Electronics > Metrology

PMHD2024 (Poor Man's Hamon Divider)

(1/2) > >>

Alex Nikitin:
Some years ago I've bought a lot of about 50 precision resistors, Vishay S102 bulk foil, quite cheap probably due to a strange value - 87.15K +/-0.01% . As far as I know, these resistors were used for a sensor calibration, each supplied in a separate nice case as on the photo below. For a long time I was trying to find a way to use these and now I've built a Hamon divider. More accurately, not a complete divider but only a base for it, a chain of 12 selected resistors in a box. I've called it PMHD2024, as per the thread title. I've used a cheap Hammond box and a cheap D15 socket.

The first unit is hand-soldered however I think now about using a PCB to simplify the build. In some way it is a very basic equivalent of the SR1010 unit, much smaller but providing for similar options. The D15 plug is also inexpensive and I will build plugs for various applications, for now only 10:1 divider plug is built (as on the photo) however a Hamon calibration plug, 1:2 divider and even a plug with a DIL switch for connecting to various points of the array is easy to build. This base unit can be used for INL testing as well, I suppose, with the right switching arrangement.

My first tests show less than 10ppm error for 10:1 division and very stable performance. The idea of using a D-type connector is to get rid of expensive 4mm terminals and switches, the performance of D-type in my experience is good enough for this application down to ppm levels. I will try to run some stability tests in the next few days.



Alex Nikitin:
Here is my first attempt in using my reptile box temperature chamber with my new HIOKI DM7275 voltmeter, to measure the tempco of the 10:1 divider made with PMHD2024.
The input was 10V from Fluke 731B (the actual value measured by Hioki is 9.99994V, so the divider error is about 10ppm to 14ppm). The tempco appears linear between 20C and 28C, about -0.6ppm/C.



P.S. the "notch" on the graph is me opening the chamber door for few seconds  :palm:

Your setup seems to be good for 0.1 ppm noisewise, except at the end of the sweep after more than two days (!) there seems to be a difference of 1 ppm. The cause could be some temperature lag. One could ground the aluminium enclosure from the outside and use the two pins for a thermistor inside the enclosure.
I'm curious to see two sweeps with the Hamon plug (division ratio 1:1, two directions).

Regards, Dieter

Alex Nikitin:
The Hioki DM7275 is a very nice voltmeter, it almost needs the 8th number, however for my needs 7.5 digits is probably enough. I am slowly building a number of shunts for it, the Hioki can store 30 setup "screens", including accurate calibration constants for shunts, for example and with the HP3245A it makes a very decent Ohmmeter. If you don't need AC capabilities, the DM7275 is a very good value for money, you get almost HP3458A noise and stability levels with no fan noise and a better connectivity. And for about 1/10 of the price for a new unit with warranty. There are some things I don't quite like about the Hioki meter but nothing that I would consider a show-stopper.

The whole cycle was about 2 hours (not two days!) and the difference is most likely just a thermal lag. The temperature sensor from Hioki meter is fixed to the divider case with a copper tape and actually works better than expected. Unfortunately, while the reptile box controls the temperature quite accurately, the rate of temperature change is not controlled and it warms up much faster than cools down. Still a useful tool and with the new meter is easy to use (Hioki takes both measurements, voltage and temperature, simultaneously).



doktor pyta:
Hi Alex,

have You checked the insulation resistance between two nearby pins of the DB connector /K617/?
Are there any D-subs with PTFE insulation ?


[0] Message Index

[#] Next page

There was an error while thanking
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod