DIY Precision Reference for DVM Calibration

If you have a home lab with electronics like DVMs this might be of interest. Check details out here at Photomacrography.

https://www.photomacrography.net/forum/viewtopic.php?p=261783#261783

Quick test circuit for LM399

A pair of HP/Agilent  34401A DVMs

Custom PCB 3D Models

Best,

Hi Johannes,

Very nice looking reference assembly, well done!!! Is that a custom 3D printed case, looks really nice and create a nice overall package. Something (3D printing) I wish I had and may consider getting involved with.

How is the long term drift with the LT reference?

I've been running the LM399 for about a week now continuous to let it stabilize, so far only about 100uv change.

There is a LT reference diode similar to the LM399 called the LTZ1000 that's even better than the LM399, but these are very expensive whereas the LM399 is still under $10.

Wanted to include reference resistors and currents in the design, voltage, currents and resistances could be compared. So 100, 1K, 10K, 100K, 1M ohm precision resistors and 100ma, 10ma, 1ma, 100ua and 10ua currents are supported.

BTW still waiting on the precision components to arrive, then will decide if the PCB should be fabricated, there is a minimum build quantity of 5, so a few "extras' will be available.

Best,

Bernard,

Agree these are very nice parts, they are the references used in the HP/Agilent 34401A DVMs shown, but special selected versions. Initial accuracy isn't that good, but drift with temperature and time is exceptional.

I also have lower resolution Fluke 77 and 87, they all agree well. The HP & Agilent DVMs calibration is a complex process and not just adjusting a pot, the calibration coefficients are stored in NVMemory. Soon I may get these DVMs calibrated, then use the LM399 reference calibration so I'll have a secondary standard on hand.

Best,

Johannes Zander wrote:

Hi Mike,

thanks for your nice comment.

I am not so scientific. I didn't do an long term drift test. I observed that it drifts between 9.9992 and 9.9995. This is precise enough for me.

I searched a bit about burn in for aging electronic devices and found this forum:

https://www.eevblog.com/forum/metrology/lm399-based-10-v-reference/250/

They talk about paralleling LM399 for reduced noice.

Good look with your research.

Regards

Hi Johannes,

I've observed the LM399 random drift over the past few days, it's now below ~30uv. So seems to be "settling in". The reading is centered around 6.85498 volts at ~1ma. This part may be an "outsider" since the norm is 6.95 volts for the LM399, and thus may not be a good representation of the LM399 (maybe a reject that I got off eBay??).

Thanks for the link. Paralleling devices not only improves low frequency noise but also drift.  If one wanted to create a serious standard voltage reference then paralleling devices seems like a good approach, but the LTZ1000A is a better device for this. These are expensive tho (why I don't have one).

BTW what DVM are you using?

Best,