EEVblog Electronics Community Forum

Products => Test Equipment => Topic started by: Shark on October 13, 2013, 01:11:09 pm

Title: Improving resolution of Uni-T UT61C temperature measurement
Post by: Shark on October 13, 2013, 01:11:09 pm
Hi everyone!

I had an idea of modifying my Uni-T UT61C to have one more digit of resolution when measuring temperature (I understand that this does not imply accuracy; I would prefer to have one more digit for observing rising/falling temperature more easily).

Out of the factory, this model is capable of displaying temperature at one degree resolution.

The multimeter is based around Fortune FS9922-DMM4 chip which has 3 modes of measuring temperature:
Datasheet can be found at: http://www.ic-fortune.com/upload/Download/FS9922-DMM4-DS-11_EN.pdf (http://www.ic-fortune.com/upload/Download/FS9922-DMM4-DS-11_EN.pdf)
Also see the screenshot I attached, taken from the datasheet.

Obviously UT61C is set to use the first mode (11111) because displayed temperature has one degree resolution and the Range switch has no function in temperature measuring mode.

I opened the multimeter and tried to make sense of the range switch. It's the first time I'm dealing with range switches and I haven't figured out how exactly the mode is set. I'm attaching pictures of how far I've come by now and I'm asking for advice on how to proceed.
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: TMM on October 13, 2013, 03:56:45 pm
In the datasheet it says that the mode that has 1/10th degree resolution is "-273.0ºC to 400.0ºC" and the mode that is 1 degree resolution is "-273ºC to 4000ºC". Then it goes on to say that the scale for the first mode is "10.0ºC/mV" and the scale for the second mode is "10ºC/mV" as if it were different? Possibly a typo or lost in translation?

It would seem logical to me that they don't change anything with the A/D conversion but instead just move the decimal place on the LCD and require the analog circuitry to be altered accordingly. Also, -273.0 to 400.0 at 0.1 resolution is over 6000 counts, so it beats me how that works...
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: Shark on October 13, 2013, 06:29:44 pm
Yeah, I noticed the same thing in the datasheet, it's funny how 10mV/°C is different from 10.0mV/°C hehehe.

To my understanding the input for all three modes, the input is the same and specified as 10mV/°C. I guess it's all about reconfiguring the range switch to select mode 10111 or 01110.
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: mariush on October 13, 2013, 07:55:23 pm
I guess you could try putting some scotch tape over the pads so that contact won't be made and see if that changes anything.
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: Shark on October 13, 2013, 08:15:03 pm
Quote from: mariush on October 13, 2013, 07:55:23 pm
I guess you could try putting some scotch tape over the pads so that contact won't be made and see if that changes anything.

This was also my first idea, I tried it last week, but it changed nothing. All it did was that it made meter not function properly in some of the cases (I mean scotch-taping different pads).

Maybe my assumption that the rotary range switch sets the mode directly is wrong. Anyone with experience on how this is usually wired in a multimeter?
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: TMM on October 14, 2013, 07:07:15 am
Quote from: Shark on October 13, 2013, 06:29:44 pm
To my understanding the input for all three modes, the input is the same and specified as 10mV/°C.
That's where i think there may be a typo. Perhaps with the 0.1ºC resolution mode it should be 1ºC/mV rather than 10ºC/mV. It would be much easier just move the decimal point and increase the gain of the amplifier circuit from 2.5 to 25 so at 400.0ºC with 0.1ºC res the ADC sees the same voltage as it does at 4000ºC with 1ºC resolution. It could explain why it can only count to 400C instead of 4000C in that mode.
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: Shark on October 14, 2013, 02:35:32 pm
Yeah, it could be a typo. On the other hand, I saw other multimeters based on FS9922-DMM4 and they have two ranges (with 1/10th degree resolution and 1 degree resolution). I think that this switching is done by the chip, without changing external (op-amp) gain.

I guess the only way I'll know for sure is if I reverse engineer the mode switch and change it into one of the two modes that enable 1/10th degree resolution.
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: Lightages on October 14, 2013, 07:03:22 pm
You need to invert 3 pins going to the chip. You need to change MEA5 or pin 93, MEA4 or pin 88, and MEA1 or pin 91. Trace which pad on the circuit board follows back to these pins according to where the selector contacts are when in temperature mode. The other pad that these contacts connect to will be the logic voltage you need to invert to get the different temperature range enabled. Once you have found those pads, then it MAY be possible to find their connections to logic 0 volts and logic 1 volts and cut a trace and reconnect them to the other opposite logic state. If these pads are connected inside a layer then it may not be possible to separate them from their logic voltage trace without disturbing other some of the other pad's state too.

If you want to get real hacky, you can lift the MEA pins on the chip and hack in your own bypass switching for the measurement functions.

Too me this seems like a lot of work when you can get a temperature measurement meter for around $16 shipped with dual probes not to mention you will make a mess of the UT61C.
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: Shark on October 14, 2013, 08:28:34 pm
@Lightages: thanks a lot for your advice. You make a good point about making a mess of the UT61C ... I'll try to be careful hehe.

I did some more poking around and found that connecting MEA4 to INT should be enough according to the FS9922-DMM4 datasheet. The thing is that '1' stands for 'Not connected' and '0' stands for 'Connect to INT'. I won't need to cut any traces; in current temperature measuring mode all MEA0 to MEA5 are not connected.
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: Lightages on October 14, 2013, 10:52:51 pm
It clearly states in the data sheet that MEA5,MEA4, and MEA1 are inverted to change the temperature mode. Where do you see otherwise?
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: Shark on October 15, 2013, 07:10:03 am
On the page 11 of the datasheet, it's specified:

Quote
(MEA1~MEA5: no connection is "1"; connect to INT is "0")

The current temperature mode has MEA5-MEA1 = 11111, so no connections.
The temperature mode I'd like to use has MEA5-MEA1 = 10111, so MEA4 connected to INT, others not connected.

Is my interpretation correct or am I missing something?
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: Lightages on October 15, 2013, 04:32:19 pm
Yes it appears you are right. It is not clear from the function matrix what which temperature function is what. I would give your idea a try.
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: Shark on October 15, 2013, 08:03:26 pm
Okay, so this evening I took the time, wired everything up and tested connecting INT to MEA4. I was successful in switching the multimeter into 1/10th degree mode (10111), but as TMM pointed out initially, it is required for the analog circuit to provide a 10-times bigger input voltage. An example, instead of 22°C I got 2.2°C.

Now I'll go scratching my head for some time hehehe.
Title: Re: Improving resolution of Uni-T UT61C temperature measurement
Post by: Lightages on October 15, 2013, 08:51:12 pm
Have a look at pages 15 and 30 for hints.