Chronograph Madness

[Gwhite]

I'm aware of the Nate Chrony, but haven't really looked into them in any detail. For tuning purposes, 1% accuracy of absolute velocity is fine, but I need REPEATABILITY better than that, which is a number nobody else specifies, but the Nate appears to. The Combros were apparently live fire calibrated, and I can even adjust them myself IF I had a decent reference.

I've tested multiple pistols with the Combros that showed repeatability of around +/-1 fps, which is suitable for my needs. I find it hard to believe that multiple air pistols could magically cancel out variations in the Combros, so I believe that they are at least that good. Unfortunately, they don't make 'em anymore, their mounting makes them easy to destroy, and there is some evidence that their absolute velocity numbers are suspect, at least to the level of a few %.

For now, I need to do some more testing to see where the HT-X3006 fits into the picture. If that flunks out, the Nate looks like a good alternative. The specs look like they would definitely do the job, but it means re-doing weeks of designing 3D printed adapters for all of the various air pistols I have to support.

 

[NateChrony]

...... I need REPEATABILITY better than that, which is a number nobody else specifies, but the Nate appears to.

For now, I need to do some more testing to see where the HT-X3006 fits into the picture. If that flunks out, the Nate looks like a good alternative. The specs look like they would definitely do the job, but it means re-doing weeks of designing 3D printed adapters for all of the various air pistols I have to support.

The only other Chronograph i've seen with repeatability accuracy is the Pro-chrono DLX, but they specify it in timing ns... kinda frustrating, you'd have to do the maths on sensor spread & calculate a %.

Adapters - all the work is still valid, you'd just add a 20.1mm round end to slide into a NateChrony, which also uses adapters!

World first - work has started on open sourcing the adapters & helping people make NateChrony compatible adapters:

GitHub - natechrony/Adapters: Adapters for the NateChrony

Adapters for the NateChrony. Contribute to natechrony/Adapters development by creating an account on GitHub.

github.com

 

[Gwhite]

The only other Chronograph i've seen with repeatability accuracy is the Pro-chrono DLX, but they specify it in timing ns... kinda frustrating, you'd have to do the maths on sensor spread & calculate a %.

Adapters - all the work is still valid, you'd just add a 20.1mm round end to slide into a NateChrony, which also uses adapters!

World first - work has started on open sourcing the adapters & helping people make NateChrony compatible adapters:

GitHub - natechrony/Adapters: Adapters for the NateChrony

Adapters for the NateChrony. Contribute to natechrony/Adapters development by creating an account on GitHub.

github.com

All my HT-X3006 adapters are already out there: https://www.thingiverse.com/thing:6977381

I think I can just create a single dovetail adapter that fits the Nate, which would allow reusing the mounts I designed for the HT-X3006.

I want to be able to use the sights on the pistol when firing. What is the distance between the sensor bore-line and the top of the Nate Chrony? If it's more than ~ 1.5 cm, how big is the sensor area? On the HT-X3006, I can offset the pellet path a few mm to get the sights above the housing.

Thanks!

 

[NateChrony]

All my HT-X3006 adapters are already out there: https://www.thingiverse.com/thing:6977381
I think I can just create a single dovetail adapter that fits the Nate, which would allow reusing the mounts I designed for the HT-X3006.
I want to be able to use the sights on the pistol when firing. What is the distance between the sensor bore-line and the top of the Nate Chrony? If it's more than ~ 1.5 cm, how big is the sensor area? On the HT-X3006, I can offset the pellet path a few mm to get the sights above the housing.

The distance from center to top is ~15mm & bottom is roughly the same, however alignment is absolutely critical (new software has alignment detection though - so you can detect if there's an issue)
I would not recommend to use the outer surfaces on the Chrony as a reference point, and only recommend using the center bore as the reference.

Also looking at the designs - they're most likely not rigid enough & you'll see errors in fps with the NateChrony. To get the accuracy, I have to detect on a certain part of the pellet, if there's vibrations/movement the front sensor may detect on a different part of the pellet to the rear - at best it'll be variances of ~5fps, worst it'll be 50-70fps (assuming ~500fps speeds).
The adapters will need some form of bracing/contact top/bottom to secure them enough.

 

[Gwhite]

I'll have to check the height of the sights relative to the bore for the various pistols. 15 mm may be too much to use the sights properly on some of them. The 3D mounts are reasonably rigid, and the HT-X3006's are quite light. The cable hanging off the Nate Chrony might be an issue., but it's pretty easy to stiffen the mounts. Aligning to the bore should also be pretty straight forward. It would take about a 1 degree error to change the path length between the sensors by 0.1%. I think I can live with that...

Given that the shape of the front of pellets varies all over the map, I would think the best detection would be off the back end. You'd want to detect the un-breaking of the beam, rather than it's initial interruption. It sounds like the Nate Chrony has a very small (round?) optical beam. A thin, but fan shaped beam/sensor arrangement would reduce the alignment position sensitivity. The big old fashioned open optical units had lenses to help with that.

 

[OldVet]

#1 Combros Chronograph CB-625 MK4. AT DEFCON AIRSOFT.
#2 Combros Chronograph CB-625 MK4
Deluxe kit at Desertcart Oman.
#3 Combros CB-625 MK4 Chronograph
On eBay UK.
Honestly you are all are so smart , but this post was a very interesting read. I hope the above helps you
I believe the first 2 are new ones , and I hope it's the correct model.
Good luck teach.

I have a 3D printed sight with a tube on top, but if the LabRadar beam is off center from the face, that doesn't help much. When I get time, I will drag mine back to the 50 foot indoor range and try again with an SD card so I can monitor the SNR. I really need to fix more pistols, not become an expert on the idiosyncrasies of my chronographs...

I've ordered another HT-X3006, but with the Wi-Fi option so I can collect long strings of data and get them into a spreadsheet more easily. I may try to put the two in series to see if they read the same on a given shot. That would give me more confidence that they are working well. It will take a few weeks to get here from China, but that gives me time to 3D print a bracket to hold a pair. I will test them with one of the Hammerli pistols that showed very small velocity variations.

Given that I apparently can't use the LabRadar in my basement shop, it has VERY limited utility for me. Between the Garmin, the FX, the Athlon and the LabRadar LX, there are quite a few options for small radars with wider beams. I have yet to hear anyone bad mouth the Garmin, and it sounds like a few people have had success with one at short range indoors. The LX and Athlon are both so new that there's very little real info, especially for my use case.

 

[Gwhite]

I really like the Combros, and actually have 5. One I bought direct from the source many years ago, one was new old stock from a Canadian vendor, and three came off of eBay in the US. Two of those have been shot, and someday if I get bored, I will see if I can combine them to get another working one. They appear to be VERY repeatable, but they don't agree very well, so the absolute velocity is suspect. As long as you are VERY careful to align them, and can live with some absolute velocity uncertainty, they are still pretty nice. Some people claim they lose velocity when the batteries aren't fresh, but I haven't explored that.

Besides some additional testing with a couple of my most stable air pistols (fortunately, I have about 50 to chose from...), I may get a Nate just to use as my "gold standard" to evaluate the others I've got.

 

[NateChrony]

I really like the Combros, and actually have 5. One I bought direct from the source many years ago, one was new old stock from a Canadian vendor, and three came off of eBay in the US. Two of those have been shot, and someday if I get bored, I will see if I can combine them to get another working one. They appear to be VERY repeatable, but they don't agree very well, so the absolute velocity is suspect. As long as you are VERY careful to align them, and can live with some absolute velocity uncertainty, they are still pretty nice. Some people claim they lose velocity when the batteries aren't fresh, but I haven't explored that.

Besides some additional testing with a couple of my most stable air pistols (fortunately, I have about 50 to chose from...), I may get a Nate just to use as my "gold standard" to evaluate the others I've got.

I'm somewhat confident in saying I've most likely shot through more chronographs than any person - because I live shoot calibrate every unit that is made.
I cannot stress enough how critical mounting is - it has taken many, MANY iterations, design updates and testing to get even a simple mount/test rig that works well to the accuracy i demand.
Something like a combro would benefit from a large/high-grip surface area with strong bands to ensure it doesnt move. Target pistol should be ok - maybe a strip of rubber to make sure between the unit & barrel for grip.

This is only for muzzle mounted units - although if a chrony is on a table which is moving/wobbling, the same will most likely apply. Careful on things like blast from air blast or powder burners.
I've been able to demonstrate something like a vibration will introduce a ~5-10fps error @ 1,000fps, larger sensor spacing will be much more

 

[Gwhite]

I've been tied up with other stuff, but I went through my notes and found a couple of promising pistols that I rebuilt several months ago. One of them is a Hammerli 480K that was upgraded to the use the Hammerli AP40 regulator system and cylinders. I typically only tested the velocity for 10 shots, which I've concluded may not be enough to get really solid results. I have several pistols that showed variations on the order of +/- 1 FPS when measured with a Combro. I measured one of those with the HT-X3006, and here are the results:

The pistol started out acting very erratically. I think this is from sitting around for about 3 months. I've seen several pistols that calm down after about 15 shots, and this one behaved that way. After shot #16, the velocity over the last 12 shots varied only slightly more than +/- 1/2 m/sec, which is quite good for a PCP competition air pistol. The last 10 shots I did with a Combro when it was freshly rebuilt (and had been fired a fair amount in the process) had an average velocity of 145.6 m/sec, vs 144.7 m/sec for the last 12 shots with the HT-X3006.

I'd like to have a bit more data like this before I decide to use the HT-X3006 as my primary chronograph for tuning, but at this point, it's looks promising. My earlier tests with a more erratic pistol made it difficult to draw any conclusions, so I view this as a step forward.

In the meantime, LabRadar tech support has gotten back to be with some questions & suggestions. At some point I may drag that back to the 50 foot range at the college and try that again. As the weather improves, I could also try it at an outdoor range with even fewer obstacles nearby to confuse it.

 

[NateChrony]

Radars...
There was an interesting thread on AGF UK about radars, an old guy that worked on radars for 60 years popped in - he basically laughed at using a radar on something the speed, size & shape of a pellet for accurate readings.
Given it is pretty clear its possible to do it - I think he has a very high bar for accuracy & consistency, hence the statement. But some interesting insight from an industry expert in the technology.

That said - with airgun pellets (slugs not so much) they decelerate a good amount, one of my customers can notice a 2fps difference with/without their moderator on. (less so at lower velocities - which you're at with pistols)
Due to this - the acquisition time of the pellet is critical, when did the radar actually measure the pellet speed?
If it acquired the speeds at different distances pellet to pellet - then you'll get different readings even if the pistol was bang on.

Some units will show the acquisition measurements over time/distance to assist with this.
The Garmin unit "looks" to be the most reliable - notably it does a lot of "processing" between shots, I think its doing a LOT of smart things with the data to help with things like the issues above.

I looked into doing a radar unit - but given the challenges above and costs, its just not something I see as a good solution - especially for airgunners.

 

[Gwhite]

While going over the data in the previous plot, I noticed that the velocities tended to only take on certain values. That suggested that the chronograph had limited resolution, and that the results are "quantized." I did the original measurements in FPS, and I took that data, and sorted it by ascending velocity. I then subtracted the velocities offset by one position. That gives the velocity changes. Here's a plot of the results:

There are a few intermediate values as the differences get bigger, but there are NO adjacent velocities that change by less than 0.3 FPS, and a lot of them change by exactly that amount. There are also a couple that change by twice that.

My suspicion is that the sampling rate is designed to produce 0.1 M/sec velocity resolution, and the 0.3 FPS quantization results from the internal conversion from metric, which is then truncated to 1 decimal place. It's not a big deal as a practical matter. If I had done the tests in M/sec, I think it only reads out to 0.1 m/sec. The main display shows 2 decimal places, but if you review a list of shots, it only shows 0.1 M/sec resolution.

 

[NateChrony]

yeah the uncertainty - covered that in this video a little:

I've looked internally at a few of these & Its most likely due to the lack of processing power for the given the sensor distance. You can increase the distance between sensors, or increase the processing power, both have draw backs.

 

[tibor]

Athlon crony..

 

[Gwhite]

My circuit design skills are out of date, but to save power, I wouldn't expect them to try to do the actual time interval measurement in a "processor." A dedicated CMOS interval counter could be clocked at a high rate without having to run everything else that fast. The HT-X3006 sensors are 5.5 cm apart. The specs claim it will measure velocities up to 1219 M/sec (which is nonsense as far as air guns go, as Nate's video points out) to 0.5% accuracy. At 1219 M/sec, it takes ~ 45 microseconds to go 5.5 cm, and you need to measure that to 1 part in 200. That says the minimum clock for the counter is ~ 4.4 MHz if everything else is perfect (like the sensor spacing). That's actually pretty slow by modern circuitry standards. A 10 MHz timer clock would give you more than a factor of two margin. It actually looks like there are "ultra low power" microcontrollers out there that can clock at many 10's of MHz, so maybe the processor can handle it.

Realistically, I think the precision of the sensor spacing is more critical if they don't want to have to calibrate each one. 0.5% of 5.5 cm means the sensors need to be soldered in place to better than around a quarter of a mm. I think a good fixture should be able to do that. I suppose at some point, the accuracy of the location of the semiconductor chip relative to the package would be the limiting factor.