LabRadar: Great Data - Terrible Implementation. (partial solution inside)

[shinbone]

How does the Wiser mount attach to the rifle? The website shows a picatinny rail on the underside of the forearm. My guns are mostly light weight hunting rifles built for all-day carry over long distances, and do not have such a rail. The only thing I got is a simple sling stud.

 

[Gone Ballistic]

I purchased a picatinny rail mount that mounts over the forearm swivel made by Caldwell. Really simple to mount and it also has a sling swivel to attach your sling to if you want to keep it on for mounting a bipod or something else. I actually purchased 4 of these units so I could keep 3 attached to the 3 rifles I shoot the most often. They are around $14 each on Amazon. Simple to mount in a couple of minutes. Then the pod mounts on the rail very solid and has a great bracket that holds your information receiver.

 

[shinbone]

Excellent. Thanks for the info!

 

[jim_hahn]

Well, its almost two years since last post on this thread, but this new member (aka me) will try anyway.

I have been playing with my new LabRadar and trying to get info on how the sw works. I also am playing with Excel.

Some questions:

1) I don't see the gross errors in calculating V0 that you see/saw. Do you feel LabRadar has updated their sw?

2) When you cull data, what 'rules' are you using to find/exclude outliers?

Thanks,
--jim

 

[entoptics]

Well, its almost two years since last post on this thread, but this new member (aka me) will try anyway.

I have been playing with my new LabRadar and trying to get info on how the sw works. I also am playing with Excel.

Some questions:

1) I don't see the gross errors in calculating V0 that you see/saw. Do you feel LabRadar has updated their sw?

2) When you cull data, what 'rules' are you using to find/exclude outliers?

Thanks,
--jim

I thought problem was minimized for awhile after a software update, but that may be because I got a better setup for keeping the unit aimed and steady. I still run everything through my spreadsheet, and see the errors still, though perhaps not as bad after said update? I haven't updated again in quite some time, so it may be better with the latest version.

Here's the link to my spreadsheet.

Lab Radar Rescue.xlsm

drive.google.com

My "rules" involve variation from the best fit (highlighted pink cells, >0.15% off best fit), as well as an initial visual culling. I'll remove any that are obviously WAY off the line, then I sort by variation with the macro button, and cull the worst offenders. Doing the initial visual cull can be important. Particularly if there's several outliers at the beginning or ending of tracking, as they can really pull the best fit line off the majority of data, causing the >0.15% metric to falsely flag otherwise good data.

 

[jim_hahn]

Might be better to use email than the forum. Not sure if anyone else still cares and email does not require logging into LRH.
hahn under 02493 at yahoo dot com

OK, so I got your sheet downloaded. The button that says "Sort by variance" invokes a macro, right? I have not used macros much. Are there other macros and how/why would I invoke them?

So if I just copy my data into the sheet, will it automatically (ha,ha) recompute things? Need to go do some stuff right now and won't get back to it for a bit.
--jim

 

[entoptics]

Might be better to use email than the forum. Not sure if anyone else still cares and email does not require logging into LRH.
hahn under 02493 at yahoo dot com

OK, so I got your sheet downloaded. The button that says "Sort by variance" invokes a macro, right? I have not used macros much. Are there other macros and how/why would I invoke them?

So if I just copy my data into the sheet, will it automatically (ha,ha) recompute things? Need to go do some stuff right now and won't get back to it for a bit.
--jim

There's two macros embedded in the sheet.

See post #4 in this thread, and it should tell you all you need to know. There's a video showing the usage as well.

 

[jim_hahn]

Looks good. Congrats! Maybe I have to learn VB and macros.

Video has no audio which makes it tough!

Keyboard command to run macros did not work. Probably because stupid Microsoft changes things every version. Opening the macro pulldown and selecting which macro did work, and then 'run' it.

Took me a bit to figure out both sheets must (?) be open to run the macro. And Excel attention needs (?) to be focused on the tracking sheet.

The green and red boxes truncated my V to only 3 digits.

Just started playing with it, so I may have other comments.


Want to check my hypothesis with you on how LabRadar works:

At the core measurement level, all it is doing is capturing V (based on doppler return) every millisecond.

It DOES NOT capture distance downrange. In other words, x(t) is not captured.

Just like we know it back calculates vMuzzle, it also back calculates the 'distance' column in the track file. So, x(t) is a derived column, not measured.

LR knows distance from muzzle to LR because user enters it as a setting. Given ballpark value of c, if you are one foot away, LR adjusts its time values accordingly by about one ms. The error introduced here, if any, can be ignored - it's only a couple fps.

When trigger signal arrives, an internal counter is started. For some number of ms (10 to 20ms) the LR is not seeing radar returns until the bullet gets in the radar cone cleanly. Then it's off to the races filling its internal array. I suspect this is done in software as hardware would be painful to design cheaply.


To state my above point on x(t) differently: If you know v(t), you can integrate to determine x(t). Just like if you know x(t) you can differentiate to get v(t). (Or a(t) for that matter.)


Separate point based on above: YOUR analysis should not use the distance column because it is based on LR's computed values for distance. If you suspect LR is producing 'some' bad data, your algorithm should not use data (the distance column) based on said bad data. I have no idea if your alg does in fact use the LR distance data.

Again, nice job!
--jim

 

[shinbone]

Jim - Interesting comments on the functioning of the LabRadar. Since radar systems are easily capable of ranging a target, on what basis do you conclude that the LR does not make such a measurement?

 

[jim_hahn]

a) If they measured x(t) they would not have needed to measure v(t).

b) Although I am an EE I know nothing much about radar. But I discussed this with some EE friends who do have first hand knowledge of radar and they said a system that ranged (x(t)) doppler v(t) would be challenging. It would involve pulses being sent - which in and of itself is not tough but processing returns would be tough.

I don't really know the answer to if they range. But they seem reluctant to discuss such details. I am betting a $600 unit uses the simplest system it can, and that would be solely doing doppler. I would be eager to be shown wrong.
--jim

 

[shinbone]

Jim - I myself don't know, either.

I've chatted with the LR guys a few times, and it is true, they are very reluctant to discuss any technical details on their system.

 

[ptosis]

a) If they measured x(t) they would not have needed to measure v(t).

b) Although I am an EE I know nothing much about radar. But I discussed this with some EE friends who do have first hand knowledge of radar and they said a system that ranged (x(t)) doppler v(t) would be challenging. It would involve pulses being sent - which in and of itself is not tough but processing returns would be tough.

I don't really know the answer to if they range. But they seem reluctant to discuss such details. I am betting a $600 unit uses the simplest system it can, and that would be solely doing doppler. I would be eager to be shown wrong.
--jim

My understanding is that the only thing Labradar actually measures is radial velocity, that is -- velocity of the projectile to or from the device. All the rest (= figures written to track files) -- time, distance, linear velocity -- are not measured but calculated.

 

[jim_hahn]

My understanding is that the only thing Labradar actually measures is radial velocity, that is -- velocity of the projectile to or from the device. All the rest (= figures written to track files) -- time, distance, linear velocity -- are not measured but calculated.

No. Time has to be included in what LR does. Every millisecond it takes a reading of the radial velocity and stuffs it into an array indexed by milliseconds since the unit was triggered.

 

[ptosis]

No. Time has to be included in what LR does. Every millisecond it takes a reading of the radial velocity and stuffs it into an array indexed by milliseconds since the unit was triggered.

Of course, the device does have an internal clock. What I meant is that the starting "zero" point is calculated as the point in time when the blast arrives to the mic minus the time it took for the blast wave to get from the muzzle to the mic (as implied from the distance to muzzle as configured in the device settings).