Tasi 632A luxmeter, some tests

Not long ago clemence discussed with me buying this luxmeter, and looking at the meter and specifications I got enthousiastic and hopeful that this could actually be a good chinese luxmeter, or actually taiwanese, for a reasonable price. Especially the claim of a spectral accuracy of 8% caught my attention, in the first place because they actually listed a number for that accuracy (instead of the usual vague claim that the spectral reponse is in accordance with the human eye response (photopic) curve, even Extech does not tell you more than that), but also because 8% is really good. Still a bit unsure because the Tasi brand also sells those 16 dollar luxmeters that do not look really different from this one, I bought one for $168 here: https://www.aliexpress.com/item/TASI-632A-Digital-Light-Meter-USB-DC-Power-Luxmeter-Meters-Digital-LCD-Backlight-illuminometer-Storage-128/32791440073.html

Yesterday I got it in the mail and luckily I had time for tests. :-)

The first impression was not good, it came wrapped in the usual plastic foam , in which I found an all too familiar crushed thin white cardboard box. In that box was, loose without retail packaging, the luxmeter, a sensor with a lead, a USB cable (the meter uses the outdated mini-USB connector) and a 1-page photocopied manual.

Not let down by the packaging, as a flashaholic I'm used to sometimes very good stuff packaged like that :) , I inspected the outside of the luxmeter: thick plastic and a thick rubbery bumper, big sturdy sensor with a massive lead, the connector looks like the well-proven connector commonly used in personal computers (some assembly damage to the housing there). This luxmeter looks ready for some heavy duty use!

Here's (most of) the manual. Quite informative but still sloppy, with some unclear language and repeats:

The screen is nicely back-lit (push button for that) with a time-out of 30 seconds, which is useful. The readings are slow compared to my Mobilux: just 1 read-out per second. This can be annoying when you do many quick measurements or monitor fast events.

I will not go further into the operation, and also not into how to use the USB-connection because I'm not very good at that . Perhaps clemence can chime in in this thread with some details if he has received his copy.

EDIT 7sep2017: Clemence evaluates the connection to a computer in post #11, thanks!

My first and only interest is: how good is the spectral response and calibration, and does it work well for how we flashoholics use a luxmeter.

Here is some testing against two other luxmeters that I own, my top-range Mobilux class A luxmeter, and the über-cheap (12 dollar) Uni-T UT-383 luxmeter.

(I had done some checks on the Uni-T meter before, posted in the Mitko-thrower thread, these tests are in line with those measurements)

What I did is comparing the readings of the three meters in a as precise as possible way for a wide variety of natural and artificial light sources. I did not pay special attention to the linearity of the response (i.e. does the meter correctly measure 10 times as much if it sees 10 times as much light?) because thusfar I have not seen great linearity problems in even the cheapest luxmeter that I have.

Here's the rough data FWIW:

For each source I placed the sensor on the exact same spot in the beam, took readings from all three luxmeters starting with the Mobilux, and ended with a second reading with the Mobilux to see any output deviation. All light sources were not truly constant but constant enough for the comparison.The actual values of the numbers are not relevant here, often medium settings were used, and I used small light sources at closer distance. What matters is the type of light source and how different the luxmeters read the light source.

On the separate piece of paper are readings outside, first set in the shadow of the overhang on my balcony, second set aimed directly at the morning sun. Because sunlight varies per second (there's always some varying haze in the air here), I did four and five sets of readings alternating the luxmeters. This gave a good average IMO.

Here's a table in which the above measurents of the Tasi and Uni-T meters are represented as a percentage of the Mobilux reading.

Table: % output reading deviation

compared to the mobiLux meter

mobiLux

luxmeter

Tasi 632A

luxmeter

Uni-T UT383

luxmeter

50W halogen bulb 0 +4% +1% -3%
outside, clear sky in the shadow 0 +10% +7% +4%
outside, clear sky aimed at the sun 0 +7% +4% +6%
Crelant V4G with Nichia 219C 2700K 92 CRI 0 +6% +3% -11%
X6 with XP-L Hi 3000K 0 +3% 0% -10%
Lucky S. mini20 with Oslon Square 4000K 92 CRI 0 +3% 0% -2%
Supwildfire thrower with dedomed XP-G2 3D 0 +5% 2% -14%
Supfire F5 zoomie with dedomed XP-G2 S4 2B 0 +2% -1% -13%
Ultrafire C8 with dedomed cool white XP-E2 0 +4% +1% -6%
SWM D40A with cool white (6500K ?) XM-L2 0 -3% -6% +5%
XM-L colour zoomie, 625nm red die 0 +27% +23% -6%
XM-L colour zoomie, 532nm green die 0 -8% -11% -15%
small flashlight with 470nm ice-blue XP-E2 0 -32% -34% +295%
XM-L colour zoomie 450nm blue die 0 -53% -54% +690%
flashlight with 400nm purple generic chinese led 0 -76% -77% +5000%

So one of my fears appears not true, the spectral response is quite different from the cheap Uni-T meter, and better. Over the whole spectrum it does not differ dramatically from the MobiLux meter that I use as my standard of a good spectral response (I'm sure that is not perfect either but the spectral response from the MobiLux meter is the closest I have to the defined spectral response that is the basis of the lux), while the Uni-T meter severely over-reads in the blue region of the spectrum. The Tasi meter compared to the MobiLux meter under-reads the blues, but not dramatically. It does cause some under-reading of very cool white leds, like from my SWM D40A, but in general the white light sources are measured well.

For all white light sources, except very cool ones, the Tasi-meter reads higher than the MobiLux meter so part of the deviation can be a overall calibration difference. If you let the tasi-meter read 3% lower than what it does, the general comparison to the MobiLux meter becomes better, apart again from the very cool white leds. I substracted 3% reading to get the green numbers. This is relevant for me because I want to use the Tasi-meter for an integrating sphere, in which the absolute calibration of the luxmeter is irrelevant anyway.

I'm not sure yet how to interpret the daylight measurements, they seem somewhat confusing but I may just have to think more about it.

As a small additional test I checked the readings of the three luxmeters pointed straight at the sun and at the (randomly chosen) angle of 55 degrees. No large differences here, at 55 degrees compared to nicely perpendicular to the sun, the Mobilux reads 64%, the Tasi 67%, the Uni-T 62%. No significant reading differences are expected from this in any use I can think of.

Conclusion.

This Tasi 632A luxmeter is clearly not perfect, but I think it is value for money if you are a bit more serious about your flashlight numbers, it is built well but especially the spectral response (=optical filter in front of the light sensor) is some levels better than that of the cheap chinese luxmeters, although the blues seem to read a bit low (relevant for very cool leds). Perhaps the light sensor is based on one of the newer ambient light sensors made for mobile phones that I found to have very good specifications for spectral response (look for the newest Texas instruments ALS!). I have never owned any other mid-range luxmeters like the various Extech meters, but the little information I do have suggests that the spectral response of this Tasi is better than those (don't kill me if I'm wrong).

If you are not that serious about the numbers that your flashlight produce but would like ball-park figures that still give way more information than what your eyes can see, go for a cheap luxmeter like the Uni-T UT383, who cares about 15% :party:

bump, finished the write-up

Thanks for the great detailed review mate.
In conclusion you say the light sensor in the new phones these days are good. Does that mean luxmeter app on a smart phone is better than my cheap UT383? lol

Could be, or not.

Light sensors in phones differ greatly. All I say is that there are some being made with great spectral responses. But they are not in every phone, and I don’t know how well the phone manufacturers make use of the good specs (for example they could screw-up by the way it is built-in optically). To use your phone as luxmeter the least you need do is making a diffuser before the ALS, and you have to figure a new calibration.

As an example, my pretty high end Xperia Z3 Compact has a very average ALS with aweful spectral response. The type of ALS can be found among the information in your phone, and you can then look up the specsheet on internet.

Thanks djozz for the detailed review and comparison on the luxmeter :wink:
I am not expert on these things, in fact “I know nothing” about lux/lumen/candela meters or whatever you folks use to measure that!
So, I fit in the category “not that serious about the numbers”….(yet!) :stuck_out_tongue:
Therefore, my question for you is more about the Uni-T UT383, I would try one to begin with as it is cheaper! Any reliable suggestion from where I can buy it online?

Also, just for testing/play around, can any of you recommended me a “luxmeter” smartphone app (Android) to start with?

I would appreciate your help on this :wink:
Thanks in advance :+1:

EDIT: Just saw them on AliExpress stores :wink: Gonna try one! Thanks anyway :wink:

Thanks a lot! I appreciate the lux meter tests, since it seems to be difficult to find those. +–10% is very good for the price. I wonder how much the filters change through time, especially if you forget them in the sun and don’t use a cap in front of the sensor.

@MascaratumB, I bought mine on aliexpress too, could well be the same seller.

I have never seen any luxmeter review that goes beyond mentioning the specs. It seems to me that there’s the light professionals who know everything about the subject, and then there’s the rest of the world who have no idea what and how they are measuring. And a handful people like me have some clue and want to find out a bit more, without a full grip on the subject and without much equipment for testing.

About the filters: I don’t know more than you probably, colour filters often contain organic compounds that may be destroyed by light over time, hence your concern probably, and why luxmeters always come with covers over the sensor.

Thanks for the review djozz. Does SWMBO have the same enthusiastic thoughts about the meter as you do? :slight_smile:

When she came very tired back from work today, I tried the making of this review as an interesting conversation piece but for unknown reason she was not intruiged by spectral responses and proceeded straight to the couch ???

I must subscribe on this and will read later… :+1:

OK, I finally received mine…

Both of my order and Jos didn’t include the 12V power adapter and software CD. Tried to download it from the official website: http://www.china-tasi.com/en/Product/Digital_Light_Meter/2014/0508/133.html with no luck. Looks like the admin linked the software button to the manual file.
I contacted the store to get the file link. She sent me a corrupted winzip archive at first, later she sent me the good one. Here’s the download link: TASI600series

The program written for all TASI 600 series and designed for WinXP/Vista. The nostalgic UI would not appeal to modern people but I like it. Simple, too simple perhaps. Functions well.

I use Win10 (accidentally upgraded from Win7), I installed using the Vista version

Immediately after connected to the USB port the zero calibrated readings always jump to much higher value. I have to re-calibrate after the USB connection. Open the pull down menu to open the graphic UI.

Unlimited recording time with minimum 1 second interval.

When recording you might want to turn off the auto range feature in the handheld unit. Otherwise the transition between scales (Example: from 20lx to 200lx) will create a spike in the graph. In the pic above the flashlight was in low mode at 49,1 lx. When I switched to turbo mode the reading supposed to be 400-ish lx but the UI had to switch from 200lx scale range to 2000lx scale range. For a second the reading jumped to 4000-ish lx and then settled to the correct value. With (correct) manual ranging this will not happen. This is not really necessary as we can always sort the abnormal spike later in the excel table.

The file can be saved as to .xls or.csv. Very useful for long run time tests. Don’t forget to turn the APO (Auto Power Off) off in the handheld unit to prevent auto shut down after 15 minutes.

So far I like this light meter. Combined with Jos’s report this one has the much better quality (than most cheap MIC products), consistency, and nice features for it’s price.

- Clemence

Thanks clemence, that sounds very useful. Now I have to dust off my ancient windows laptop and see if I can do that too :slight_smile:

I really like that small Uni-t lux meter… Especially when I found it gives “right” numbers when I turn FC function on(seriously… without that turned on I have mixed results and have to recalculate lets say +13% for XP-G2 from your comparison test but when I turn FC on results settles where it logically should be…)
It is also very easy for handling… I can only recommend it and I really don’t need better light meter than it for this hobby. It is better than more expensive TES light meter(i paid around 200$ several years ago) i have at home.
Your percentage data in comparison with real lux meter is also very valuable…

But why not give Tasi a try? Someone will surely like it.

Got an uni-t ut383 and played with it today. Nice cheap meter. I got the best results with the F.C. function off. Compared it to other meters and known specs and it measures spot on for floody lights and about 10% to low for intense hotspots. I only have warmer tinted lights don’t know if this influences anything.

I think every blf should have one for this price.

Well… I don’t know how to explain that but for me(my all lights are over 200+kcd ) it works best with F.C. function on giving higher and more constant results. 1FC=10.76391 lux and online converter here: Footcandles to lux (lx) conversion calculator
I use regular mode for measuring intensity of IR lights(it cant really measure that but it can tell what is happening when you put directly on sensor and if lux raises and stays at certain value it means that IR light is OK if it drops a lot below that value it means something is done bad in process)

Anyway I really like it now it is very good and cheap product same like Uni-t clamp meter brother.

It would be really weird and worrying if setting the meter on fc or lux would matter, the meter should just do the conversion for you, nothing else.

You are right on this. There is exactly a factor 10.76 between the two settings. Does not matter which you use.

It is easier to read if nothing else… There is less digits to follow.

Which also means less precision.
If you want easier to read you can also just change the multimeter range…
Instead of using x10 you can use x100 or x1000

Not on the uni-t ut383. It has an autoscale.