Thanks djozz! Mail is out and i will post the result when/if i get a answere!
Update from TASI:
They told me that the version with the bad sensor is the old version. And they are not selling this version anymore (they wanted to know when and from which shop the old ones was bought).
They also said:
You can buy directly from us, we are the manufacturer, we can assure you our quality. Please don’t worry about the sensor we will send you our updated new product (USD17/PC).
So it seems the new Meters should be with the good Sensor ( But aslong you are not buying directly from TASI you can not be sure to get the new one).
That is a surprising outcome to me, in principle encouraging but time will tell if that is indeed the case.
Thanks Docc for contacting Tasi, even if it is not true what they say it helps if manufacturers know that their customers notice and care about the build quality of their products, which is not obvious in case of luxmeters.
I also hope that it is the truth! I not only asked which one i get i also linked the test 
so they know it is not only one person asking. And that many more people will be aware of their answers.
I asked how much shipping will cost (and ~shipping costs for the bigger countries). If it is not to much (compared to buying on Aliexpress) i will order one directly from TASI. Then we will see which one they deliver. I also asked if they sell directly on Aliexpress (could not find them on Ali).
As I predicted, the bad one was probably an older lower quality version.
Let’s see if buying directly from them gives you the new one.
Good thing it’s easy to tell which is which.
Thanks for the info 
Ok i was expecting to pay a bit more than on Aliexpress. But a total price of 47$ inkl shipping is a little bit to much for my taste (cheapest price on Aliexpress is less than 19$ inkl. shipping.
I think we will have to wait and see what we will get from Aliexpress when the Sellers old stock is sold and new ones are orderd from Tasi.
@ Djozz
Is there any difference in labeling on the two 8133? Something we can ask the Seller before we order (except the red cap at the top)?
I received my TA8133 ordered from BZG Electronics Co,.Ltd this morning
The top of the case doesn't have the red part and guess what ...
When comparing results from various light in my cheap integrating sphere with a calibrated Minolta T-10A, the TA8133 seem to be 30% to 50% off :-( :-(
^ :person_facepalming:
I’m sorry to hear that.
Real sorry Kame Sennin.
But you saved me. I asked BZG Electronics yesterday and they answered that they have the one with the red top in stock.
Now i ordered from frank hu’s store and hope to get the right one!
I have a 6133 on the way from Frank Hu that just left China. Will report back when it arrives.
I think i have bad news. This morning i was contacted by frank hu’s store (some payment issues). I should cancle and please reorder (happend before on Ali). I took the chance to ask him about the 8133. He confirmed our worst fears. Every new delivered Tasi meter in 2018 is the one with black top. And he also has only the new (black ones).
So this makes me think about the answere i got from Tasi directly! (was it a straight lie?)
Are you sure he knows what you’re talking about?
Both of them have a black top, the difference is that the better one has a small red piece in the middle of the black top.
Maybe he misunderstood you?
I send him two pictures. One with the red top and one with the black. He said the black one is new one.
If that’s the case then how did he send a “good” one to djozz a few weeks ago?
I do not know. Old Stock? If you are interested you can ask the seller. The only thing i know is, that everybody is telling something different.
See BZG Electronics. I asked with picture which one they have in stock. Answere was: The one with the red top. Kame Sennin ordered from there and got the bad one.
Now frank hu’s store said he has only the black ones and that it is the new version (knowing that i will not buy this version). But djozz got a good one from there.
And Tasi says black one is old, red one is new version.
I think the situation is simple, aslong you are not having someone in china that can personaly confirm the content of the package you can not be sure what you get!
Maybe ask the seller to send a pic of your item before they ship it?
I was so pissed of this TA8133 thing !
And that classified ad about a T-10M not far from were i live was still here ...Not the latest model like the one i have at work but the previous one, the price was ok compared to the price of a new one but still too expensive for me so i hadn't called but hey it was still around after a full month in the ads and i was REALLY pissed so i called the guy and offered the maximum i could "reasonably" spend for something that i don't really need.
And he eventually agreed :-)
So here is my new toy :
I checked the 3 cheap Tasi’s (and Mobilux) against 4 of the test-lights also used in the OP to get an idea of their differences.
Here’s the data, to explain what is in the table:
*for most lights I checked a few output settings. After every series I re-measured with Mobikux (number between parenthesis) to see any output drop of the flashlights in the course of measuring with 4 meters (there was some).
*the green numbers in the Mobilux row are +3.4% corrections to get the Mobilux to the most recent calibration.
*the green numbers in the Tasi rows are percentage deviation from Mobilux
*the red numbers in the last row was the measurement of the good 8133 divided by the measurement of the 8132. To see if they are similar apart from a calibration difference.
Can’t blame you if you do not follow the clutter above, the conclusions are:
*the re-measuring of the TA8132 gave roughly the same results as in the measurements of the same luxmeter in the OP, but there’s a few percent difference here and there, so this type of comparison is not extremely accurate (assuming that the TA8132’s performance is consistent, which I think is likely)
*the “good” TA8133 has a ~3% lower calibration than the TA8132, so its calibration is not as close to the Mobilux as the TA8132 , but otherwise the error-characteristics are the same. I expect that the wavelength responses are very close to each other.
*the “bad” TA8133, when at all it measures (it starts producing numbers above 60 lux :confounded: ), the numbers are way too low, with lower light levels giving larger errors than higher light levels so this luxmeter is not even lineair! It seems I got a total dud here.
That should be an extremely good luxmeter, I just read the specs. Well done.