a different PCB layout. This was a bit disappointing as I hoped to be able to use the DL9307 to try an determine what was wrong with the MG302 wireless communication.

With the HV board removed, a insulation sheet is found between the back of the board and the processor board that sits beneath it. This remains attached in the MG302 unit but can be peeled off. In the DL9307 unit, it fell off when the board was removed. The insulation sheet is also roughly cut out in the DL9307, showing torn edges and an access hole for the trimmer pots not cleanly cut out.

Underneath the insulation board, a MOSFET can be clearly seen and there was a noticeable difference between the soldering of this, with the MG302 have a fully soldered tab, but the DL9307 attached with much less solder.

With the HV board removed, access to the processor board is obtained. Layout of the boards looks to be the same, some of the components within the DL9307 are a little more misplaced than in the MG302 and the internal EMF shield looks to be a little more battered.

A slight difference was noted around some shunt resistors in series with the 10A fuse. These look to be hand soldered in and may be part of the calibration of the units. They do not look to be part of the insulation test circuit, so I don't think that they are the culprit for the large difference into short circuit current between the two units.

Some heat damage was observed on one of the connectors and a case boss on the DL9307 unit, that were not present on the MG302. It does not appear to be causing an issue with the connector, but is another pointer to the apparent lower quality of the DL9307 in comparison to the MG302.

Stripping down the unit further, allowed the processor board to be removed to reveal the inside of the front case. I did find that there were rubber off cuts from the case within the MG302, but the DL9307 was all clean.

Looking at the DL9307 on the left, it seems to me that at some point, the EMF shield has been bent upwards to gain access to the components below. Looking under the shield, I could seem some resistors paralleled up, so may be some more calibration adjustment has been carried out.

Overall, neither of the units offer a 'Rolls Royce' approach to insulation testing and of course, I only have one unit of each of them, but based on what I have seen, the DL9307 certainly seems to have a lower quality build when compared to the MG302. I do not know if the units are completely built by CEM, or whether they just supply components and Extech and Di-LOG have their own plant, or contract out, the assembly and final checks.

Meter Failure

Two issues have occurred with the DL9307 during the review process. The first that occurred was the failure of the tilt stand, where one of the hinge points snapped off whilst I was removing the battery cover. I am not sure exactly why this happened, I don't believe the stand was damaged from manufacturer, I think I must have just twisted it badly whilst working on the unit. Repair however, was reasonably simple, using araldite to secure the broken part back onto the stand.

The second failure was more serious and occurred during the multimeter function tests, when the unit failed whilst measuring 1000V AC. A small pop was heard and the side of the meter felt warm. It was also observed that the voltage reading on the DL9307 was now much lower than the applied voltage.

Dismantling of the unit revealed a physical failure of SPG4 on the input circuit of the DL9307. The identification markings on the damaged component, revealed that it was a varistor type VDR 07D621K, manufactured by Bourns. Some tracing out of the circuit showed that SPG4 was connected in series with SPG3, so I decided to replace both of these, incase SPG3 had been stressed.

I could only get the Bourns varistors of that type on an extended delivery within the UK, so temporarily I opted for a varistor from EPCOS. This had the same clamping voltage as the Bourns unit, but had a lower energy dissipation. EPCOS do manufacture a varistor of a higher energy dissipation, but it is physically bigger and would not fit within the meter case. I therefore, only intended this to be a temporary repair, to see if I could get the unit back up and working. If it works I will get the right varistors, or may even scrap the unit and transfer the wireless module over to the Extech unit.

The repair did return the unit to a working state, all be it slightly out of tolerance. This was not expected, as in the past, I have only seen varistors across meter inputs and not specifically part of the measuring circuit. However, the repair allowed me to finish obtaining the test data that I wanted.

Conclusions

The DL9307 pretty much performs as per the MG302, which is of no surprise. It was disappointing that I could not utilise the wireless modules to try and identify the fault on the MG302. Although the units were purchased in 2018, within 4 month period of one another, I believe the MG302 is old stock that a supplier was getting rid of. I had to wait for the DL9307 to come into stock, so believe that this is likely to be a 2018 build.

Most of the functions between the two instruments were comparable. The one big difference being the very high short circuit current of the DL9307 that is not IEC compliant and has a unique profile during sustained short circuit faults. Insulation test accuracy was found to be better on the MG302, accuracy of the multimeter functions was split across the meters. As it is the same design, the DL9307 suffers from the same issues, with a low insulation resistance capability for the type of testing that I do. This is such a shame, as the data recording and 4 to 20mA functions are implemented really well and are useful to me.

The software appears to be the same for both of the units, the DL9307 did communicate with the MG302 software. The functionality is the same, the software is basic and very easy to use. This however, limits the reporting capabilities within the software and relies heavily on the ability to export the data to spreadsheet software. This becomes a necessity when multiple tests are saved directly to the meter, as it can only be downloaded at once and must then be split up within a spreadsheet to make sense of it.

The Di-LOG DL9307 has turned out to be quite a messy meter. The meter appears to be a slightly later version of the Extech MG302, both manufacturers by CEM. I was quite surprised at the difference in quality between the units, with the DL9307 revealing a lower quality build that also appears to have had rework carried out within it to an even lower standard. Added to this that it comes without a case, and within the UK, comes at a higher price than the Extech, and lack of support from Di-LOG, I really don't see why any one would purchase this unit.