If you have been following my recent high-output/multi-emitter light reviews recently, you may have noticed some discussion (on my part) of the accuracy of my standard lux light meter for throw measures.
I have a custom set-up with a permanently-mounted sensor for my lightbox, where I continually monitor the internal calibration for my relative output measures. I derived estimated lumen conversions from that setup using the method discussed here: How to convert Selfbuilt's Lightbox values to Lumens
But in terms of direct beam intensity measures and beam distance, I have been using one of the inexpensive lux lightmeters available at a number of deal sites (the common Ceto model). The problem with these sorts of meters is you never know how good their internal calibration is. Note that all meters will go off calibration with time, but I can compensate for that with my reference standards. I recently picked up a second inexpensive meter (the V&A VA800), and found it typically gave ~15% higher lux readings than my long-standing Ceto meter.
As a result, I decide the break open the piggy bank and get a good quality Extech light meter (the EA31 in this case). But I also paid the extra $75 to have its specific calibration tested and certified according to NIST standards.
Note that there is nothing magical about NIST certification - it just means this specific unit has been tested (and tuned if necessary) so that its calibration matches an agreed-upon reference standard. But I do have a nifty looking calibration certificate (good for one year). Again, all meters will drift out of calibration over time, but at least I have a good starting point with this meter - and I can monitor its performance over time to reference standards I've set up.
The other thing I've done recently is to invest in a copy of the ANSI FL-1 testing standard. The actual standard is a protected document, and details the laboratory methodology used for testing (available from NEMA for $60). Although I am most definitely not a proper laboratory, I can at least try to follow the methodology as closely as I can in my testing. The details matter here - the closer I can replicate testing conditions, the more consistent my results will be.
Up until now, I've been working back from intensity measures at 5m (for convenience sake) using my old Ceto meter. From now on, all beam measures will be taken at the full 10m distance and worked back, using the NIST-certified Extech EA31 meter. To designate this change, I will be color-coding my summary tables with orange-highlights for the new measures.
Below is a comparison of my high output lights, broken down by single or multiple emitters. Let's see how the new meter and methodology compares. Note: blue is the old Ceto meter, orange is the new EA31-NIST meter.
I'm showing the percent change for each light, as I believe it is instructive. I know from directly comparing various lights using the same technique that my Extech EA31-NIST reports about ~10-12% brighter than my Ceto meter (and ~3-5% dimmer than my V&A VA800 meter).
However, when using this new 10m measuring distance, I am seeing anywhere from ~10-25% increase in raw lux measures (converted back to 1m). In particular, the multi-emitter lights are all showing a fairly consistent 21-26% increase in beam intensity. I suspect a big part of the reason for this is how long it takes the beams to fully converge in multi-emitter lights - 5m just wasn't enough. This is likely also a contributing factor for some of the higher output single-emitter lights.
The other factor is the larger hotspot when projected to 10m - this gives me more opportunity to hunt around for the area of peak intensity. Recall that the ANSI FL-1 standard for beam intensity/distance is based on absolute peak values. I suspect some of my earlier 5m measures were lower because the diffuser cup over the light meter sensor was integrating too wide an area.
You can really see this effect on aspherics - my peak intensity measures from regional hotspots are much higher now (one example is shown above). That said, I believe you really need to go out to 30m to get a true measure of aspherics - these 10m readings are just too heavily influenced by those hotspots. But it illustrates the principle of why some lights are increasing in intensity readings at 10m compared to 5m.
I will begin updating all my tables in upcoming reviews with the new measures, and wanted to provide a reference here for those who were curious about the change.
Keep in mind one thing hasn't changed – I typically only have one sample of each light I test. ANSI FL-1 requires a minimum of 3 lights be tested for beam intensity/distance (under fully controlled conditions). As always, you should NOT take my results as representative of a given model. It is simply the most accurate picture that I can provide of the one given sample in my possession.