Trying to add pressure sensors to LDC testing.

[WhatUPSbox?]

Over in the 3D printing gate and elsewhere there have been some pretty intricate LDC designs made. For my education, I wanted to look at the very basic features to see what effect they have. I previously posted a brief description of my approach https://www.gatewaytoairguns.org/GTA/index.php?topic=212304.msg156510848#msg156510848 Basically an LDC made from segments that can be mixed/matched and some audio instrumentation. I've added some ports for pressure sensors to see if adding that data helps in figuring out what drives what.

Warning, this is heading off into the technical weeds and may not come out with anything useful.

The first image shows the basic building blocks of the test LDC. There is a mount to the barrel that has a laser gate used for timing and a pressure port right at the exit of the barrel to try to get the pressure behind the pellet. There is a short section that is a little larger than the pellet so the gas can't expand too much right after the pellet skirt exits the muzzle. Once you enter the LDC there are options for segments with cones or just hollow volumes. Some have pressure ports. At the other end is a cap which can have either a radial or an axial pressure port.

The hardware is shown in the second photo. It is mounted to a 2240 with a 1377 barrel. The 2240 has a power adjuster but is otherwise stock.

The third image shows an example of a test shot. The pellet is exiting the muzzle at time = 0 when the laser gate trips. In this case the pressure sensors were placed in the first segment and radially in the cap. They basically provide a time trace of the static pressure at those locations.

Note the pellet spends a little under 1 milli-sec in the LDC (starting at 0)

In the fourth image, one of the pressure sensors is moved from the first segment to the location right at the muzzle.

Not sure if this deep in the weeds is of interest or if any conclusions will come of it but it keeps my brain busy. Doing the pressure measurements with a hobby budget is tricky (for example I could not use my 4 channel DSO because of noise floor), so I have a bunch more checkout to do and the curves in the images are meant to be examples. Hope to combine this with some of the audio response measurements in the other thread.
 

[rkr]

That's one part of silencer design, dropping the pressure peak at the end cap exit which usually means longer duration for the peek. One easy way of doing that is to use bigger hole in the cap. From noise perspective it's only part of the story, you also need to look at the noise made by the pressure blast hitting the walls of your silencer.

[WhatUPSbox?]

Yes, part of the motivation for looking at the pressure was that in my microphone only testing the noise spike was starting before pellet exit from the LDC. The hope was to start separating those two effects.

[WobblyHand]

Yes, part of the motivation for looking at the pressure was that in my microphone only testing the noise spike was starting before pellet exit from the LDC. The hope was to start separating those two effects.

Stan, you are doing a nice job at this!  Fascinating.  I hesitate to say this, but could be interesting to apply an accelerometer or similar sensor to detect vibration of the LDC cylindrical body, sort of a ping detector.  I'd bet it really rings.  To reduce the cylinder ring, I'd imagine one would need to increase mass or stiffness, by construction or material choice, or significantly increase damping.  TPU might do this to some extent.  That's why I was printing TPU LDC's.  A lead loaded rubber material would probably work as well.

Had a crazy idea for an LDC material, but no one seems to make the material, which is tungsten loaded TPU.  I did search for quite a while.  Even asked Prusa if they knew of any sources.  They didn't, and told me they wouldn't be manufacturing tungsten loaded TPU either.  I'm sure someone would or could do it, but it would be custom.  Would increase cylinder damping a lot.  I have found tungsten loaded PETG, but it's expensive, and stiff.  The density is about 4gm/cc vs the ordinary density of 1.27 gm/cc for PETG.

[WhatUPSbox?]

I hesitate to say this, but could be interesting to apply an accelerometer or similar sensor to detect vibration of the LDC cylindrical body, sort of a ping detector.  I'd bet it really rings. 

Yes, the hope is to integrate some of the other sensors into the set-up. If you look at the segment with the port for the pressure probe, there is a flat spot on the opposite side for either an accel or at least a contact mike. One thing I need to check is how sensitive the pressure sensors are to vibration. On one of the test shots I clipped the front cap and the sensor had a response.

The tricky part for me has been the data acquisition side. The pressure sensors have a sensitivity of about 3mV/psia. Unfortunately my 4 channel Rigol DSO only goes down to 10mV/div and is noisy at that setting. I don't want to spend the money (even surplus) for a lab instrumentation amplifier. I tried the AD620 based modules on Amazon but they had their own problems. I may try an authentic AD620 at some point. For now, I have a humble 2 channel hobby DSO that you see in the picture. It does have some useful features. It is battery powered and relatively low noise with a 5mV/div sensitivity and claims 12 bit. It also accepts an external trigger that I can trip with the laser gate. It will store and export 4 pairs of waveforms (1024 points). That's where the Excel plots came from.
As I add the mic/accel channels, those can go to the Rigol and share the trigger. I just have to make sure I avoid the EMI issues. I'm not an EE so this moves more slowly.

[WobblyHand]

I hesitate to say this, but could be interesting to apply an accelerometer or similar sensor to detect vibration of the LDC cylindrical body, sort of a ping detector.  I'd bet it really rings. 

Yes, the hope is to integrate some of the other sensors into the set-up. If you look at the segment with the port for the pressure probe, there is a flat spot on the opposite side for either an accel or at least a contact mike. One thing I need to check is how sensitive the pressure sensors are to vibration. On one of the test shots I clipped the front cap and the sensor had a response.

The tricky part for me has been the data acquisition side. The pressure sensors have a sensitivity of about 3mV/psia. Unfortunately my 4 channel Rigol DSO only goes down to 10mV/div and is noisy at that setting. I don't want to spend the money (even surplus) for a lab instrumentation amplifier. I tried the AD620 based modules on Amazon but they had their own problems. I may try an authentic AD620 at some point. For now, I have a humble 2 channel hobby DSO that you see in the picture. It does have some useful features. It is battery powered and relatively low noise with a 5mV/div sensitivity and claims 12 bit. It also accepts an external trigger that I can trip with the laser gate. It will store and export 4 pairs of waveforms (1024 points). That's where the Excel plots came from.
As I add the mic/accel channels, those can go to the Rigol and share the trigger. I just have to make sure I avoid the EMI issues. I'm not an EE so this moves more slowly.

Cheap decent data acquisition can be an issue.  I have a two channel 200 MHz Siglent O-scope with a trigger channel that I bought a while back.  It's great for the cost, but one always needs more channels.  For multi-channel work, it might be worth rigging something up.  Have to look around to see if there's any decent stuff out there.  I'd think if all the channels were operating at 300-500 KHz it would be more than good enough.  The question is always - how many are needed.

If you have circuit questions, let me know.  I was a practicing EE for my career.  Wasn't a full time analog circuit designer, but made more than my share of radar front ends for the automotive market.  Quite a bit of cost pressure, but the job still needed to be done.  We detected and tracked picowatt returns, including babies, strollers, children's toys, as well as vehicles.  Anyways, it's an offer.

[WhatUPSbox?]

Bruce, thank you for the kind offer. I may bounce a circuit concept or two off of you. Right now I'm trying to work through some EMI shielding/mitigation options. For the pressure measurements, I'm using battery power for everything (DSO, laser, and bridge excitation through a linear regulator). I use twisted pairs where possible and the noise is driven by the DSO and is a little over an equivalent psi p-p. When I add the acoustic sensors and the Rigol DSO, it will share the trigger circuit and I expect the noise to go up.

[rkr]

To dampen the ringing on silencer shell a piece of bike inner tube or shrink tube on the outside is normally used. Neoprene covers also work if tight. In addition felt on the inside against the tube slows down the air before it hits the shell.

[WhatUPSbox?]

So I already failed my first EE101 pop quiz.   

People should ignore the specific curves I included in the first post (though the barrel exit one is probably close).

I do need to add amplifiers between the sensors and the scope to provide isolation. I was OK when I was testing with just one sensor but with two (and the laser circuit), there is coupling with the bridge circuits because of the common ground.

I'll repost when fixed

[WobblyHand]

So I already failed my first EE101 pop quiz.   

People should ignore the specific curves I included in the first post (though the barrel exit one is probably close).

I do need to add amplifiers between the sensors and the scope to provide isolation. I was OK when I was testing with just one sensor but with two (and the laser circuit), there is coupling with the bridge circuits because of the common ground.

I'll repost when fixed

Sorry to hear that.  It happens, especially if looking at small signals and big ones at the same time.  I call it the "learning experience", aka school of hard knocks. 

My radar chronograph front end was like that - kicked my butt for a while, until I got it.  Was a seven stage active filter, and it took me a long time to design and debug, especially since the simulator just didn't take into account some real world things, like leakage currents.  Actually it was the device models that didn't have those effects included.  So you sort have got to learn as you go along. 

Might want to consider differential amplifiers since they tend to be more independent of ground.  You'll get there because you're motivated.

What specifically are you having an issue with?  Got a schematic?  Or some chicken scratchings?

[WhatUPSbox?]

Bruce, Here is a schematic of my current (problematic) set-up. If I test with just one sensor and trigger off of its signal rather than the laser, I think I'm OK. It is just a differential measurement across the bridge. Once I add the second sensor and/or the laser module, the common in the DSO ties it to the bridge and the output is affected.

I am looking at instrumentation amplifiers. Either the Analog Devices AD62X series or the AD8221. That will turn the bridge outputs into single ended inputs into the DSO.
Of course the right answer is to use the Endevco 136 amplifier but even surplus they are out of hobby budgets.

I'm not quite sure if and where I should attach the cable shield that is in the Endevco cables. 

[WobblyHand]

Stan, yes, not a good idea to common those bridge circuits.  An instrumentation amplifier would be best for this, although I'd think using an op amp as a differential amplifier would work as well.  But to make it work well, you'd need to match the resistors carefully.  Even then, an instrumentation amp would be much better.  How much gain do you need?  What kind of bandwidth?  1 MHz ok?

Here's https://www.analog.com/media/en/technical-documentation/data-sheets/LT6370.pdf which might be suitable if you can get it in a decent package.  Made for bridge applications, I think.  Maybe $6 each, I'd guess.  But something like this might work.

This might be better.  https://www.analog.com/media/en/technical-documentation/data-sheets/ad8422.pdf.

For all of them, check the gain bandwidth product to see if you can get both the gain and the bandwidth you need.  If GBW=1MHz, and you ask for a gain=2, you will only get 500KHz of bandwidth.  So for BW=1MHz and a desired gain of 10, the GBW has to be greater than 10MHz.

[WhatUPSbox?]

Bruce, thank you for the recommendations. I'll have to decide on packaging. The AD620 and AD626 I can find in DIP. I do have a few sop8 breakout boards so the AD8422 is an option.

[WobblyHand]

Bruce, thank you for the recommendations. I'll have to decide on packaging. The AD620 and AD626 I can find in DIP. I do have a few sop8 breakout boards so the AD8422 is an option.

I'm finding that getting DIP packaging for any modern parts is impossible.  I've even had to resort to using DFN packages which have no leads and the pads are on the bottom of the part.  For prototyping I had to buy assortments of breakout boards, for DFN-8, SOT-6, SOP8, SSOP8, etc.  I simply couldn't procure the parts in the larger packages, they are no longer being made.  I use a hot air gun to solder the SMT parts to the boards.  Definitely harder than it used to be - in the old days.  Now I have to use a microscope to solder parts.

[WhatUPSbox?]

Bruce, any thoughts on the AD8226 compared to the AD8422? Slightly slower bandwidth but the lower cost makes for extra soldering spares

[WobblyHand]

Bruce, any thoughts on the AD8226 compared to the AD8422? Slightly slower bandwidth but the lower cost makes for extra soldering spares

If it was me, I'd pick the AD8422.  At G=10, you would get nearly 850KHz BW, which significantly greater than the AD8226.  So you'd be able to get near 1.2us rise times with the faster part vs, 6.25us or so?  Depends on your measurement goals.  I'm assuming a gain of 10 will be adequate for the instrumentation noise to be greater than your scope noise, but don't actually know what you need.  (You want roughly 10dB more front end noise than the scope.)  The more gain you need, the more smeared your output will be.  So if you are looking for better rise times, then choose the larger gain bandwidth part, ie, the AD8422.  It will allow you more gain should you need it for your scope.

[WhatUPSbox?]

Thanks for pointing that out I was comparing the G=1 values and didn't revisit the G= 10. I'll probably get a couple of each and compare. Though, I'm not sure what I will do with pressure data above 10-20 Khz.

[Motorhead]

Tho do take into account .... Equal caliber PCP's operating with the same pellet/slug at similar velocity .... will have a vastly different sound signature IF & WHEN they differ in valve dwell 

The longer dwell PCP & in turn the Less efficient will be louder period !!  Put more air into the LDC raising the internal pressure post shot, relative to the shorter dwell PCP.

[WobblyHand]

Thanks for pointing that out I was comparing the G=1 values and didn't revisit the G= 10. I'll probably get a couple of each and compare. Though, I'm not sure what I will do with pressure data above 10-20 Khz.

What's the bandwidth of the pressure sensor?  By that I mean, how fast does it response to a pulse of pressure, not it's pressure range.  There's not much point of amplifying stuff beyond that bandwidth because the sensor won't respond.

[WhatUPSbox?]

The spec for a somewhat newer version of the same model shows a resonant frequency (f_n) of 500 Khz. The endevco doc shows a response curve with 4% amplification at f_n/5. So, I would say 100Khz and a rise time consistent with that (2.5 micro-sec according to their doc). In the data I was capturing with the handheld DSO a useful shot cycle duration was 10-20 milli-sec and I was getting 1024 samples. Of course one can capture a shorter duration but you start loosing context and the DSO runs out of sampling rate. With amplification, I may be able to go back to the 4 channel Rigol with a much higher sampling rate. So I don't expect the sensor or the amplifier to be the limiting factor. 10-20 Khz would give me plenty of info. The pellet is moving at about 6000 inches/sec.