Voltcraft VC940 (alias Unitrend UT71E) problem - false readings

[csadam]

Hi Guys!

I've bought a defective VC940. The symptoms are the following:

• in ohm range,  everything is OK
• in DC Volt range, wrong values displayed, but linearity is OK
• in DC mA range, similar wrong values as in DC volt

When I say "wrong" values, I mean the following:

Input value	Displayed value	difference to the previous step
0 V	0.0000V	-
1 V	0.1593V	0.1593
2 V	0.3194V	0.1601
3 V	0.4799V	0.1605
0 mA	0 mA	-
1 mA	0.142mA	0.142
2 mA	0.303mA	0.161
3 mA	0.464mA	0.161
10 mA	1.584mA	-

It seems that the value of one unit (1 V) is taken as 0.16V by the multimeter. So if I divide the displayed values by 0.16 then I get the right value. I hope it's understandable

I have the schematics and the calibration steps posted in the other topics but none of them helps. The displayed value does not change in Volt mode by turning the VR2.
There are no visible damages on the PCB or on the components. The fuses are OK.
Could you please help me? What could be the problem?

Thanx,
Adam

[Lightages]

If making adjustments to the trimmers do nothing, then you have:

1. one or more defective components
2. one or more bad solder joints
3. a defective circuit board
4. burned traces
5 maybe all of the above

If this is new, send it back. If you purchased used and as defective then maybe you can start checking component values and see if you can trace down the bad part. If you don't know how to do this without help, I can offer some in the way of guiding you to the obvious places to look.

[csadam]

Thank you, I've started checking the components, and all of them were OK.
Then I searched the components which responsible for the reference voltage. I've found that the reference voltage is on GND.
And finally I've fixed it!
While I was measuring it for a time, at once the ref voltage magically appeared and the displayed values are correct since then.
I don't really understand it,  maybe I have magic repair powers...  Or bad solder joints under the VR2 or the R31.

I attached a picture. The critical components for the reference voltage are the

• C22: 1uF
• R31: 4.7k (upper dot)
• R32: 25K  (bottom dot)
• VR2: 200 (blue pot)

The voltage between the GND and the upper dot is -196mV, and at the bottom dot is -204.5mV. The VR2 makes the -200mV reference for the contol IC from these values.

[Lightages]

I am glad you got it working, but don't be too happy. As any owner of the UT71X series will tel you, they aren't reliable.

[csadam]

New problem appeared, or old problem but I just noticed it now

I measure 1.000 volt. It displays 1.000 volt. OK.
But when I activate the PEAK mode the displayed value changes to 1.400 Volt. Then I deactivate the PEAK and it shows 1 volt again.

So someting is wrong with the peak calculation. I dug deep into the internals of the multimeter, and I've sniffed the communication between the measuring IC (Cyrustek ES51966 F) and the controller IC (MSP430F149). Surprisingly the PEAK function does not use the measuring IC-s peak capabilities, because no such command goes into it when the peak mode is enabled. Moreover the measured counts, which are sent by the measuring ic to the controller, are not changed at all during the peak mode.

So I've narrowed the problem to the control IC. I can think two possibilities:
1. there is problem with the firmware
2. the calibration data in the EEPROM is damaged

I'm not sure how can I step towards from this. I would be grateful if someone could send the data in the eprom of his VC940 or UT71E. The data is read from the eeprom via i2c bus at the boot, so it is relative easy to get the data with a logic analyzer.

I can help with the details of this if someone have enough courage to do it

[Lightages]

While I admire your willingness to get your hands dirty with this, I hope you aren't doing it for anything more than to learn. Given all the information I have seen online about the UT71X series of meters and from my own personal experience, any work put into keeping one of these meters working properly is false economy. You can never trust it on the field so what is it really worth?

[csadam]

I can't answer why. I'm a hobbyist, nothing serious. I have a cheap logic analyzer for years and have not used it for anything until now. It is also inspiring to see how a multimeter works in inside. And... because I have no money for a serious multimeter which can measure power

[saturation]

Its interesting to consider that at least for DMMs,  being able to have faith in it is one of the most difficult items to quantify.  It comes from developing a reputation in the field particularly among folks who use it in mission critical applications.  Otherwise,  in new models seeing how its built [ its a reason Dave's tear downs are invaluable], what safety and environmental standards its made to conform too help create that trust between the user and the device.

See this current thread on how Flukes are tested:

https://www.eevblog.com/forum/general-chat/fluke-in-house-testing-video/msg99160/#msg99160



 

[T4P]

Its interesting to consider that at least for DMMs,  being able to have faith in it is one of the most difficult items to quantify.  It comes from developing a reputation in the field particularly among folks who use it in mission critical applications.  Otherwise,  in new models seeing how its built [ its a reason Dave's tear downs are invaluable], what safety and environmental standards its made to conform too help create that trust between the user and the device.

See this current thread on how Flukes are tested:

https://www.eevblog.com/forum/general-chat/fluke-in-house-testing-video/msg99160/#msg99160

That's why i don't trust my 20$ cheapies ... at all 
I will do a review and it's gonna probably be a 15min rant because i really don't like the meter at all .

[csadam]

Interesting video.
But I'm not sure how this Fluke promotion relates to my problem 

[Ricardo]

New problem appeared, or old problem but I just noticed it now

I measure 1.000 volt. It displays 1.000 volt. OK.
But when I activate the PEAK mode the displayed value changes to 1.400 Volt. Then I deactivate the PEAK and it shows 1 volt again.

I own a UT71E Multimeter and it has the same problem as you describe with the peak mode. This mode is useless, I think it is a software bug. It also has other strange problems, perhaps you might check yours:
1: When doing AC-Current measurement, if a small value is measured in a fixed measurement range,  it is not displayed if value is < 200 digits, displayed value is 0. E.g. if 10A AC range is used, the Multimeter can not display values <200mA in this range, it jumps to 0, while >200mA is displayed correctly.
2. In tempereature measurement range, there is a discontinuity fom 0°C to -2°C, no changing values are displayed between, it jumps from 0.0 to -2.0 or vice versa. The problem is also shown on fahrenheit-dispay, jump points are 32.0 to 28.4.

I must say as other users do, the UT71 series seems to be very unreliable, I would not buy it again.

[Bored@Work]

1: When doing AC-Current measurement, if a small value is measured in a fixed measurement range,  it is not displayed if value is < 200 digits, displayed value is 0. E.g. if 10A AC range is used, the Multimeter can not display values <200mA in this range, it jumps to 0, while >200mA is displayed correctly.
2. In tempereature measurement range, there is a discontinuity fom 0°C to -2°C, no changing values are displayed between, it jumps from 0.0 to -2.0 or vice versa. The problem is also shown on fahrenheit-dispay, jump points are 32.0 to 28.4.

Sounds like they try to fake a stable reading of Zero. I think the UT71E uses an off-the-shelf multimeter IC. You could try to find the datasheet and check if this is a known feature of the chip, and if it is maybe a configurable option.

[saturation]

Sorry to jump the gun, it represents Fluke's attention to detail, at least with DMM. 

Keeping an in house safety testing lab is expensive unless they use such labs frequently, which implies they do more frequent quality checks on their production, which suggests high quality control.  Most folks oursource such testing to dedicated labs.

http://www.dtbtest.com/

If Fluke can spend this much on safety testing, its very likely less costly quality testing of boards, components, and cases, are done with much attention too. 

Interesting video.
But I'm not sure how this Fluke promotion relates to my problem 

[csadam]

New problem appeared, or old problem but I just noticed it now

I measure 1.000 volt. It displays 1.000 volt. OK.
But when I activate the PEAK mode the displayed value changes to 1.400 Volt. Then I deactivate the PEAK and it shows 1 volt again.

2. In tempereature measurement range, there is a discontinuity fom 0°C to -2°C, no changing values are displayed between, it jumps from 0.0 to -2.0 or vice versa. The problem is also shown on fahrenheit-dispay, jump points are 32.0 to 28.4.

Hi Ricardo

I tested your 2. problem and it is the same at me. So it is very likely a software issue in the controller.

BoredatWork,
I've done that. The measuring IC takes a command to switch to a mode, then it starts to send the measured counts. In temperature measurement mode it is switched to 400mV DC range. There is no compensation at measuring level. Every correction is done by the controller.
Seems you are right with the fake zero, but there is also a resolution issue.

I've simulated -2°C to 0°C and sniffed the communication. I've found that the the display switches to 0 when the measured counts are around 0 and 8. But there is many missing values there, the measured counts vary between -4, 0 and 6 when the display switches to 0°C. Then the measurements linearly increase until 16  and just then switches the display to 0.2°C. So it shows constant 0°C. from -4 to 16 measured values.


Based on the above I suspect buggy firmware. So my next step is to get an MSP JTAG and read the software. I hope it is not protected

Thanks for the replies