So I finally got to get my hands on an LXT myself and did some side by side testing. I want to thank Kyle Grob of KGmade for sending me his LXT to use. Kyle and I have spoken on the phone a few times but we have never met. When I asked him if he would send me his LXT he didn’t even hesitate in saying “sure no problem”. He said he has done it many times and the 2209 has always given him lower numbers than the LXT but they were close.

To get things started I simply took the LXT out of the box turned it on and set it up next to the 2209. The first couple tests were what I have experienced in the past, the LXT gave some pretty random very low numbers. I called Kyle and told him what was going on. Kyle got me on the phone with one of his guys Chad that runs the LXT the most. Chad got me into a new screen to use that held the Peak Impulse number. When in this screen you have to press 4 buttons between each shot you take to reset the LXT. When in this screen the numbers were a bit closer to reality. In my testing the LXT always gave low numbers ranging from .1 dB to 3 dB. Much better than what I have seen in the past but nothing that I would be willing to use for stack development.  I did a 10 shot MilSpec test that Kurtis (our lead engineer) did some statistical comparison of. Here is the email Kurtis sent back to me

Here is a single string from the 2209 and LXT sampling the same shots from 20" .308. Mil-spec test with the two meters lined up parallel with the gun barrel so they were both presented the same event without one influencing the other (they were front to back). At first glance the LXT is only off by about 1 dB but when you look at each individual data point you can see that the LXT variation is not consistent. This inconsistency is caused from both the low sample rate and high rise time.

I also performed a statistical comparison of the 2 data sets. The dependent means T-test compares multiple data points of the same event and shows if the data collected are statistically the same. This test uses the mean from the 2209 (136.375) and compares deviation of each event (gun shot) to the mean of the true data set (2209).  A p value of 1 would mean the data points are exactly the same and a p value of 0 means the data points are not the same. In this case  the p value is 0.001671 which means we can say with 99.8% certainty that the data collect by the 2209 and LXT are not the same. The LXT is not a mil-spec meter and data from it cannot be trusted as "true" to the event.

Here is the data string





After a few days on the range doing some side by sides after learning the better way to run the LXT I can see how some of you say they are always pretty close and you don’t understand the crazy results I and others from the industry have seen with the LXT. I will admit it is better than what I have seen in the past but not something I could use to develop baffle technology. Anyone using an LXT and not doing the 4 button dance between shots can totally discard the numbers. I don’t know why the numbers are so different between screens but they are. Peak Impulse should be Peak Impulse.

For Kyle to say 2209 is always quieter than LXT for him and me seeing the reverse along with the comparison from Kurtis as well as a few days on the range doing side by sides I could not trust numbers from an LXT.

After having some hands on I can see how the LXT is very easy to use completely wrong. It makes the numbers from the Guns & Ammo 338 article and other times I’ve tested side by side make more sense to me. Even when used correctly it is insufficient for my needs as a manufacturer and should never be used to give industry numbers.

Excellent info. Thanks for taking the time to conduct the experience to, and then to do the write-up.

The meter just had some software to help with looking at wave forms or something. KG can reply to that. I do not think it is good enough to use. Like I said my results were opposite of the ones Kyle expected.That says a lot right there. The string I used was the closest to being the same. I actually was doing some stack development work and was able to collect a lot of data. The meter is useless for that for sure.

The information you posted was good, but I am not sure I agree with your conclusion. From what you recorded, I think I am correct iin assuming you were using either A or C weighting and not unweighted (also called Z-weighting). When using A or C, the response and rise time are degraded to that of a 25-or-so microsecond meter. In your string on the 2208, there was a shot-to-shot variation of slightly over 2 dB with similar variation on the LxT.

Had you used unweighted, the difference between the LxT and 2209 would have been on the order of roughly 8 dB rather than the 1.06 that you showed. In that case, I would agree with you without question.

With A or C weighting, the results you obtained with the 2209 are in line with my LD-800B, which has a measured rise time of 13 microseconds. The rise time is inversely related to the highest frequency the meter will measure without distortion, and this frequency defines the sampling rate.

Obviously, the LxT, 800B, and 2209 are not as accurate as the Pulse system, especially unweighted. We had a Pulse demo here at Gemtech a couple of years ago, and compared it to the 800B. Obviously, unweighted, the Pulse read higher. But with C-weighting (closer to correct than A), the difference was minimal. BTW, the 1/4" pressure mic with its 5 microsecond rise time will limit the Pulse's response -- you need the 1/8" pressure mic.

Is the LxT ideal? Definitely not, but it is a lot better than 95% of the meters out there on firearm sounds -- especially when looking at the meters that a number of LE crime labs use. It at least has a peak detector. The industry standard (and permitted in MIL-STD 1474D and before) permit the use of A or C weighting. For these parameters, it is acceptable. The 2209 is considered pretty much the standard of our industry. While the LxT reads slightly over 1 dB low in your example, your round-to-round ammo variation is greater than that. Is averaging 10 rounds adequate? Sure. Would the Std Dev be better at 100 rounds averaged? Possibly. Do you really lose that much averaging only 5 rounds? Not really.

For most purposes, I believe the LxT is adequate, and for the price, it is a better choice than most of what is on the market. If you want really accurate hearing damage risk, bypass the 2208 or 800B and step up to the Pulse system and do not use weighting.

Just sitting here for a minute i may have missed a number, but the 2209 shows 3db extreme spread and the lxt ~2.2db extreme spread implying the lxt was potentially 25 percent  tighter on analysis for r&d purposes.  The 1.3db disparity is the difference between the 2209 and lxt reported average result (which would effect an advertised specification), but the r&d message would appear to be be 25% clearer here. That should be a set of numbers to compare with the next set of numbers, so the exact value of all numbers  being 1.3db off is irrelevant really to r&d.

Thats the way the data could be crunched if you remove the assumption that the 2209 with the man readable analog dial has to be more accurate, because these results imply slightly better accuracy in the lxt could be possible.

I would walk away from that test saying either meter could be used for r&d if this is indicative of performance but 4 buttons is annoying, and the b&k2209 should be used for specifications simply because the lxt is under-reporting by about 1.5db.

Granted this topic is a source of much contention and for that reason i feel the decision to use an lxt would be better avoided for pr reasons. Thats been the clear message im reading between the lines.

Just know it'll read bwteen 1.5~3 db low and inconsistently so. If you want to be taken seriously, use the proper equipment.  For industry professionals, the 2209, Pulse or any meter that meets the  MILSTD1474 spec should only be considered.

Somewhere there is a sound engineer in a lab who knows exactly how accurate the 2209 family is across the 110-150DB range where the sound suppressors are typically tested.  Whoever that guy is could say, "yeah all over the place" or, you know really that's similar accuracy because....

I'm not a sound engineer so I don't know what the meter to meter consistency of the 2209 family unit to unit is, so it's hard to know.  Looking at 2.2 vs 3DB extreme spread, and 1.03DB undersampled I could say yeah that performance is pretty accurate and very consistent with Phil's PDF which suggested it isn't a poor system.  You could slip calibration 1DB and center that and you might be within +-.5DB for all I know, and with possibly tighter extreme spread.

For years people have been talking about atmospherics and temperature and saying the only relevant test is the same place, same meter, same day, and I've always thought that was mostly bullshit because I see the 2209 performing pretty stable day to day in various atmospheric conditions, but those arguments make the "this is a horrible result argument" seem less accurate as a generalization.   Were those people really saying a test of a product here, vs there on 2 2209's was irrelevant because of a difference of less than 1 decibel in the results?  I would think not personally.

Nobody ever really said Al Paulsons 2209 results were irrelevant when he was metering with a ~2.5" foam windscreen on the end of his microphone.  I'm sure that changed sound more than 1DB.  I just took his results and added 2DB for my ballpark assessment of where a product would be to the rest of the people who metered without 2.5" foam windscreens.   The reason why the windscreen makes no sense is that the wind is almost never a problem even at 110DB even with no windscreen, certainly never in the 130's.

Here it was 1.06.  I think Phil had 1.5, 1.3, .4, and .5DB undersampling.  The average of those is 0.925DB- fairly consistent with this 1.06DB result.   A 1DB adjustment to calibration puts all those numbers within -.5DB+0.6DB of the 800B or 2209 numbers.    I think inside +-1DB would be more accurate than the historical accuracy of 2209 testing by accepted sources.

This is where it really gets into the minutia and where the old standards like using CCI Standard Velocity .22lr, and M855, factory 147grain ball 9mm, factory 230 grain ball .45,and M118LR 7.62x51, become important for one result to have even this level of accuracy to another.  Platforms have to be std too for this kind of relevance.  A lot of things that just don't often happen.