LCR Cal Test PCB

[KungFuJosh]

PCB finished size ?

150mm x 120mm, not too small.

[Martin72]

That leaves the vexed question of calibration....
I see several possibilities.
The first option is to send in the assembled board and have it measured.
Second possibility:
Those of us who are "blessed" with an LCR bridge that has 0.05% basic accuracy and is therefore well below the component tolerances, use their device as a reference - But would also have to have it calibrated first to be on the safe side.
Then, of course, there is the third option:
You don't care, the main thing is that you have something to measure.

[mawyatt]

If the PCB can fit inside those small USPS boxes that ship for ~$5 anywhere in US, you could have folks around the country make measurements with different meters for comparisons, kinda like folks do with the Voltage Source Cal group.

Also consider a thermistor on the PCB.

Best,

[KungFuJosh]

If the PCB can fit inside those small USPS boxes that ship for ~$5 anywhere in US, you could have folks around the country make measurements with different meters for comparisons, kinda like folks do with the Voltage Source Cal group.

Also consider a thermistor on the PCB.

It should definitely fit in the small flat rate priority box which is ~$8 to ship anywhere in the US.

I've never used a thermistor before. What type should I get, and how should I set it up? ...and is it really better than measuring ambient temperature?

Thanks,
Josh

[mawyatt]

Wow that flat rate has gone up since last we used!!

Thermistors are used to measure temperature of the thermistor element, and are usually inexpensive (<$1) and most folks have a DMM that can use them. If the thermistor is mounted to your PCB then the readings should give a rough indication of the PCB temperature.

Best to study up on thermistors (lots of on-line tutorials) as there are an almost infinite variety of such!!

Thermistors are non-linear and most DMMs use Steinhart-Hart equation:

1/T = A + B(lnR) +C(lnR)^3, where T is Kelvin and A, B and C are Thermistor Parameters.

Common types are NTC 10K and 5K @ 25C.

Here's an example SMD from TDK.

https://www.digikey.com/en/products/detail/tdk-corporation/NTCG163JF103FT1/614632

Best,

[KungFuJosh]

Here's an example SMD from TDK.

Thank you, that's what I needed. I understand the (very) basics of thermistors, but there's a wide range of accuracy, value, etc. I didn't want to guess at what specs would be acceptable.

I don't want to use anything smaller than 0805, I assume this should be fine?: https://www.mouser.com/ProductDetail/594-NTCS0805E3103FLT

Should I use the same connection setup as the other solo items?

Thanks,
Josh

[mawyatt]

Suspect that should work with most DMMs. Since the expected PCB temperature should be 25C +-5C this should give a reasonable indication of PCB temperature.

Of course humidity also becomes a factor, but that's another level to consider as you dive down the Rabbit Hole of Precise Measurements

Edit: Should have mentioned no need for 4 wire measurements to Thermistor, 2 wire measurement should be fine.

Best,

[KungFuJosh]

I'm changing the 1mF to 2824 (was 2812). That way people can choose between either 10% at 6.3V or 20% at 10V. The old one was 6.3V 20%.

[KungFuJosh]

Assuming no glaring issues, this may be the final version of the V5 board. Feedback is always welcome.

I also attached 2 BOMs. One for Mouser, and one for JLC PCBA service. Most of the JLC parts seem fine, a few I'm not sure about (haven't heard of a couple brands before). They also didn't seem to state the value for their thermistor. I might mix and match parts from both sources, or skip the PCBA.




ETA: After seeing what it would cost to preorder parts for PCBA, I will not be using that service. Yikes. 

Thanks,
Josh

[mawyatt]

Just noted, don't forget holes in the corners for nylon standoffs.

Best,

[KungFuJosh]

Just noted, don't forget holes in the corners for nylon standoffs.

Fiiiiiiiiiiiine. 😉

Is 3.5mm good for you?

[Martin72]

@Josh:

1mm drill hole..

Will 1.5mm still work for you? I don't mind changing it, do you have the sheet for the pins?

They will then sit loosely in it, but yes, 1.5mm is also possible.
Data sheet/origin is at work, I won't get to it for another 2 weeks.

[KungFuJosh]

I noticed an issue with my dissipation factor test results. For some reason when I ordered the parts for V4, I substituted the 50Ω with a 47Ω and the 160kΩ with a 158kΩ. Those two test results were lower than they were supposed to be, which is not surprising.

I corrected those values on the Mouser BOM, and here's the project link: https://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=6851743830

All parts come out to less than $30, which is nice. Unless you want test pins, and then it will cost twice as much. 😉😉

The 1mF cap is backordered until August, but I'm guessing that won't matter.

Here's (hopefully) the final V5 board version:


Thanks,
Josh

[KungFuJosh]

I ordered the new PCBs and parts. I should be able to make a new board next week. I ordered some extra PCBs in case anybody in the US wants one.

[KungFuJosh]

I got the PCBs in. Now I just need to throw some stuff on there.

[Martin72]

That went very quickly

[mawyatt]

PCBs look nice

Check out the repeatability of those Short and Open Calibrations spots. Do a Short and Open Cal after everything is stable, then make Short and Open readings at various frequencies, record, then disconnect/reconnect and repeat, then swap Kelvin Leads (Red and Black), do again.

This should give a good indication of how well your Short and Open Cals are behaving.

Best,

[KungFuJosh]

PCBs look nice

Check out the repeatability of those Short and Open Calibrations spots. Do a Short and Open Cal after everything is stable, then make Short and Open readings at various frequencies, record, then disconnect/reconnect and repeat, then swap Kelvin Leads (Red and Black), do again.

This should give a good indication of how well your Short and Open Cals are behaving.

Thanks!

I'm too lazy for that. I'll just send you a PCB and you can let me know. 😉

[Martin72]

How many of them do you have?
 

[mawyatt]

PCBs look nice

Check out the repeatability of those Short and Open Calibrations spots. Do a Short and Open Cal after everything is stable, then make Short and Open readings at various frequencies, record, then disconnect/reconnect and repeat, then swap Kelvin Leads (Red and Black), do again.

This should give a good indication of how well your Short and Open Cals are behaving.

Thanks!

I'm too lazy for that. I'll just send you a PCB and you can let me know. 😉

Ok will do. If you want we'll populate the PCB and record the measurements, then return to you for your verification measurements.

Best,

[KungFuJosh]

Ok will do. If you want we'll populate the PCB and record the measurements, then return to you for your verification measurements.

I already populated one, but I can send it to you along with an unpopulated PCB for you to keep, and you can ship mine back after testing if that works for you.

I'll hopefully have time to run through the tests in the next day or two.

How many of them do you have?
 

I might have ordered 20. 😉

Thanks,
Josh

[Martin72]

I placed an order with Mouser earlier - what was the thermistor good for again?
I forgot to order it earlier...

[KungFuJosh]

I placed an order with Mouser earlier - what was the thermistor good for again?
I forgot to order it earlier...

The thermistor was to make Mike happy. 😉😉

...but I guess also to test the temperature of the PCB.

[KungFuJosh]

Here's the results comparing my ST2832 and ST42 on V5 of the test board. Most tests were matched at 360mVrms (ST42 tested at about 358mVrms).

The DC bias values are based on 1kHz or 10kHz frequency for AC. It matters for some reason. 🤷

I think I did step DF test at 1Vrms. I forget, that was way back to yesterday.

Thanks,
Josh

[Martin72]

Interesting results.
It is also easy to see how the reactive part of the resistance increases at higher frequencies, as a result of the ever decreasing real value.
Or how at higher frequencies (100, 200kHz) the accuracy decreases (up to 4%).
I want to reproduce this as soon as possible.