I never stated someone MUST go out 10m or more for accurate readings. 90% of my throwers will get accurate readings at 5m. A few won’t. My SBT70 powered SR90 simply will not. You call call it user error, crappy meter, whatever you prefer, but that’s a fact. And I also use two different meters, one is fairly cheap, but I also use a Extech meter.
And you’ll likely love this statement - ANSI ratings are not always correct. Yes I said that. I say that using a TN31 for example. Thrunite rates their TN31 XML at 75K, but they also rate it at 700m. Their TN31 XM-L2 they rate at 90k, but 840m. Obviously those numbers don’t match up.
Now why is that?? Quite simple - a serious thrower like a TN31 will not get true throw ratings at the ANSI measured 1m. And I don’t think you will dispute that 700m and 840m is actually pretty close to correct for these lights.
So what I’m saying is I could care less what a light is ANSI rated at. I care what a light can actually throw. ANSI throw is good for a baseline, but it is after all a calculated estimate.
And as far as bad readings at longer distances, you can get VERY accurate readings at 100m or more. It’s not that hard. The likelihood of your numbers being wrong at 100m on a thrower is no worse odds than they are at 1m. Or even 3m.
You do know serious light builders like Saabluster even tests at distances of 8-9m? And he has stated he needs to start increasing his distances. He knows these LED’s being created lately, meaning 200-425K and more like the DEFT-X, need more distance for accurate numbers. Do you doubt him too?? I mean simply look at some of these lights beams - you can easily see the beam does not converge on many for several meters. You can see the beam coming out of the light is more intense on the outside edges of the beam vs the center as it comes out of the light. And then it merges into one solid beam. Have you not seen this?
Maybe we should let Mythbuster doing some experiment on this? 
Yes, I know at close distance there will be a dark spot at the centre that will affect measurement. Depends on diameter of reflector, suitable distance should be chosen. We do know that the longer the distance, the more complete it converge. But I don’t think your SR90 need 10m or longer distance to get the measurement because her reflector diameter is not really very huge.
Well, if someone can come out a formula that tell us in theory that let’s say from 3m to 20m, what is the increase of candela in percentage. Then we will know is it worth to go so far distance for throw measurement.
Wow, maybe I missed it, but haven't seen much posted from you on readings off these instruments, or on specifc lights. What you have sounds like a great resource. It's all good -- but not sure why you continually state that some think they know better than the standard though. We do the best we can with what we have and look for comparitive results to see if we are in the ballpark. I have no problem listing lumens and kcd measurements because it's all relative, think it's pretty close to reality, and I know and I thought everyone knows it's based on my instruments and measuring methods which are far from perfect. I just hope more guys get into light metering and attempt to calibrate to their instrument and setups, and post results.
It would be really cool if mythbusters would test some lights. Not likely our little hobby and dispute would be worthy of them making the effort though 
That’s a shame since its one of my favorite shows.
I guess I’ve just got use to using the longer distances since I’m one of the few light nuts here who actually has a another nut in the same family, my brother, to help on testing. As you agree 1m isn’t enough for some lights and we discovered that pretty early, again on the sr90. So we started increasing the distance tested on most throwers. More than is needed on most I freely admit. Usually 15m. But like I stated above, you will never get a false throw reading testing a light a 15m, especially indoors, like we do. We do test longer than that on some lights, though mostly just for the fun of it. Just gives us another reason to run around outside shining our lights
Calm down. Why do you think you know better than ANSI standards when you don’t even know what they are? Do you think the scientists and engineers who developed the industry standard fell off of a turnip truck?
You claim: “ANSI ratings are not always correct” and then try to support that with: “a serious thrower like a TN31 will not get true throw ratings at the ANSI measured 1m.” Well there is no “ANSI measured 1m.” The minimum distance is 2m. The other official distances are 10m and 30m and then numbers must be calculated back to 1m.
ANSI ratings are MORE accurate than what you can do with your current equipment and methods because:
1) They require a meter calibrated by an accredited commercial third party service showing traceability to NIST standards.
2) They’re an average of multiple lights.
3) Readings must be taken between 30 seconds and 2 minutes. Simply using “max save” as you suggested a few posts ago can inflate readings by 20% or more.
4) Tests are standardized to laboratory conditions and include other factors such as temperature, relative humidity and ambient light.
As for readings from long distances… there’s no advantage to going beyond the point where readings converge. Luminous intensity falls off as the inverse square of the distance from the source.
On my 3rd TN31 mod, getting about 8-9% lower output. swapped heads, issue followed the head (not the batteries or carrier or tail switch). Amps at the carrier is same as other units, so, think'n it's the emitter - will swap star/emitter with my good TN31 to isolate it down. Maybe will take a LED amp reading at that time using the solder methods. Ohh - also, I removed the stock resistors on this light. Tried both a solder only (blob), then put in copper braid for jumping the pads from the resistors - either method got same results. So, can't imagine removing the stock resistors cause this lower than expected result -- on the other lights, I left the stock resistors in and simply layed a single copper wire over each, and soldered them in.
Looking at my very unofficial LED tests (https://budgetlightforum.com/t/-/20318), the U2 1A from IOS did much better than the IOS U2 1C's, also about 8%, though tough to compare with one on alum the other on a Noctigon, but at 2A, should have been about the same I believe. Now I'm wondering if the IOS batch of U2 1C's wasn't so good? Should find out this evening...
Update: Crap! Looks like Match's tests do show an alum vs copper difference at 2A -- don't have the raw #'s, but from the graphs here: https://budgetlightforum.com/t/-/16621 I'm guestimating 6% (maybe 775 vs 825 maybe?). It's in the ballpark...
I sure hope not. I just received a couple dozen of these. Looks like I need to build a small test fixture for my lux meter and power it with a wall wart to confirm emitter performance before installation. Sheesh!
Sorry - no, I'm thinking the IS U2 1C's now are about the same as the IOS U2 1A's based on the alum vs copper mounting from those tests. The ~2% difference could be attributed to a lot of things, well within margins of error and variances (converting graph to #'s, SinkPAD vs Noctigon, maybe tint??, etc.).
Weird - I don't think (not sure) we have any test results showing output from different tints of the same bin... I always thought 1A's would measure slightly brighter than 1C's, but a lot of issues there with human vision, how tint effects the meters, etc. Oh boy --- please dont' comment on this... please... 
Looks like I won’t finish mimicking your mod tonight Tom. I got hung up on the battery carrier. Before tonight, I had just cleaned it with alcohol and checked for reasonable flatness. I felt that the battery carrier mods mentioned in the OP were good ideas, but not necessary in the first round of mods.
Tonight I decided to go a head and true up the flatness of the carrier ends by running them over sandpaper on a flat surface. Seemed like over kill, but I figured it would go fast. What happened was unexpected. There is some kind of coating on the metal that I think probably creates resistance. It is probably designed to prevent corrosion. I am now going through all the connection points (there are a lot of them) within the carrier and to the carrier. And then to the tail and the plate that holds down the driver. I think this is going to significantly reduce resistance from the batteries to the driver. I may be wrong. I will know when I take the first measurement of the emitter current with the driver as I currently have it configured.
Anyway, that is were I’m at right now. I’ll hopefully update tomorrow.
Not drawn to scale because it was a 2 minute drawing in ms paint.
I know what rdrfronty is trying to say. He is trying to measure it at a point after the point converges, which is when it starts to diverge.
When calculating back to 1 meter, measuring at 2 meters will read too low, 5 meters will be too high, and 10 meters will be the most accurate (for longer distances).
When it comes to longer throwers the convergence point is actually pretty far away. On my modded TN31 it’s actually around 5.5 meters. That’s where the numbers get inflated.
I measured my TN31 at 6, 7, 8, 9 and 10 meters and when calculated to 1 meter all got the same reading, which was lower than the 5.5 meter reading.
So consequently, you can’t just measure any light at 2 or 5 meters and expect the number to be the same as when you measure it further away. It’s fine for less throwy lights, but when you bring in longer throwing lights you have to measure it further away.
Of course the numbers won’t be completely accurate, but it’s a ballpark number and they’re all relative (which is what Tom has been trying to say).
Sorry to interrupt, but does anyone know how to remove the driver?
I would like to replace the wires with thicker ones but have no idea how to take the driver out.
What I have here:
http://imageshack.com/scaled/medium/89/91p5.jpg
Thanks guys.
Hhmm, from there, the driver should just pull away. On all the ones I've worked on, no glue or anything, and the LED wires are long enough to pul the driver out pretty completely. Maybe it needs a little coaxing... You could try a pick or something in one of the holes to pry or rotate it around - be careful of course. You need access to the LED/star - I think from the pic, the head is already off? Hard to tell.
Werid… Yup, the head is off and the PCB is desoldiered.
I tried prying the holes with a tweezer, but I’m afraid it would loosen the screw hole.
Oh well, I can live with slightly thin wires. Thanks for your help!
Did you try just jiggling around from holding onto the spring? It should definitely not be glued and you could tug pretty good - think I opened 5 lights, didnt' have any problem I can recall.
Great, I’ll give that a go.
I just tried that, but no dice. I think the spring became stretched, not sure.
I think I’ll ask Vinh, and hopefully he’ll give me a reply.
My driver was stuck. I had to push it out via one of the emitter wire holes. Be careful, that driver is loaded with components. I was lucky I didn’t hurt it.
Interesting.
I think I’m going to try that, but first I would have to push one of the leads inside the hole.
If I don’t get the driver out… 
Yeah, I had to push the wire through and I was nervous about that too. My wires were quite short. To short to pull the driver out enough to see the underside. There are some empty areas. Just feel around real careful before pushing.
When you get it apart, check your pill level under the emitter base. Mine had a slight hump and a nipple in the center.
I think what you see in your measurement is not same with rdrfronty. When you say your 5.5m measurement is higher than from 6m to 10m, it is higher by how many percent?
What do you mean not same with rdrfronty?
I don’t recall the % difference as I wasn’t exactly testing it for that reason but I do recall the 5.5 meter mark being higher than my 6-10m readings.
I’d do the test to find out, but I don’t really see the point - it’s different for every light (with a reflector). Is my diagram wrong?