[brycelarson]

Egomaniacal, on 30 June 2011 - 03:48 PM, said:

That makes a lot more sense. The first spike would be the hammer and paintball being pushed forward (rig pushed back), downward spike is hitting the valve (pushing forward), second upward spike is the paintball actually accelerating (pushing the rig back), the small downward motion is the bolt being accelerated back (rig pushed forward) - is it attenuated by continued gas expansion behind the paintball? Then the final upward spike is the bolt hitting the boltstop (pushing rig back).

Looks like the human element is less than I thought it'd be. Looks nice guys, I'm looking forward to more results!

yup, exactly the way I read it. One note - we know from previous testing that the Ion cycles exactly the same way with and without paint - so the bolt return cycle isn't effected in this case by the back pressure from the gun. Your statement is prob accurate on the other guns.

paintballjla, on 30 June 2011 - 10:12 PM, said:

So you could take a reading from a different spot/directional configuration just on at the same time? I think that this would be cool just to see other directions, and how much it changes if mounted different places. And how hard would the problem of only one input be so solve is that one input into the digital oscilloscope or collection software?

correct - we have a single input o-scope. The sensors are cheap the multi input o-scope is really expensive. With this rig we will just take various measurements at various locations at various angles and lay them on top of each other.

[Doyle]

If you guys dont like the way they're doing the tests, nothing is stopping you from doing them yourselves. They aren't being paid to do these tests

While yes a multi directional setup would be idle, clearly budget restricts that. What they can do however is test the forward/backwards as previously done, then perhaps check barrel rise by doing a seperate test.

I'd like to see if the graph changes at all during rapid fire.

Keep up the good work, very exciting stuff here

brycelarson, on 01 July 2011 - 06:50 AM, said:

the multi input o-scope is really expensive.

How expensive? Perhaps another episode of DrunkWorks is in order

cockerpunk, on 30 June 2011 - 09:41 AM, said:

you are failing to discount that i am indeed, the batman. just sayin'

Does that make Bryce Robin?

[Doyle]

paintballjla, on 30 June 2011 - 02:05 PM, said:

And on the note of borrowing guns i think that it will be important to know exactly what you are getting data form, because yes it is easy to tell what physical changes have been made but with so may guns you have a whole list of things that you can change in the board that could affect how much the guns kick like dwell ect. I sujest that you request that the guns broad be reset to factory setting for a data set so a fair comparison can be made between guns.

Two indentical guns with the same settings wont shoot exactly the same. While it would be beneficial to list pressures, and settings for comparison between two of the SAME gun, the differences in settings for comparison to a different type of gun are minor. There is a lot of fine tuning that could be done using these sensors.

A good example where you will see differences is in a gun such as an Ego where the regs aren't set properly... the higher the LPR setting the more force that will be placed on the bolt.

Seems like they will be testing a lot of different guns for quite awhile

[Egomaniacal]

brycelarson, on 01 July 2011 - 06:50 AM, said:

yup, exactly the way I read it. One note - we know from previous testing that the Ion cycles exactly the same way with and without paint - so the bolt return cycle isn't effected in this case by the back pressure from the gun. Your statement is prob accurate on the other guns.

Well that's curious.

brycelarson, on 01 July 2011 - 06:50 AM, said:

correct - we have a single input o-scope. The sensors are cheap the multi input o-scope is really expensive. With this rig we will just take various measurements at various locations at various angles and lay them on top of each other.

You could cheat and use a DAQ instead of an o-scope. I've used this guy before:
http://sine.ni.com/n...g/en/nid/201986
and there are even cheaper knockoffs that come with software available on ebay. What software are you using now?

[paintballjla]

Doyle, on 01 July 2011 - 12:33 PM, said:

If you guys dont like the way they're doing the tests, nothing is stopping you from doing them yourselves. They aren't being paid to do these tests

While yes a multi directional setup would be idle, clearly budget restricts that. What they can do however is test the forward/backwards as previously done, then perhaps check barrel rise by doing a seperate test.

I'd like to see if the graph changes at all during rapid fire.

Keep up the good work, very exciting stuff here

brycelarson, on 01 July 2011 - 06:50 AM, said:

the multi input o-scope is really expensive.

How expensive? Perhaps another episode of DrunkWorks is in order

The prices range depending on specs but some can run 1000+ but depending on exact needs one may be found in the hundreds hundred. Im not sure of the specs of the one they are using now but it really completly depends on what you want/need them to do.

[y0da900]

Sorry Gordon, I don't have any thoughts for filtering after the data has been collected. I have run into similar issues with data collected at work from pressure transducers and flow meters, and it has always gone back to changing settings regarding how and what data is actually captured. Lots of that type of software will have some sort of filtering options built in, but it may not be a part of whatever package you are using. If you run into issues handling that much data, I have had much better luck analyzing it with Open Office than with Excel - it seems to behave much better with large data sets than Excel does.

But what I meant was when you generate the graphs, do not let it automatically determine your axis range. Force all graphs to generate with a time axis of 0 to 450ms (or whatever number ends up capturing all data of interest) and an acceleration axis of -n to n (again, whatever value captures the necessary data for all tests).

If you're interested, I can whip up a barrel bracket that will let you rigidly mount it repeatedly at the same point in every gun. A cocker barrel to cocker barrel adapter if you will, with a swiveling center section (to orient the axes with each gun) having a little extra meat on it that you can mount the board to and lock in place with a set screw.

[brycelarson]

Yoda - that would be really cool.

[Spitlebug]

Surprising. All said and done from shot to shot the "kick" was very similar. Not much difference at all and looking at that, having a clamped rig would probably not add anything to the test.

[y0da900]

What bore size do you want for the adapter? I can make it large enough diameter to take a cut down freak insert, but I don't think I have any boring tools long enough to make it take a full length insert. Have dimensions you can toss at me from an ifit kit or similar?

And a part number for the accelerometer so I can make sure I put enough room on there for it.

[cockerpunk]

im never a fan of giving up data (ie filtering before recording), there is no reason you can't capture everything and filter it later if you know how to building a filter (me and lurker were talking about a differentiation filter). the problem then i picking your corner frequency on your filter (basically slope). idk, i'll try things out next week when i have some time.

[Troy]

cockerpunk, on 03 July 2011 - 10:55 AM, said:

im never a fan of giving up data (ie filtering before recording), there is no reason you can't capture everything and filter it later if you know how to building a filter (me and lurker were talking about a differentiation filter). the problem then i picking your corner frequency on your filter (basically slope). idk, i'll try things out next week when i have some time.

For those of us that aren't quite as mathematically proficient as you guys are, what kind of filtering are you looking to do exactly?

I know that excel is awesome and all, but it seems to me that processing large amounts of data is always easier with a database and some code.

[brycelarson]

Were not actually looking at that much data for this test. The software were using is only grabbing 1000 data points. The shot itself is only two or three hundred samples long.

[Egomaniacal]

cockerpunk, on 03 July 2011 - 10:55 AM, said:

im never a fan of giving up data (ie filtering before recording), there is no reason you can't capture everything and filter it later if you know how to building a filter (me and lurker were talking about a differentiation filter). the problem then i picking your corner frequency on your filter (basically slope). idk, i'll try things out next week when i have some time.

If the filter is low-pass you won't lose any information that's relevant to the test. All measurable effects I'm seeing there are well below 500 Hz, and it looks like your sample rate is what, 1ksps? I don't see any line hum, so that's good, but It's probably below the sensitivity of your scope or buried in all that other crap.

There is a reason you don't want to filter after you've digitized your signal. Before it goes in the 'scope, it's an analog signal so it has all frequencies, from 0 to infinity. When you digitize that signal, every one of those contributions to the signal gets stuffed into that data point. But because your sampling rate is less than infinite, you can't distinguish between signal and noise at a higher frequency. You won't be able to get rid of any noise that's at a higher frequency than half your sampling rate. And even eliminating anything close to that frequency is difficult.

As a rule of thumb, when it comes to noise the earlier you can nip it the better. And with a properly built filter, you won't be losing any information. The other thing you could try is hooking your sensor up in differential configuration instead of single-ended (just an assumption on my part that it's not already).

If I'm guessing right at what a "differentiation filter" is, I don't think you'll have much luck because the "slope" for the noise is probably less than some of the signal itself, and greater than any details you're trying to get out of the noise. All said and done it is a pretty strong signal.

What sensor are you using (model # & MFR) - I might poke around the spec sheets to see what's up.

This post has been edited by Egomaniacal: 03 July 2011 - 11:22 PM

[Kirko017]

lol yay now the vs threads are going to be completely full of pointless arguments about kick.... because we all know paintball markers have so much dramatic kick that we have to argue about it. biggest fail this testing is going to cause.

This post has been edited by Kirko017: 04 July 2011 - 04:35 AM

[The_Economist]

Kirko017, on 04 July 2011 - 04:34 AM, said:

lol yay now the vs threads are going to be completely full of pointless arguments about kick.... because we all know paintball markers have so much dramatic kick that we have to argue about it. biggest fail this testing is going to cause.

Ummm they weren't full of threads about kick already? If any two people are bullshitting about guns, all they can compare anymore is efficiency, kick, ergonomics, and looks. Everything else is equal between two guns in the same price range. Efficiency can't really be argued since it is usually measured by Mike. Ergonomics and looks are completely subjective and therefore can't be argued in any rational sense. All that's left is kick, and people ask "Which gun is smoothest?" about every thirty seconds on this forum. At least this testing can provide some hard numbers for "smoothness" comparisons that were previously based completely on perceptions.

[cockerpunk]

Egomaniacal, on 03 July 2011 - 11:18 PM, said:

cockerpunk, on 03 July 2011 - 10:55 AM, said:

im never a fan of giving up data (ie filtering before recording), there is no reason you can't capture everything and filter it later if you know how to building a filter (me and lurker were talking about a differentiation filter). the problem then i picking your corner frequency on your filter (basically slope). idk, i'll try things out next week when i have some time.

If the filter is low-pass you won't lose any information that's relevant to the test. All measurable effects I'm seeing there are well below 500 Hz, and it looks like your sample rate is what, 1ksps? I don't see any line hum, so that's good, but It's probably below the sensitivity of your scope or buried in all that other crap.

There is a reason you don't want to filter after you've digitized your signal. Before it goes in the 'scope, it's an analog signal so it has all frequencies, from 0 to infinity. When you digitize that signal, every one of those contributions to the signal gets stuffed into that data point. But because your sampling rate is less than infinite, you can't distinguish between signal and noise at a higher frequency. You won't be able to get rid of any noise that's at a higher frequency than half your sampling rate. And even eliminating anything close to that frequency is difficult.

As a rule of thumb, when it comes to noise the earlier you can nip it the better. And with a properly built filter, you won't be losing any information. The other thing you could try is hooking your sensor up in differential configuration instead of single-ended (just an assumption on my part that it's not already).

If I'm guessing right at what a "differentiation filter" is, I don't think you'll have much luck because the "slope" for the noise is probably less than some of the signal itself, and greater than any details you're trying to get out of the noise. All said and done it is a pretty strong signal.

What sensor are you using (model # & MFR) - I might poke around the spec sheets to see what's up.

great info, ill see what i can do. thanks!

[Egomaniacal]

If your scope has an FFT function I'd love to see an output with just the sensor plugged in and not moving.

[cockerpunk]

Egomaniacal, on 05 July 2011 - 05:03 PM, said:

If your scope has an FFT function I'd love to see an output with just the sensor plugged in and not moving.

easy to do an FFT after the fact too, ill look into that as well. been busy at work ...

[brycelarson]

Egomaniacal, on 05 July 2011 - 05:03 PM, said:

If your scope has an FFT function I'd love to see an output with just the sensor plugged in and not moving.

we took some readings locked down before the testing for Gordo to start looking at the baseline.

you can download the software, take a look and tell me what you want me to record.

http://www.linkinstr...s.com/mso19.htm

it does have an FFT setting.

[Egomaniacal]

Am I completely misreading this in that you have a much higher bandwidth than you're using? Are you limiting the sample rate to make it manageable in excel?

I would think a screengrab of an FFT over say, 2.5kHz should be plenty. Just for shits and giggles you could do something bigger, too. It's running in demo mode for me since I don't have hardware so I can't quite see what the trigger/averaging is doing, but you could average it over some number of samples you think is sufficient to see any noise sources.

You've probably already realized that if you ran it at a higher sampling rate (I would think 10 ksps would be sufficient) you could probably do your filtering after you digitize your signal.

I don't suppose you have access to MatLab