Aim-TTi 1908 5 1/2 Digit Benchtop Multimeter (preview)

[voltsandjolts]

And I've never seen the term "resolution" applied to a multimeter.

Maybe you have never looked for it

[kish]

I have never in all my life seen this definition for precision.

And I've never seen the term "resolution" applied to a multimeter.

Precision: Number of digits (or "counts" in the multimeter world - it might not have a range that's a perfectly round decimal number)

Accuracy: How close those digits are to reality, often expressed as a percentage of reading and number of counts.

eg. My Brymen in CD mode has 500,000 counts and 0.02%, 2 counts accuracy.

Note that 0.02% of 500,000 is 100 counts so the meter isn't claiming that all those digits are accurate.

(although in practice they're usually very close to those 2 counts...that's digital calibration for you!)

So essentially, your Brymen is 5,000 count DMM 
(since the least significant 100 counts cannot be trusted, in 500k count mode)

[2N3055]

I have never in all my life seen this definition for precision.

And I've never seen the term "resolution" applied to a multimeter.

Precision: Number of digits (or "counts" in the multimeter world - it might not have a range that's a perfectly round decimal number)

Accuracy: How close those digits are to reality, often expressed as a percentage of reading and number of counts.

eg. My Brymen in CD mode has 500,000 counts and 0.02%, 2 counts accuracy.

Note that 0.02% of 500,000 is 100 counts so the meter isn't claiming that all those digits are accurate.

(although in practice they're usually very close to those 2 counts...that's digital calibration for you!)

So essentially, your Brymen is 5,000 count DMM 
(since the least significant 100 counts cannot be trusted, in 500k count mode)

And by your logic very expensive Keysight 34460a is 1200 count DMM because in 100 MOHm range it has 0.8% accuracy spec.
Or the legendary HP3458 is really 5.5 digit meter because best version has only 4 ppm/year accuracy spec, so you can't trust anything that is less than 10 ppm of reading on it, right? No, not really. Unfortunately, it is bit more complicated than that.

There is a reason why manufacturers give you specifications table. On every single meter, on every single measurement and every single range you have some (different) accuracy specification. It is user that has to know them all and know how and when to apply them.

[J-R]

Same for the many brymen / clones  at 50,000 count and   oh    the 500,000 count mode     loll 

Yes, that is perhaps a bit gimmicky...but also can be useful for certain purposes.

As received my BM869s was frequently 100+ counts out in 500,000 count mode.  After a quick calibration it is now never more than 15 and typically only 1-5.

At what temperature? Temperature starts to play a big role with that many digits.  Raise it a degree and you might have 100 digits.

No non-temperature-controlled meter is ever going to be 5 1/2 digits accurate but the digits are useful for seeing very small changes over time.

No, it is not that bad once calibrated, which is trivial to do for for the 5V range since there are so many easy to obtain 5V references. 

For the 15 count and 1-5 count stats I mentioned, that is for a relatively stable operating temperature of a degree or two in either direction combined with first turn on vs. extended use.  For low voltages far from the calibration point, such as 1V, (or 10V with the 50V range) it is very solid and that is where many times it is only a few counts out.

As a quick test I just did another 5V DC calibration at 70F and then tested the meter at 65F and 75F.  In that case it was about 30 counts off in either direction.

At 40F it was about 85 counts low, at 10F it was over 250 counts low.

So I think if you purchase the BM869s, buy a known 5V standard as well and you will be in better shape to utilize the full 500,000 counts.


Probably no surprise to anyone, my U1461A in 66,000 count mode with the factory calibration was about 5 counts high at 6V at 70F, basically spot on at 40F and about 5 counts low at 10F.

[Fungus]

So essentially, your Brymen is 5,000 count DMM 

No, it's a 500,000 count DMM. 

So I think if you purchase the BM869s, buy a known 5V standard as well and you will be in better shape to utilize the full 500,000 counts.

Well... mine's a BM857s but the same thing applies. 

I've got a 10V/7.5V/5V/2.5V standard that's calibrated to 4 decimal places. The meter agrees to within a digit or two so that would make it about 0.002% accurate (an order of magnitude better than the paper specs) and in normal 50,000 count mode it's accurate down to the last digit (plus or minus a couple of counts).

[J-R]

There is a reason why manufacturers give you specifications table. On every single meter, on every single measurement and every single range you have some (different) accuracy specification. It is user that has to know them all and know how and when to apply them.

I think that is the point he is making, looking at the published specs you could not really trust the meter much.  For my BM869s, that was my gripe, the DMM met specs so could not be sent back under a warranty claim, but in reality nobody in their right mind would accept 5.00112V in 500,000 count mode or 5.0011V in 50,000 count mode.  No better than some of my 6,000 count DMMs.

For example:

https://www.eevblog.com/forum/testgear/brymen-bm869-and-bm867-batch-calibration-check/

BM869 #1 reads 4 counts high, 5.0004V.  So really you can only utilize 5.000V because the 4 is clearly "incorrect".  My BM235 displays exactly 5.000V for 5.00000V, for example.  So the BM869s has no use in this specific case.

For only tracking changes, then it is not as much of an issue, so I am just focusing on the absolute values here.

Another plug for the DMMCheck Plus, which now has a Capacitance Reference Board option.  For anyone who has a mid-low range DMM I think it's a good thing to have.

[bd139]

Great review. Can you let us know how the continuity is on it. A lot of the earlier TTi meters had an abysmal implementation.

There are quite a few TTi fans in here. They make decent equipment at a not too scary price.

[eevcandies]

you gotta watch res vs accuracy...i have plenty of res, but the accuracy is only maybe 5-8 ppm

[kish]

Great review. Can you let us know how the continuity is on it. A lot of the earlier TTi meters had an abysmal implementation.

There are quite a few TTi fans in here. They make decent equipment at a not too scary price.

What would you like to know, specifically? I don't have a function generator, so I can't test duty cycle response (Hz).
But latching speed is quite good, when testing manually via probes. I would say it's on par with my Fluke 87V. I'm using Probe Master 8000-series test leads. I haven't tried the stock leads yet.

From the owner's manual:

6.7.1 Continuity Measurement
Setting continuity causes the 1000 Ohm range to be selected such that readings below approximately 10 ohms will sound the continuity buzzer. Readings above the range maximum will show {OPEN}.

[Fungus]

I think that is the point he is making, looking at the published specs you could not really trust the meter much.  For my BM869s, that was my gripe, the DMM met specs so could not be sent back under a warranty claim, but in reality nobody in their right mind would accept 5.00112V in 500,000 count mode or 5.0011V in 50,000 count mode.  No better than some of my 6,000 count DMMs.

That's true for a randomly chosen meter, buuuut ... if it's your meter that you use every day and have cross-checked against voltage references and other meters then you can know exactly how much to trust the numbers on screen.

Plus: A digitally calibrated meter with lots of digits should always beat a 6000 count meter in real life.

[coromonadalix]

+1  fungus,    you have up to 12 digits meter for tons of $$$ for metrology purposes 

If a 6000 count dmm was this precise, all  4.5, 5.5, 6.5, 7.5, 8.5 and more digits would be useless,  i would love to see the meter internal reference used  lolll  maybe an ltz1000 ?

My 34401a  would cry   

[kish]

Today, I'm looking at the temperature measurement capabilities of the Aim-TTi 1908.

The unit can be used with PT100 thermistors in 2-wire or 4-wire configuration, or with PT1000 thermistors in 2-wire or 4-wire configuration. It can measure from -50C to 400C with either thermistor. The accuracy is 0.05% +- 0.5C:



Ideally, I wanted to get a 4-wire PT1000 thermistor for the highest accuracy, but those are difficult to find at an affordable price. After some research, I found out that 4-wires aren't as important with PT1000 due to the higher resolution it provides, so I went with a PT1000.
I picked up an Atlas Scientific PT1000 temp probe (datasheet), plus a BNC to banana adapter.



As far as accuracy goes, how is that calculated? Does the error from the DMM (0.05% +- 0.5C) add linearly with the thermistor accuracy (0.15%)?

Before measuring, I first went into the utilities menu and changed the temp reading mode to 2-wire (defaults to 4-wire).
Then you hit temp option and hit right to go into PT1000 mode. Here is the temperature reading at room temp, seems "about right":

In Farenheit:


In Celcius:


As far as high temperature measurements, my plan was to test the probe against my TS100 soldering tip set to 200-400C.
Unfortunately, even with the fattest tip for the largest surface area, and me pushing down hard as I can, the temp was rising very slowly (0.1C per 1 sec). After quite some time, it settled and wasn't able to measure higher than 100-120C (with iron at 400C):



I know for a fact that the temp reading on the TS100 soldering iron is fairly accurate by using it over the years. It also melts solder like butter. Perhaps either the DMM is faulty, or the thermistor is faulty, or I'm just doing something wrong.
As as sanity check, I switched to 2-wire resistance mode while measuring temp, and the resistance values were correct (e.g. ~1250 ohms at 64C).

Temp calculation equation:

[Thomas]

You won't be able to heat that hunk of metal with a soldering iron tip. There will be a (small) gap between the sensor and the thermowell, which doesn't help either.
Have you tried without the thermowell?

[2N3055]

Test with boiling water.. Water have thermal capacity needed.

[kish]

You won't be able to heat that hunk of metal with a soldering iron tip. There will be a (small) gap between the sensor and the thermowell, which doesn't help either.
Have you tried without the thermowell?

Yeah, these results (except the ambient) are without the thermowell, soldering tip directly against the thermistor metal can. With the thermowell, it takes even longer to heat up (understandably so).

[Thomas]

Yep, that makes sense.
For measuring soldering iron tip temperature, you'll need a small mass sensor.

Accuracy errors are added, because both can be on the edge of their spec, in the same "direction".
Not very likely, but it can happen.

[voltsandjolts]

I bought the TTi1908P, with the ethernet and GPIB interfaces.
Really it was the battery option that sold it to me. I do 90% bench work and the battery covers me for the rest.

First impression, out of the box, it really does feel like a quality bit of kit.
Nice metal enclosure, buttons feel good and the display is clear with good viewing angles.
Safety fuses are accessed via a small 'hatch' on the underside.
Can't see any 'warranty void if broken' labels...just two screws to undo at the back and the lid comes off.

PCB assembly is a tidy build. Gold plated board, brand-name components.
Chunky vishay in-guard/out-guard isolators, NXP MCUs, dale and caddock resistors.
AC RMS converter under the shield includes AD8436.
ADC is TI ADS1246 24bit sigma-delta, using 2.5V reference provided by a SOT23-5 marked ACQD == MAX6071AAUT25

Better put it back together and try using now

[voltsandjolts]

So far I only have two negative comments about this meter;

1/ The 500 datapoint log memory is way undersize for this day and age. Was it actually worth the effort implementing all the firmware to support datalogging and only have 500 datapoints?

2/ The TTi PC Software is not great, functional on Windows for getting readings and retrieving log data to CSV but the charting is laughably bad.
To help with this, I wrote a 1908 driver for the popular Test Controller software which is available on this forum.
The driver will hopefully be included in the next release, until then just drop the TTi1908.txt file into the drivers sub directory.
Test Controller here:
https://www.eevblog.com/forum/testgear/program-that-can-log-from-many-multimeters/msg2981004/#msg2981004
Driver submission is here:
https://www.eevblog.com/forum/testgear/program-that-can-log-from-many-multimeters/msg3591749/#msg3591749

To begin, run Test Controller then go to load devices tab, click ADD button and find TTi1908.
Edit the COM port / tty appropriately. Baud rate doesn't matter, the 1908 implements CDC but ignores baud rate.
Click RECONNECT and the 1908 should go into remote mode.

On the command tab click MODE and select the meter mode you want VAC/VDC/.. etc.

Live readings should now be available, to log and chart data click the LOG button and select the log rate.

[gamalot]

you gotta watch res vs accuracy...i have plenty of res, but the accuracy is only maybe 5-8 ppm

I have this NF WF1946B function generator too, and my wife said it is the most beautiful one of my 50+ benchtop test and measurement equipment.   

[voltsandjolts]

Support for this DMM is now included in Test Controller V1.7
Download link at the bottom of this page:
http://lygte-info.dk/project/TestControllerIntro%20UK.html

[voltsandjolts]

Something to be aware of with this meter is the lack of jack detection on the current inputs.
This means there is no warning (on display or from buzzer) when you've selected a non-current range but leads are connected to the current jacks.
Same when you select the 10A range but leads are in the 1A/mA jack.
So far, I've been lucky but one day....

[Kleinstein]

Jack detection is not common with bench meters.  There is at lest 1 that at least uses LEDs (as some kind of ring at the terminals) to indicate which terminals to use. This is nice, but the exception.

Some of the bench meter allow to use amps and volts more or less simultaneous, though with reduced accuracy from a common terminal.

[voltsandjolts]

I generally use this meter running from the internal battery, using it as a replacement for a handheld DMM, and had that mindset here so hence my quibble.
Need to re-train the brain!

[PeterKlop]

Hi.
I was wondering.
The display of this meter looks like really low contrast with the blue numbers on the blue background. The fluke meter has much more contrast in these pictures. Are these pictures representative for what you see with your own eyes?
If you don't mind i ask you this question because you have bought the meter and the thread starter (kish) doesn't seem te be active on the forum anymore.
How do you like the display?
Kind regards,
Peter

[voltsandjolts]

Hi Peter,
I bought my meter in May 2021 and the display is not blue at all.
With the backlight off it looks like "standard" LCD (contrast is OK in daylight).
With the backlight on digits look black against white background and contrast is excellent.

On a web search, I could not see any blue display versions...maybe it was an older fitment?
To my eye, the contrast looks just like the attached picture.