DIY milliohm/volt meter kit

[jaxbird]

Here a quick test before final assembly:



Almost everything worked when I first powered it on, no excess current draw, programming the mcu was fine, usb worked, battery charging was fine etc. But I did have a problem with the milliohms measurements, took a bit of investigation, source current was ok, and it turned out to be a 0603 capacitor on the sense inputs that was shorted Easy to find as the kit comes with a complete schematic. 

Just moved it around a bit with the iron and everything was fine.

Then a good flux cleaning and ready to put it in the box.

There doesn't seem to be any support for the display, other than it will just sit on the board. I used a couple of small strips of thick double sided tape to secure it at a good height to fit in the box close to the top plate.

Here the complete device:



Small detail, the included screws for the top plate were too short, but I had some other ones that would fit, so no big deal.

Edit: Note the kit does, of course, not come with a battery, so you'll need to find one yourself. I found the battery in my old Sony/Ericsson phone still worked fine, so I used that.

[Vgkid]

Thanks for the mini review, it looks pretty neat. I guess i will need to start messing atound with smd components soon.

[jaxbird]

Thanks for the mini review, it looks pretty neat. I guess i will need to start messing atound with smd components soon.

Thanks, yeah, it's a useful little device. But I'd recommend at least some experience with SMD before having a go at this kit.

For milliohms, I used to use a power supply in CC mode adjusted as close as possible to e.g. 100mA or 1A using a reasonably accurate meter, then with a pair of Kelvin clips measure the voltage over the resistor.

Now I don't need to do that anymore Plus it doubles as a nice logging DC voltmeter.

From what I've worked out these are the main features:

MilliOhm Meter:

Measurement source current is fixed at around 100mA at max 3V.

There are 2 ranges: 0.2R and 20R.

In the 0.2R range, the resolution goes down to 1 micro ohm.

At the 20R range it's 10 micro ohm.

Overrange is > 50%, as in, you can measure a 30 Ohm resistor, on the 20 Ohm range.


Voltage Meter:

150,000+ counts, 4 ranges: 0.1V, 1V, 10V and 100V

Again with > 50% over range. As in, 10V range will measure +/- 15V

Inputs are protected by PTCs and ADC/mcu protected by resistors and diodes.

Common for both Ohms and Voltage are 2 speeds, Slow and Fast. Slow is approximately 1 update per second and fast is updating something like 14-15 times per second.

The display has got a nice uniform backlight.

USB for charging and logging. Output is very simple, just the value currently being displayed and at the selected speed range through an USB/UART 115200 baud 8N1 (not isolated).

It's also got battery voltage monitoring with a visible indicator.

Auto off after 10 minutes (can be disabled).

Power consumption is about 15mA in Voltage Meter mode. And ~110mA as Ohm Meter. Add about 35mA for backlight.

Not sure what the accuracy is, but all ranges can be calibrated against a reference. (positive, negative and zero for voltage and zero plus value for ohms)

The standard reference is a MAX6192A, ADC is 24bit in the C8051F350 and the current resistor is a low tempco type.

It appears to have good accuracy.

[kripton2035]

nice review, how much did it cost to you at end ?

[sorin]

schematic
http://www.filedropper.com/ohmmeterf350v21


on taobao is sold as a kit, and cost ¥170  around $28
http://item.taobao.com/item.htm?id=17790993945

[kripton2035]

on taobao is sold as a kit, and cost ¥170  around $28
http://item.taobao.com/item.htm?id=17790993945

no one sell it with an english page and accepting paypal

[iloveelectronics]

schematic
http://www.filedropper.com/ohmmeterf350v21


on taobao is sold as a kit, and cost ¥170  around $28
http://item.taobao.com/item.htm?id=17790993945

In defence of my pricing that Taobao price doesn't include the programmer and the acrylic case, nor shipping (within China)

[sorin]

In defence of my pricing that Taobao price doesn't include the programmer and the acrylic case, nor shipping (within China)

What is your price (if you want to make it public)?

[sorin]

on taobao is sold as a kit, and cost ¥170  around $28
http://item.taobao.com/item.htm?id=17790993945

no one sell it with an english page and accepting paypal

Use a taobao agent which accept paypal
You need to add 2-3$ for shipping inside china, ~5$ for international shipping and +10% commission for the agent, +0.3% paypal fee.
Total around 40$

[jaxbird]

nice review, how much did it cost to you at end ?

Hi Kripton2035,

Thanks, I think I paid $62 including shipping and tracking.

I really don't think there would be much saved from buying on Taobao, the complete kit there is something like 250 Yuan, then plus fees and shipping, it would be close. I've only had good experience with buying from Franky, fast shipping good communication etc.

Whether it's worth $60 something, is up to the individual. I'm pretty happy with it, as I didn't have any meters that would do milliohms with good resolution, and it does work very well.

A nice detail is that the kit comes with full source code, so very easy to customize it to specific needs. I had a quick play and created a modified version, that adds an extra digit of resolution in slow Voltmeter mode.

[iloveelectronics]

In defence of my pricing that Taobao price doesn't include the programmer and the acrylic case, nor shipping (within China)

What is your price (if you want to make it public)?

I think I did make it public earlier in this thread. I'm taking pre-orders at $62 shipped worldwide while finishing the translation (half done there). Jaxbird didn't mind the Chinese-only documentation so I sent him the first unit and he managed well with Google translate

[KD0CAC John]

I may have missed it , but was there some info on ranges of measurement ?
I picked up couple used / questionable HV test probes , in 30,000 volt max on one , and none of my gear will measure the resistors in the probes .

[iloveelectronics]

I may have missed it , but was there some info on ranges of measurement ?
I picked up couple used / questionable HV test probes , in 30,000 volt max on one , and none of my gear will measure the resistors in the probes .

For resistance there are 2 ranges, 0.2 and 20 ohms.

[dannyf]

0.2 and 20 ohms.

I would take a look at the design / implementation of the 0.2ohm range. If a separate precision resistor is used, I would like to see the pcb design for that. If it is done via the pga, I would think the pga's gain error would render the 24-bit adc largely ineffective.

[jaxbird]

0.2 and 20 ohms.

I would take a look at the design / implementation of the 0.2ohm range. If a separate precision resistor is used, I would like to see the pcb design for that. If it is done via the pga, I would think the pga's gain error would render the 24-bit adc largely ineffective.

It does use the full 128x gain on the 0.2R range, but it also limits the counts to just ~300,000 compared to the 20R range using 1x gain and ~3,000,000 count resolution.

Both ranges are stable on the last digit.

[KD0CAC John]

I may have missed it , but was there some info on ranges of measurement ?
I picked up couple used / questionable HV test probes , in 30,000 volt max on one , and none of my gear will measure the resistors in the probes .

For resistance there are 2 ranges, 0.2 and 20 ohms.

I guess that's a good example of short-term-memory loss , as soon as I saw your reply , I thought to myself , ya saw that at least a couple times
Old-Farter

[robrenz]

Does it do offset compensation to eliminate thermal EMF issues?

[jaxbird]

Does it do offset compensation to eliminate thermal EMF issues?

No, it doesn't, the source current is on continuously.

I guess it would need to do a measurement without the source current and then use that as compensation to eliminate thermal EMF.

[robrenz]

Does it do offset compensation to eliminate thermal EMF issues?

No, it doesn't, the source current is on continuously.

I guess it would need to do a measurement without the source current and then use that as compensation to eliminate thermal EMF.

Correct, micro Ohm measurements without some type of offset compensation is meaningless. That would be my first programming change (if I were a programmer)

[dannyf]

I guess it would need to do a measurement without the source current and then use that as compensation to eliminate thermal EMF.

The chip can do the calibration (both offset and / gain). Not sure if the software takes advantage of that.

[jaxbird]

Correct, micro Ohm measurements without some type of offset compensation is meaningless. That would be my first programming change (if I were a programmer)

It would also require some hardware modifications to fully implement thermal EMF error compensation.

I did a quick experiment, simply disconnecting the current source leads to get some idea of the error. In this case, measuring a 3W 0.1R resistor, the error is about 25 micro Ohm.

I guess using this procedure one could manually zero out the error.

[jaxbird]

The chip can do the calibration (both offset and / gain). Not sure if the software takes advantage of that.

The 8051F350 ADC does have a nice set of features:

https://www.silabs.com/Support%20Documents/TechnicalDocs/an217.pdf

I haven't checked but I assume it does an internal gain/offset calibration during initialization, but no continuous re-calibration. 

[robrenz]

IMO you don't want recalibration. You calibrate once to get correct gain and offset of the voltage readings. Then when doing a resistance measurement you take a reading with no current applied and store that value. Now take a reading with the current on. algebraically subtract the "no current" reading from the "with current" reading and display that value. This should repeat at least once per second.

[dannyf]

The 8051F350 ADC does have a nice set of features:

The whole 35x family is amazing.

You calibrate once ...

Read the datasheet and see how your implementation could be simpler than recalibration.

[jaxbird]

IMO you don't want recalibration. You calibrate once to get correct gain and offset of the voltage readings. Then when doing a resistance measurement you take a reading with no current applied and store that value. Now take a reading with the current on. algebraically subtract the "no current" reading from the "with current" reading and display that value. This should repeat at least once per second.

It is more a recalibration of the ADC's internal gain and offset. It will drift slightly with temperature.

But, the part that requires a bit of thought is not how to do the compensation in software, whether subtraction or just updating the offset, it's more how to implement the changes in hardware.

A simplified schematic of the circuit looks something like this:





It would need to switch the source current from the dut to a dummy load for the reference +/- voltages to remain approximately the same.