UT171C Teardown (Lots of Pictures) & Various Notes

[Digital Corpus]

Before anyone asks, I paid a solid $240 USD w/ shipping for the meter through Alibaba. Anyhow...

My first meter I purchased was an Extech 450. Not the worst and I didn't need much from it. I decided it was time to upgrade and because I'm a night owl, the OLED display is an actual selling point for me and I wait to see how burn in on the display plays out over the next several years. Two other requirements for me were data logging and thermocouple attachment, to which this accepts a K-type thermocouple and has a 10K count memory that technically can be offloaded onto a PC.

The brother, UT171B, was torn down here. I won't go over the construction details because the two meters are nearly identical so any of my notes will focus on differences between them and general notes/thoughts about it.

First up will be photos, including some macro shots.
(Click on the photos to embiggen and may be up to 18 MP & 1.6-7.2 MiB)

Front side of PCB & Backside of PCB
   

Macro of top portion of PCB by UART LEDs:


Macro of the HY3131 DMM front end, a 50,000 count chip for a 60,000 meter:


Macro of Cortex M3 processor:


Macro of Cortex-M3, STMicro 45PE40 4Mbit SPI Flash (Right), and generic (?) 256 kb I2C EEPROM (Left):


LiPo Charging Circuit sans a glass diode on the input:


Inside of front case & Inside of back case
   

HRC Fuses with isolation:


Larger HRC fuse access:


Smaller HRC fuse access:


Dembossed Details on back panel:


Under the RF can:


Silkscreen quality:



I didn't think to take a shot of the membrane buttons, but there is a positive lip on the side of the print so if you accidentally splash a little liquid onto the front, don't worry about it making its way to the PCB. I did not remove the OLED screen because it is mounted on 2 squared strips of adhesive backed foam; very precisely mounted so I didn't want to disturb it. There were a few ICs underneath, to which I assume is part of the RTC drive, but taking a macro shot at the right angle and right light isn't something I wanted to spend 30-90 minutes on with a small likelihood of identifying the chip.

Display
Now, the screen since it's a fairly unique feature. It's a yellow-on-black, negative display with a refresh rate 120-125 Hz regardless of brightness. It'd take more time than I'd care to figure out the precise duty cycle but from analysing a few photos I've taken, it looks to be ~10%. You'd think that this creates a relatively smooth display, though due to the short DC %, flicker is visible. Then again, everyone's flicker threshold is different and I am accustomed to spotting PWM in LED light sources. Under standard conditions, I notice the refresh rate, but don't find it distracting.

New display data is written every 60 ms +/- 5 ms jitter regardless of if the data has changed. Visually, the data updates 4-5 times per second depending on the mode you're in e.g. all DC volts ~= 4-5 Hz, True RMS AC+DC ~= 0.2 Hz, thermocouple ~= 0.667 Hz. As for the bar graph display, It's response time seems to equate to the 60 ms timing, aka ~16.67 Hz. There are 3 default brightness levels and you get to them by a 2 second press of the Esc button. They happen to be 10%, 40%, and 100%. You can go into the menu and custom set the brightness if you *really* want to manually adjust the brightness, you can do so from 10% to 100% in 2% intervals. However, if you use the Esc button to change your brightness level, it defaults back to 10-40-100 so the manual setting is kinda useless, but I bet it's in there for those with a touch of OCD... Now, because this isn't and LCD screen, you can read it as long and you don't have a reflection interfering with the characters, which translates to ~160? vertical and horizontal. When you turn the display off, the updating freezes and the screen fade off in about 1s. If they really want to make it look fancy, it should turn on the same way instead of insta-on.

Using some *rough* relative measurements with a lux meter, the brightness I register @ "10%" is about 16 lux. At 40% I get 24 lux, and at 100% brightness I get 38 lux.

Probes
As for the probes, well let's say this. I immediately tried continuity testing them against each other, but even rubbing them together like a fire starter prevented the UT171C from even triggering a beep; they're in tomorrow's trash. I picked up a 2 sets of ProbeMaster's probes that Dave reviewed, 24" and 72" leads since most of the time I only need the short stuff. The delay is still a touch long, but the bar graph updates before, about 1/3rd the time, the beeper goes off.

Beeper
Speaking of the beeper, it sounded a bit, odd. When I opened up the case to investigate it worked just fine and is also quite loud. What appears to be happening is that the sound is being reflected straight back into the cavity of the beeper and causing harmonic interference. There are a few ways to fix this, though with the responsiveness of the meter, I turned the key beep off so I only hear it with the continuity test.

Buttons
There is no hardware debounce on the buttons and I was hoping there was because the delay before a registered keypress is a good 20-30% longer than I'd wish. The RECALL button is the only one that doesn't have any secondary or tertiary functions. To canceL the relative mod, you have to so a 2 second press of the relative button. To enter data logging mode, a 2 second press of the STORE is needed. A single press of the blue button allows bumping the almost every mode into its secondary mode and for some, such as mV, mA, & uA a 2 second press is required. As for the visual layout of the meter, everything is generally okay except for the silkscreen. If you take a look at that photo, the lettering has some holes in it. Because of hot they're printed on a curved surface, they also distort when looking from an off-angle and that is more common than not. I hope the ink is high quality or those labels will deteriorate much faster.

It does take a little while to get use to the layout and UI. The setup menu is sparse but everything else is accessible outside of it which is nice, imho. I've had the meter for about 2 weeks and so far there is only one quirk I'm not fond of and I'll mention that later.

Signal Generator
You'll also note that instead of NCV there is a built in signal generator, to which I missed this when looking over it its specs. It's a basic generator and it's not even that great, let's be honest. It only outputs a 0.8 V p-p square wave. On screen it is good from 0.1% to 100% duty cycle in 0.1% resolution. And the frequency range is from 0.5 Hz to 4800 Hz in .1 Hz resolution. In actuality, it's really only good from 0.2% to 98% when at higher frequencies. Also, the 0.1 Hz resolution doesn't mean anything after ~ 2400 Hz. There is aliasing apparent from a not enough bits being dedicated to this output, but for basic stuff it's okay so it has its place.

I do regret to say that I do not have any reference sources for voltage, resistance, current but I'd expect it to match the UT171B.

Battery & Power
As for the LiPo battery, as long as it's decent quality, you'll be fine. It is two 103450 LiPo batteries in series and I cannot seem to find a replacement anywhere online except for random possibilities through similar Chinese trade sites. The charger supplied sucks about 6 W tops and the battery is supplied a constant current charge of 460 mA up to a set point where it then tapers the current down to nothing until the battery is charged to 8.4V. Sorry, I missed the point where it crosses. Overall, the charging circuit has an efficiency anywhere from 85-90%, fwiw. Power is applied through the mA and Com jacks and the other two posts in the adapter are not connected to anything. I've measured 65 mA current draw with the screen on 100% brightness with the batter at ~7.8 V and 55 mA when on 10%. You can set the screen to turn off within a set time of inactivity by keypresses or mode switching to which is 5 min to 30 min or off. Auto power down is customizable is the same manner. For some reason there is power consumption when the meter is turned off, I've measured 8-11 mA. Currently I've only measured this with the UT210E though I hope to be able to grab some more precise, in-line readings to establish battery life.

With the addition of a Rigol DP821, I have had a chance to measure power consumption a bit more closely. I explicitly tested current across the li-po range. I did 5 sets of explicit measurements while rounding to the nearest milliamp. These were with the screen off from auto low power, @ 10% brightness across the voltage range, @ 40% brightness across the voltage range, @ 100% brightness across the voltage range, and every 10% @ 8.4V input. The voltage regulator is of the linear variety and is "W1" and is the same as "W4" on the back side of the PCB; package markings are "PAW". This is likely the TLV70436DBV. First a few specific measurements:

• Peak consumption w/o the beeper/buzzer: 80 mA @ 8.4 V
• Minimum consumption w/ screen off: 44 mA @ 6.9 V
• Battery voltage is checked every 15 seconds after initial reading
• Full battery voltage range: 8.1 V to 8.4 V
• "66%" battery voltage: 8.0 V to 7.8 V
• "33%" battery voltage: 7.7 V to 7.4 V
• "Empty" battery range: 7.3 V to 6.9 V
• "Boot"-able voltage range: 4.2 V to 6.8 V will let the meter power on but upon initial voltage check it'll beep 3 times and auto power down for protection of the battery.
• The beeper/buzzer requires ~25 mA @ a nominal operating current of 63 mA.
• The average power consumption across the entire brightness range and voltage range, which is interpolated, is 69 mA.

Voltage	Batt. Lvl.	Off (mA)	10% (mA)	20% (mA)	30% (mA)	40% (mA)	50% (mA)	60% (mA)	70% (mA)	80% (mA)	90% (mA)	100% (mA)	Stand By (mA)
8.4	100%	49	64	66	68	70	72	73.5	75	77	79	80	38.1
8.3	100%	49	64	66	68	70	71.5	73	74.5	76	77.5	79	37.7
8.2	100%	48	64	66	68	69	70.5	72	73.5	75	76.5	79	37.4
8.1	100%	48	63	65	67	69	70.5	72	73.5	75	76.5	79	37
8	66%	48	63	65	67	68	69.5	71	72.5	74	75.5	78	36.6
7.9	66%	48	63	65	67	68	69.5	71	72.5	74	75.5	78	36.2
7.8	66%	47	62	64	66	68	69.5	71	72.5	74	75.5	78	35.9
7.7	33%	47	62	64	66	67	68.5	70	71.5	73	74.5	77	35.5
7.6	33%	46	61	63	65	67	68.5	70	71.5	73	74.5	77	35.1
7.5	33%	46	61	63	65	66	67.5	69	70.5	72	73.5	76	34.7
7.4	33%	46	61	63	65	66	67.5	69	70.5	72	73.5	76	34.2
7.3	0%	45	60	62	64	66	67.5	69	70.5	72	73.5	75	33.8
7.2	0%	45	60	62	64	65	66.5	68	69.5	71	72.5	75	33.4
7.1	0%	44	59	61	63	65	66.5	68	69.5	71	72.5	75	32.9
7	0%	44	59	61	63	65	66.5	68	69.5	71	72.5	74	32.5
6.9	0%	44	59	61	63	64	65.5	67	68.5	70	71.5	73	32
6.8	APO	44	58	60	62	64	65.5	67	68.5	70	71.5	73	31.6

Will the meter win any rewards for power efficiency? Heck no. But it's cutoff voltage is 3.7 V per series Li-po cell so you'll have this battery last you quite some time. Granted, the full rating of the battery is 1800 mA but that's down to 2.7 V. At 3.7 V, that is ~85% capacity and will serve to maximize the battery's life.

With a 13.32 Wh li-po battery, this gives us an effective nominal battery life of 22 continuous hours. At peak, average consumption this drops down to 20 hrs, but if used as a data logger with the screen off, about 33 hrs. Under normal indoor use, you don't need anything more than 10% screen brightness which will net just barely 25 hrs.

Also, the meter consumes quite a bit of power after it Auto Powers Off with an average current draw of 35 mA. This will kill a fresh battery in just under 44 hrs. The low voltage cutoff is not functional in this state and the meter drain the better flat. Make sure you turn off the meter!

Component Identification & Specs
As for oddball components, I'll start with under the can.

• ES636 is made by Cyrustek and is an "RMS-to-DC" converter.
  
• The small SC70-5 marked JASR chip next to is is TI's TS5A4594, an analog switch. If someone can reference if this is an actual switch, that'd be neat.
• Next up are the previously 'unmarked' ICs in the UT171B, the DF105's, to which there are 2 in both meters. They are 600V, 1A bridge rectifiers. It's hard for me to tell and I didn't think to check the last time I had it open, but I really hope these two are in . Since it passed inspection, I'm a little less worried, but still wondering if I were to need to measure 1000 VAC if I should swap these guys out with higher rated equivalents.
• The HY3131 outputs 5 sps @ 120 dB so with the display updating 4 times instead of 5, it must be an aliasing issue with the 16.67 Hz display write frequency. As previously mentioned, its a DMM analog front end rated for 50,000 count. I'm not familiar with test equipment to know how they get 60,000 out of it. According to the datasheet, when calibrated to 30,000 counts it has an accuracy of +/- 0.01% +2 counts, which is better than what the meter is rated for.
• Just above the can on the far left edge, there is a 4 pin footprint for an I2C hookup. I didn't probe around to see what they are in turn hooked up to, but it'll either be the I2C EEPROM or the ARM Cortex I assume.

Data Logging
Here is the first gotcha, and I assume it's not needed much so this is why it's setup as such. In order to transfer data to the PC, you need to turn on the USB function via the setup menu. However, if the meter power cycles, auto or manual, USB gets turned off. I assume it has to do with power consumption, to which I've not measured it when the interface is on to be truthful.

From the meter's UI, I can select and interval of 1 to 240 seconds between samples. Based on available free memory, it will them tell me the maximum duration I can log in minutes, but I can make it shorter. However, I can select 1 min of logging with a 240 second interval . If you hit Esc when on the Duration setup it'll take you back to the Interval setup, which is a nice refinement. Once you hit Ok on the duration, logging will immediately commence. It's nice too that you get a countdown timer so you know when logging will end, though it'll round to the nearest minute. To provide isolation the USB adapter is a simple USB-to-UART via IR LED board powered by the Silicon Labs CP2110. I have not clocked it's UART speed but it is supposed to top out at ~1 MHz. When reading a long batch of thermocouple samples, +9000, I calculated that it can pull about 6.4 sps onto the computer over the USB cable.

The other option for logging is simply just plugging the UT171C into the computer and connecting to it. Immediately you get 4-5 sps output just being pulled right into the program without user prompting when you click connect. Now, with their app, which is only available on the mini CD provided, you can do all of this and monitor the meter via the PC. You can even pull down the data samples previously stored onto the meter. Here is the next gotcha.

You'll naturally want to click on the "Pause Receive Data" button before moving over to pull the previous samples off of the meter. However, when you do this and download the samples everything will look fine until you go to save those samples. When you click the "save" button, it wipes the pull from the meter on the PC, reverts back to the last live reading by the meter, and saves only that one reading.

Don't worry, the samples remain on the meter until you explicitly delete them, though if you have say, 3000 samples, that will be 7-8 minutes of your life you need to relive. A full 10K samples takes just under 30 min to download. If you log data here and there, I advise you download and delete if this time is too long, however, this will wear out the flash memory faster. Until someone else reverse engineers the protocol and makes a custom program for this function, this is no way to not begin at the very beginning of the sample memory when you download onto the PC.

Anyhow, in order to reliably download the samples and save them on the PC, this is what I do:

• Switch to a mode that pegs the meter in overload, OL is displayed
• The simplest one is mV and long press the Ok blue button to switch to the thermocouple
• Make sure you click "Start Receive Data" and then "Clear All Record" to ensure the buffer on the PC is clear
• Click on "Read ALL Data"
• Make your favorite pastime drink or snack
• Once the automated process finishes, click the "Read Data in Index" button for any sample, it really doesn't matter
• Click "Save" when you're done

If you don't do step #6, as soon as you click "Save", you'll see that the buffer is clear automagically


Calibration
Last bit is how to access the calibration menu, which is done like on other UT DMMs

• Turn the unit off
• Hold down RANGE + STORE
• Turn the unit on
• You're greeted with a 5 digit passcode entry screen

Anyhow, that is all for now. I hope you guys find this useful.

[mos6502]

As for the probes, well let's say this. I immediately tried continuity testing them against each other, but even rubbing them together like a fire starter prevented the UT171C from even triggering a beep; they're in tomorrow's trash.

Don't be so quick to dismiss them. My UT139C leads were the same. I just cleaned them with a brass wire brush. By the way, chemicals/solvents didn't help at all. Only the wire brush did the trick.



Now they're just as good as any of my Fluke leads. The wires aren't that much worse either: the Uni-T leads have a total resistance of 37 milliohms, vs. 30 milliohms for the Fluke TL75.

In fact, the leads were probably the best thing about that meter ...

[retiredcaps]

Thanks for the notes, teardown and high resolution pictures! 

Is it me or does the picture at

http://mobrienphoto.com/Misc-Pics/Electrical/UT171C/_OB_9270-hi.jpg

show a lot of dirt, debris, stray solder bits on the pcb (especially near R66 - bottom right hand corner of pic)?

[mos6502]

Thanks for the notes, teardown and high resolution pictures! 

Is it me or does the picture at

http://mobrienphoto.com/Misc-Pics/Electrical/UT171C/_OB_9270-hi.jpg

show a lot of dirt, debris, stray solder bits on the pcb (especially near R66 - bottom right hand corner of pic)?

Notice the corner of the electrolytic cap is also melted. Par for the course for Uni-Turd.

[Digital Corpus]

Thanks for the notes, teardown and high resolution pictures! 

Is it me or does the picture at

http://mobrienphoto.com/Misc-Pics/Electrical/UT171C/_OB_9270-hi.jpg

show a lot of dirt, debris, stray solder bits on the pcb (especially near R66 - bottom right hand corner of pic)?

The fun part about macro photography is that things your normally don't see because plain as day. It's dust, nothing more. There are no solder balls from solder paste or th such on there. I did move outside for the photos so the dust was most likely picked up there. Welcome to why clean rooms exist.

Looks like their QA on the production line changed out 2 components before shipping it to me too. Look for the two circled items.

[Digital Corpus]

As for the probes, well let's say this. I immediately tried continuity testing them against each other, but even rubbing them together like a fire starter prevented the UT171C from even triggering a beep; they're in tomorrow's trash.

Don't be so quick to dismiss them. My UT139C leads were the same. I just cleaned them with a brass wire brush. By the way, chemicals/solvents didn't help at all. Only the wire brush did the trick.



Now they're just as good as any of my Fluke leads. The wires aren't that much worse either: the Uni-T leads have a total resistance of 37 milliohms, vs. 30 milliohms for the Fluke TL75.

In fact, the leads were probably the best thing about that meter ...

Thanks for the note, but I'm not fond of having to remove parts of the plating or finish in order to have the test leads function better. Maybe it's just the automotive or photography sides of me, but unless it's a peel off protector, they should function 100% out of the box.

[techfan]

Hello, in which store alibaba buy the multimeter ?? Thanks for the help.

[3nigm4]

Looking around I see that the HY3131 is also inside Agilent U1282A, that is a 60000 count too 

[3nigm4]

Have someone tested UT171B/C frequency range?

[joeqsmith]

Spaghetti layout looks worse than the UT181A.   

[Digital Corpus]

Still enjoying this meter. Just updated the initial post with power consumption stats. It tracks with my DP821 to the millivolt and within 500 uA in the mA range which is plenty enough resolution for me currently.

[coromonadalix]

Is it possible to re-up the images,  seems unavailable

[eddiea6987]

any idea what OLED screen that is? I have an Agilent U1253A OLED meter which the screen is hardly visible in the dark 
but getting a replacement is impossible to hte point where I am trying to get one shipped from Germany it uses the  SSD 1303 controller which is discontinued for the most part.
So i was wondering what screen this uses because it could be a possible alternative when I throw this meter away but I would want one with readily available replacement screens

[Digital Corpus]

Nope, but the following is printed on the assembly:
"100106702f B 3" on the back of the glass

"1930114308
102182602R" on the FPC.

The driver is the SDS1305T7 and its a 2.7", 128x64 screen.

The FPC assembly is 2 part that is glued/taped to the custom, 22-pin cable and originates as a 31-pin cable with a ~0.65 mm pitch. The 22-pin piece is a ~0.3 mm pitch FPC.