Hioki DM7276-01 Precision DC Voltmeter (9 PPM)

[Alex Nikitin]

Here is another measurement result using the DM7275, in this case I was trying to do a 1000:1 transfer of resistance, from Fluke 742A-10K to my 10 Ohm DIY shunt (made using a TO-3 packaged Vishay Foil resistor). The Keithley 263 is used as 1mA current source to supply the Fluke and 10R shunt in series, the voltage on the Fluke was monitored by HP3456A and was stable at 10.00046V (double checked with the Hioki), the Fluke 742A-10K has a measured resistance of 10.00005K so the current was about +41ppm from 1mA. As a control option the ESI SR1010 resistor was used as a 10R shunt in place of my DIY one. Despite some meters of cables and numerous connections, the circuit offset at the shunt was measured below +/- 50nV average, 250nV p-p variations (the Keithley 263 has some useful features to change the current polarity and also switch it on/off so it was easy to measure offsets without disturbing the wires and connectors). The graph below shows two 15min runs with two 10R shunts in this configuration. The temperature variations (measured by Hioki) were below 0.5C). A linear trendline was added to each run to indicate a possible drift in time. Vertical scale is 100nV /div (=10nA/div, or 10ppm of value) , horizontal is samples/time , 1min/div, (=15 samples).

Cheers

Alex

[HighVoltage]

Your 1000:1 transfer results are very impressive with the old Keithley 263 as a 1mA source !
Please show a picture of your complete setup.

[Alex Nikitin]

I had to tidy up the cabling a bit  and let it settle for 10 min or so. Also today the lab temperature is somewhat higher and the HP3456A shows 10.00044V on 10K Fluke (and that is the HP tempco shows, the Keithley is stable as Hioki shows 10.000463 on the Fluke).

The Keithley 263 is a very stable and capable instrument even at higher currents, my unit is repaired and fully recapped, and for most part I use it over the newly acquired HP3245A, mostly for convenience reasons but it is not that far in DC performance and similar in the LF noise area.

Cheers

Alex

P.S. - on the photos, the first one is measuring the offset only (the current source is OFF), the second is +1mA current, the third is -1mA current. The Keithley is very symmetrical in polarity switching, due to its architecture.

[Alex Nikitin]

Here is a longer run (1mA from K263 into 10R DIY shunt) in the night time, the temperature change was about 1C over that period. The orange line is averaging for 5min (75 samples) . Vertical scale 50nV/div, horizontal - time.

Cheers

Alex

[Alex Nikitin]

Here in an interesting result with the same 10R DIY shunt and Keithley 263 supplying +/-100nA, so the voltage measured is +/- 1uV, with the HIOKI on FAST (NPLC1) speed (~30s/sec), 10 sec OFF, 10 sec +100nA, 10sec -100nA and last 10 sec OFF. The orange line is 5 samples moving average.

Cheers

Alex

[Alex Nikitin]

I rather like the temperature measuring feature. Strangely, if you copy data to a USB stick, the temperature measurements are what is on the screen, with 0.1C resolution. However if you request the data remotely, the resolution is 0.01C which does make a lot of difference! Below the results of an overnight run with the Hioki measuring the Fluke 731B 10V output and the room temperature at the same time. I plan to use the DM7275 together with a small DIY "temperature chamber" for tempco measurements on resistors and references.

Cheers

Alex

[EC8010]

Going back to the lovely internal pictures, the difference between the 7276 and 7275 is that the black dot is in the A box on the voltage reference board whereas it's in the C box on the 7275. Looks like they graded the stability of the reference board. Which seems entirely reasonable. Perhaps more significantly, it might mean that the DMM's drift is dominated by the voltage reference rather than input stage and ADC noise/drift? If so, it might be possible to improve a 7275.

[HighVoltage]

Going back to the lovely internal pictures, the difference between the 7276 and 7275 is that the black dot is in the A box on the voltage reference board whereas it's in the C box on the 7275. Looks like they graded the stability of the reference board. Which seems entirely reasonable. Perhaps more significantly, it might mean that the DMM's drift is dominated by the voltage reference rather than input stage and ADC noise/drift? If so, it might be possible to improve a 7275.

There is almost a $1000 price difference between the DM7275 (20 ppm) and the DM7276 (9ppm) models
Resolutions are the same.
You might be right, same as the differences in the Keysight 3458A models.

BTW, I used the Contact Check Function of the 7276 for the first time consciously.

[chekhov]

BTW, I used the Contact Check Function of the 7276 for the first time consciously.

Eh, having so much space on display they show "NoCntct" instead of "NoContact", lame 
Also, from the first seconds this is a little bit of cheating on low voltage input - "guessed" DMM in 10M input mode.
If it was in auto or was a different similar class DMM - it's evident when the input is not connected by readings jumping all around. Of course, for automated measurements that could be much more versatile function.

[MegaVolt]

BTW, I used the Contact Check Function of the 7276 for the first time consciously.

How does it work? What would it show if the battery was working?

[chekhov]

Here is a little description from 158 page of user manual, provided in the beginning of this thread:

[MegaVolt]

Here is a little description from 158 page of user manual, provided in the beginning of this thread:
(Attachment Link)

If we put a thin insulator between the terminal and the battery, will the test pass?

[wanghar]

How to calibrate a DM7276/7275?  Does this descriped anywhere in the manual?

[Alex Nikitin]

How to calibrate a DM7276/7275?  Does this descriped anywhere in the manual?

Calibration is described on page 212 of the user manual, however Adjustment (next page) is password protected and not accessible. It would be good to know the adjustment password!

Cheers

Alex

[wanghar]

 Is this Hioki's style that not allowing users  to do calibration  by themselves? 

[Alex Nikitin]

Is this Hioki's style that not allowing users  to do calibration  by themselves? 

Not just that, so far I can not get any calibration information on my new meter from them or from the seller (Farnell UK) . And I am not even sure where I could send the meter for calibration! (OK, calibration is simple and any decent calibration lab should be able to do it, however if the meter needs adjusting, it might be a problem) .

Also the 3-year manufacturer's warranty on these meters is not transferrable from the original owner, so if you sell your meter or try to buy one second hand, there is no warranty even for a relatively new meter. Don't get me wrong, I like my DM7275 , use it all the time and won't swap it for almost any other meter, including the HP3458A, however the customer service from Hioki Japan is essentially non-existent and even from the distributor I can not get much sense so far in regard of calibration data.

Cheers

Alex

[HighVoltage]

I had tried a few times to the password from Hioki Japan but so far I had no luck.

[Alex Nikitin]

I've measured the DCV output of my HP3245A from -10V to +10V in 1V steps with the HIOKI DM7275. The measurements were made at 10V range, SLOW speed, 10 averages per point, with the input cables polarity reversed for the second run. Vset = voltage output setting on HP3245A, V = measured voltage with the correct polarity, Vr = measured voltage with the reversed polarity. V+Vr is the difference between measuring the same output voltage in two polarities. There is a small meter/cables offset about -2.5uV which I did not null. Vertical scale is 0.5ppm = 5uV per division.

Cheers

Alex

[Alex Nikitin]

And here is a 24h measurement run of 10V output from HP3245A, there is a visible tempco of about -0.5ppm/C for the two units and connecting wires. The room temperature (orange line) varied from 24.5C to 25.7C. Right now I'm running the same measurement for 1V output.

Cheers

Alex