Possible data output from UNI-T UT33A+ multimeter?

[takio]

You need a negative measurement to see where the sign is in the data.

Great catch. Was never going to remember that. Will add +/- measurements of the same value. Placing my bets on the “KL” bit.

[magic]

Placing a bet on two's complement

[MikeK]

REL may be in there too.

[takio]

Looks like both win

I'll repeat all the readings here for convenience: new ones are in bold text and denoted with an asterisk *

The meter only supports REL for capacitance, so threw in a combined measurement for that, in hopes of spotting the REL state in the data.

Also repeated some readings and got exactly the same output for the same values, so it's definitely consistent.

10.90 V AC output was indeed incorrect, I must have made an error in noting it down. Measured again and seems correct now.


*************************************

10.90 V AC *
AB CD 01 03 08 00 04 42 00 52 

2.50 V AC
AB CD 01 03 08 00 00 FA 01 06

*************************************

0.005 V DC
AB CD 01 06 04 00 00 05 00 10

2.009 V DC
AB CD 01 06 04 00 07 D9 00 EB

5.00 V DC *
AB CD 01 06 08 00 01 F4 01 04

- 5.00 V DC *
AB CD 01 06 08 FF FE 0C 02 18

6.00 V DC *
AB CD 01 06 08 00 02 58 00 69

- 6.00 V DC *
AB CD 01 06 08 FF FD A8 02 B3

10.00 V DC
AB CD 01 06 08 00 03 E8 00 FA

12.00 V DC
AB CD 01 06 08 00 04 B0 00 C3

24.0 V DC
AB CD 01 06 10 00 00 F0 01 07

*************************************

14.6 mV DC
AB CD 01 05 02 00 00 92 00 9A

35.0 mV DC
AB CD 01 05 02 00 01 5E 00 67

116.0 mV DC
AB CD 01 05 02 00 04 88 00 94

*************************************

1.3 Ohm
AB CD 01 09 01 00 00 0D 00 18

10.3 Ohm
AB CD 01 09 01 00 00 67 00 72

1.003 kOhm
AB CD 01 09 02 00 03 EB 00 FA

100.0 kOhm
AB CD 01 09 08 00 03 E8 00 FD

1.000 MOhm
AB CD 01 09 10 00 03 E8 01 05

*************************************

Continuity: OL
AB CD 01 0D 04 00 7F E8 01 79

Continuity: 0.0
AB CD 01 0D 01 00 00 00 00 0F

Continuity: 20.2
AB CD 01 0D 01 00 00 CA 00 D9

Continuity: 30.0  (no beep)
AB CD 01 0D 01 00 01 2C 00 3C

*************************************

9.79 uF *
AB CD 01 0E 01 00 03 D3 00 E6

100.2 uF *
AB CD 01 0E 01 00 03 EA 00 FD

90.5 uF *
AB CD 01 0E 01 00 03 89 00 9C
[100.2 uF capacitor measured with meter Rel'd at 9.79 uF]

*************************************

[MikeK]

A zero (0.00) reading with and without REL might make it obvious.  HOLD may be in the data too.

[magic]

Looks like both win

So it seems

It may be that the data value is signed 24 bit, but this model doesn't have enough counts to ever output positive numbers longer than 16 bits (~60000), so only 0xff shows up in negative readings.

[takio]

Perhaps using the same protocol/format as higher-spec IC's from the same maker? That could hint at being compatible with existing software — and possibly USB adapters. Wishful thinking.

Meanwhile, some more output:

*************************************

CAPACITANCE MODE

Short across probes:
O.L
AB CD 01 0E 01 00 FF FF 02 0E

HOLD on, short across probes:
O.L
AB CD 01 0E 01 00 FF FF 02 0E

No probes attached:
0.100 nF
AB CD 01 0E 01 00 00 64 00 74

REL on, no probes attached:
0.000
AB CD 01 0E 01 00 00 00 00 10

HOLD on, no probes attached:
0.100 nF
AB CD 01 0E 01 00 00 64 00 74

*************************************

CURRENT MODE: uA / mA / A

0.0 uA DC
AB CD 01 0F 02 00 00 00 00 12

0.0 uA AC
AB CD 01 0F 02 00 00 00 00 12

0.00 mA DC
AB CD 01 07 08 00 00 00 00 10

0.00 mA AC
AB CD 01 07 08 00 00 00 00 10

0.000 A DC
AB CD 01 0B 20 00 00 00 00 2C

0.000 A AC
AB CD 01 0B 20 00 00 00 00 2C

*************************************

Still can't make sense of those last two bits. Thought it might have to do with current measurement somehow, but don't see it. Also strange that the AC and DC current measurements are identical for the same value, no?

[magic]

See if there is a difference between 0mVAC and 0mVDC.
Maybe it just doesn't distinguish between modes selected with the yellow button.

About the final bytes we know that the value is the same for 10.00 VDC and 1.003kΩ
Also, 02 xx only appeared with negative readings and shorted probes so far.

The final byte could be a checksum. Completely repeatable, completely meaningless.

[magic]

It's a checksum. This C program prints full message based on its first 8 bytes.

Code: [Select]

#include <stdio.h>

int main() {
for(;;) {
unsigned ab, cd, ef, gh, ij, kl, mn, op;
scanf("%x %x %x %x %x %x %x %x",
&ab, &cd, &ef, &gh, &ij, &kl, &mn, &op);
unsigned csum = ab + cd + ef + gh + ij + kl + mn + op - 0x178;
printf("%02x %02x %02x %02x %02x %02x %02x %02x %02x %02x\n",
ab, cd, ef, gh, ij, kl, mn, op,
csum / 0x100, csum % 0x100);
}
}
Example run:

Code: [Select]

AB CD 01 0E 01 00 FF FF
ab cd 01 0e 01 00 ff ff 02 0e
AB CD 01 0E 01 00 00 64
ab cd 01 0e 01 00 00 64 00 74
AB CD 01 0E 01 00 00 00
ab cd 01 0e 01 00 00 00 00 10
I tried it on all messages you posted and it got all of them right.

Perhaps using the same protocol/format as higher-spec IC's from the same maker? That could hint at being compatible with existing software — and possibly USB adapters. Wishful thinking.

It's more likely a debugging aid left by firmware developers or used for production testing and calibration.
But it doesn't matter: UT33A+ is now officially the cheapest logging DMM on the market

[Messtechniker]

UT33A+ is now officially the cheapest logging DMM on the market

Precisely what I was hoping for.   

[magic]

BTW, I noticed that 0x178 is the sum of 0xAB + 0xCD. So instead of subtracting 0x178, we could simply skip the AB CD header in checksum calculation, because it looks like the header is a fixed part of the protocol.

To validate a packet:
1. make sure the fist two bytes are AB CD
2. calculate the sum of the next 6 bytes: ef + gh + ij + kl + mn + op
3. compare the sum against the final 2 bytes: qr, st
 

There doesn't seem to be any more hidden information there: just range dial position, decimal dot position and LCD readout.

[takio]

Yep, same deal for AC/DC on mV:

*************************************

0.0 mV DC
AB CD 01 05 02 00 00 00 00 08

0.0 mV AC
AB CD 01 05 02 00 00 00 00 08

*************************************

No reference to AC/DC status. Suppose it's left to the operator to know which mode the meter is set to

[AB CD] could, theoretically, relate to the DMM model, the only global constant. It could tell the computer what data to expect. On this board from the UT33D+, a sister model from the same series, lo and behold: the same TXD/RXD/VCC/GND pads in the same location. This one is a manual range meter. Don't think there would be a reason for it to output differently from the auto ranging model, but the board layout is very different, especially around the IC.

Also: the yellow button does different things on different models. E.g. HOLD is on the blue button for the 33A+ model, and on the yellow button for the 33D+ model. Might help explain why they left that out of the data, and reason to expect equal output from all variants.

There's probably a bunch of other meters using this same IC out there, UNI-T meters are often rebranded by Tenma, Wolfcraft... wonder if it could be identified from a firmware dump. I'm still hopeful there might be a compatible cable+software interface out there. Most if not all meters sold with PC connectivity are way above 2000 counts, of course. But if I can find some specs or reverse-engineering notes on IR connection cables, no reason not to give it a go.

Still very happy with what was found, thanks to all of your help! This is a nice little meter, especially with logging capabilities for about 15 USD/EUR.

/ / /

One thing I absolutely love about this DMM is having one single function per dial position (except for AC/DC, but hey...). "Booting" straight to the mode you want without having to fiddle with options or doing three presses to get to continuity mode, is very old-school in the best way possible. Makes this a great little "knock-about" tool for basic tasks. It even has provisions for DIY extra input protection (for different market compliance requirements I guess).


Image from https://budgetlightforum.com/node/58485