[SOLVED] ADC voltmeter and ammeter layout query

[d123]

Hi!
Could somebody tell me what mistakes I may have made in the layout (and possibly reasoning/understanding, especially of the OpAmp) for this circuit please? I've omitted external components for both ICs which would clutter the schematic and are not an issue.
The 0.01 Ohm resistor and trimpot are just an idea (maybe a bad one?) which may not be included, just in case when bread-boarding I see the ammeter is too inaccurate and needs tuning.
The current shunt monitor is an INAID286 current to voltage device (100V/V), the ADC is a TC7107 (don't tell KlausST!), the DPDT switch is a break-before-make rotary switch. The purpose of the voltmeter and ammeter is to monitor V and I on a power supply.
I understand that for a 200mV ADC scale, the shunt should also manage 0 - 200mV, so the INA286 will output from 0 to 20V (which the voltmeter divider of 1M/10K will reduce back to 0 - 200mV again).
I have some doubts (besides not being sure I've understood the range I'll get from the current shunt monitor OpAmp I selected for this purpose): Would it be better to have separate devices for each function/Will this set-up introduce further inaccuracy on what will obviously be a not very good homemade power supply, voltmeter, and ammeter? The voltmeter is at worst 15mV out compared to a multimeter, so I'd like to maintain that "accuracy", rather than lose it by cobbling an ammeter onto its input.
Thanks.



- - - Updated - - -

...I put a fuse in the wrong place and omitted another, so this is the corrected schematic.

 

[KlausST]

Hi,

(don't tell KlausST!)

Busted!

The main problem may be the lower side switch connection. Instead of using the switch just connect GND, ADC_IN-, and REF1 continously.

***Ooops, i accidentally pressed the "submit" button

* Although not wrong, i don´t think it is a good idea to amplify a signal by 100 and after that feeding it to a 1/100 voltage divider. You loose precision and may catch noise.
* did you recognize that with no load your voltage reading is OK, but with load it will be wrong up to 200mV (voltage drop of shunt)?

I recommend to build a voltage divider maybe 100k/1k and a capacitor in parallel of about 10uF (caramics, or better MKT) for voltage measurement.

To avoid that large voltage drop with current measurement i recommend to use a much smaller value shunt. Maybe 10mOhms. Making 1V FS Amplifier output. Then use a 4:1 voltage divider to get 200mV range.

Use a SPDT switch to choose center of "voltage reading divider" and center of "current reading divider".

Good luck

Klaus

 

[d123]

Caught me out, "...found in (cheap) multimeters"!
And thanks, that's great to know and very interesting, really helpful. Thank you.

 

[d123]

Hi again.
Shame, the short answer was one I can manage! The other advice sounds maybe hard for me to implement towards the end of what you say about the 4:1 divider, I've only been fiddling with circuits that aren't kits you solder brainlessly since last year so I'm still at the stage of "It works, what a relief, maybe I'll tinker a little to improve it but not too much." (!).

I have only tested the voltmeter on a 7805 output/a 7660 -5V output/the rail voltages from 9V to 12V/an LED, so I dread to think if it's going to lose about 200mV with real loads. A bit late to change anything with regard to divider resistors as I've soldered that part already...However, I will bear in mind your suggestion for 100K/1K for a future occasion, and your other useful advice, thank you.

I haven't tried the ammeter yet - I ordered the shunt resistors at the week-end, so have a tiny IC soldered and have been filling time re-reading the 286 datasheet (know the enemy!) and reading a little about current shunt monitors, etc.
Just to be sure I understood - Is this what you mean about tying ref, GND and vin- together and using a SPDT on the +voltage or current signals?:

I'm slowly trying to wean myself off through hole, it's like a vicious cycle of needing something else to finish something else... And I like MKT and polystyrene, but generally have to use normal PET to save on costs, they may not be the best, but from what I've seen on my (bad) circuits the three types have helped a lot make up for my lack of skills.
Many thanks for your helpful advice, it's much appreciated.

 

[FvM]

Amplifying the shunt voltage by 100 simply doesn't work because INA286 output is limited to 5 V.

 

[KlausST]

Hi,

Yes, didn't think about this.....another point for using a 10mOhms shunt.

Klaus

 

[d123]

Hi!
...But surely by wishing devices can function based more on naïve expectations than actual operating parameters, can't they?

Your observation is much appreciated, that aspect about the possible output swing based on supply voltage went right over my head for concentrating on the shunt resistor and getting it to fit the scale of the voltmeter I've already made.

Not to worry, in my case the main point is learning (oh, and playing with fun little "toys" that can do interesting things), this time I'll just have to have a separate voltmeter and ammeter and use a 7106 I have for the ammeter.

Thank you very much.