Thoughts on my piston valve design
- mrfoo
- Specialist 2
- Posts: 273
- Joined: Wed Apr 29, 2015 11:05 am
- Has thanked: 80 times
- Been thanked: 89 times
Surely that would mean that as soon as your chamber rises over atmospheric, your piston is going to start moving, with full pressure on the entire area of its sealing face. Your previous design opens once pressure rises over a "step" pressure defined by the pilot pressure and the differences between the two piston diameters, and with far less force.essentially sealing the chamber
Oh, no, scratch that. I'm an idiot. You're still gonna need more pressure in the pilot for a given piston release pressure, though, as the ratio of forces acting on the piston are different.
- Moonbogg
- Staff Sergeant 3
- Posts: 1733
- Joined: Mon Oct 13, 2008 10:20 pm
- Location: SoCal
- Has thanked: 158 times
- Been thanked: 100 times
Mr. Foo pointed it out, the entire front face of the piston is exposed to chamber pressure and the cannon won't work well that way. It opens too easy. You need the small sealing step on the front side of the piston to cause the pressure to build before opening. Once the valve opens even a little, then the entire front face is exposed to chamber pressure and it pops open from that point.stup86 wrote:There's something about the orientation of the piston that isn't sitting right with me.
I feel that having a design where the piston moves vertically, essentially sealing the chamber rather than the barrel would be better. As a picture is worth a 1000 words, here we go:
What's the general consensus of the 2 designs featured?
Regards
Stuart
Your design looks like a traditional safety valve that you see used on ships and other things with steam pressure and air pressure that needs safety relief valves. They are mostly all vertical like yours, so I get the appeal to go that route, but they don't allow the huge unbalanced surface areas on the front and back that we need for our purposed. I suppose you could design it into the new configuration though. It might get a bit tricky and expensive perhaps. I haven't thought about it. Give it a good think and see what you come up.
In terms of performance differences I'm not imagining there being any big differences.
-
- Private 2
- Posts: 20
- Joined: Sun Jun 09, 2019 5:29 pm
- Location: Bristol
- Has thanked: 2 times
- Been thanked: 2 times
Well, I'll just park the idea and file it under "stupid".Moonbogg wrote:Mr. Foo pointed it out, the entire front face of the piston is exposed to chamber pressure and the cannon won't work well that way. It opens too easy. You need the small sealing step on the front side of the piston to cause the pressure to build before opening. Once the valve opens even a little, then the entire front face is exposed to chamber pressure and it pops open from that point.stup86 wrote:There's something about the orientation of the piston that isn't sitting right with me.
I feel that having a design where the piston moves vertically, essentially sealing the chamber rather than the barrel would be better. As a picture is worth a 1000 words, here we go:
What's the general consensus of the 2 designs featured?
Regards
Stuart
Your design looks like a traditional safety valve that you see used on ships and other things with steam pressure and air pressure that needs safety relief valves. They are mostly all vertical like yours, so I get the appeal to go that route, but they don't allow the huge unbalanced surface areas on the front and back that we need for our purposed. I suppose you could design it into the new configuration though. It might get a bit tricky and expensive perhaps. I haven't thought about it. Give it a good think and see what you come up.
In terms of performance differences I'm not imagining there being any big differences.
I had some parts arrive today to build "Iteration 1" - it's a burst disc cannon to experiment with whilst the piston valve is being designed and made.
Here's a pic of it quickly put together (chamber only). Still need to use thread sealant and get some electrics sorted:
Regards
Stuart
- Moonbogg
- Staff Sergeant 3
- Posts: 1733
- Joined: Mon Oct 13, 2008 10:20 pm
- Location: SoCal
- Has thanked: 158 times
- Been thanked: 100 times
Nothing stupid about doing thought experiments. At least half the fun is designing stuff. Actually, for me I'd have to say designing the thing is more than half the fun, lol. Maybe like 70% fun designing and 30% fun shooting the thing.stup86 wrote:Well, I'll just park the idea and file it under "stupid".Moonbogg wrote:Mr. Foo pointed it out, the entire front face of the piston is exposed to chamber pressure and the cannon won't work well that way. It opens too easy. You need the small sealing step on the front side of the piston to cause the pressure to build before opening. Once the valve opens even a little, then the entire front face is exposed to chamber pressure and it pops open from that point.stup86 wrote:There's something about the orientation of the piston that isn't sitting right with me.
I feel that having a design where the piston moves vertically, essentially sealing the chamber rather than the barrel would be better. As a picture is worth a 1000 words, here we go:
What's the general consensus of the 2 designs featured?
Regards
Stuart
Your design looks like a traditional safety valve that you see used on ships and other things with steam pressure and air pressure that needs safety relief valves. They are mostly all vertical like yours, so I get the appeal to go that route, but they don't allow the huge unbalanced surface areas on the front and back that we need for our purposed. I suppose you could design it into the new configuration though. It might get a bit tricky and expensive perhaps. I haven't thought about it. Give it a good think and see what you come up.
In terms of performance differences I'm not imagining there being any big differences.
I had some parts arrive today to build "Iteration 1" - it's a burst disc cannon to experiment with whilst the piston valve is being designed and made.
Here's a pic of it quickly put together (chamber only). Still need to use thread sealant and get some electrics sorted:
Regards
Stuart
- jackssmirkingrevenge
- Five Star General
- Posts: 26203
- Joined: Thu Mar 15, 2007 11:28 pm
- Has thanked: 569 times
- Been thanked: 343 times
Smoother gas flow should result in better performance, however usually it's not something terribly significant.stup86 wrote:Also, I think I'm going to change the tee used in my design to swept tee as shown below
hectmarr wrote:You have to make many weapons, because this field is long and short life
-
- Private 2
- Posts: 20
- Joined: Sun Jun 09, 2019 5:29 pm
- Location: Bristol
- Has thanked: 2 times
- Been thanked: 2 times
So, as I'm trying to get the piston assembly weight down to as little as possible, I've started to (try to) analyse the part in SolidWorks. My company used to use SolidWorks before moving to Catia, so I only have access to a basic version - meaning SimulationXpress is all I can run.
I HAVE NEVER PERFORMED ANY STRESS ANALYSIS OF ANY KIND, so the stuff I'm presented with upon completion of the simulation doesn't mean alot.
I only pay attention to the safety factor and then look at the highly stressed areas of the von Mises.
One thing, I'm not sure of though is part of the setup of the simulation and this is the fixture of the part. The example image within SolidWorks shows the following:
However, this doesn't really help as the piston assembly ins't fixed to anything...
Anyway, ignoring the fixture that I used, I inputted the following pressures on each side of the piston:
For anyone that isn't aware how this part fits into the overall assembly, please see image below for (hopefully) some clarity:
I used 1200psi on the combustion side as the max combustion pressure that I could get on HGDT for a 10x propane mix was around 1150 psi. The 150psi of the pilot side is a complete guess, I can't imagine I would need to go that high but I see it's better to overestimate that underestimate.
Based on these figures I got a factor of safety of 4.4. But I'd like to ensure I get the fixture point as accurate as I can.
Any suggestions? Also, what factor of safety would you aim for?
Currently I don't plan on going too much above 10x due to the pressure rating of the components I have bought (I'm currently at 75% of the max design pressure, 300% of the max working pressure! )
Thanks, as always,
Stuart
I HAVE NEVER PERFORMED ANY STRESS ANALYSIS OF ANY KIND, so the stuff I'm presented with upon completion of the simulation doesn't mean alot.
I only pay attention to the safety factor and then look at the highly stressed areas of the von Mises.
One thing, I'm not sure of though is part of the setup of the simulation and this is the fixture of the part. The example image within SolidWorks shows the following:
However, this doesn't really help as the piston assembly ins't fixed to anything...
Anyway, ignoring the fixture that I used, I inputted the following pressures on each side of the piston:
For anyone that isn't aware how this part fits into the overall assembly, please see image below for (hopefully) some clarity:
I used 1200psi on the combustion side as the max combustion pressure that I could get on HGDT for a 10x propane mix was around 1150 psi. The 150psi of the pilot side is a complete guess, I can't imagine I would need to go that high but I see it's better to overestimate that underestimate.
Based on these figures I got a factor of safety of 4.4. But I'd like to ensure I get the fixture point as accurate as I can.
Any suggestions? Also, what factor of safety would you aim for?
Currently I don't plan on going too much above 10x due to the pressure rating of the components I have bought (I'm currently at 75% of the max design pressure, 300% of the max working pressure! )
Thanks, as always,
Stuart
- jackssmirkingrevenge
- Five Star General
- Posts: 26203
- Joined: Thu Mar 15, 2007 11:28 pm
- Has thanked: 569 times
- Been thanked: 343 times
That looks about right. One good thing about hybrids is that the pressure spike is just that, not a constant pressure in the chamber that you would get from a high pressure pneumatic launcher.
Yours truly certainly falls into the former category at times, with such brilliant ideas as using a thin alloy tube from a printer cartridge as a hybrid chamber with predictable results certainly not something I'd recommend to anyone else.
One thing to consider about your design, maybe an o-ring isn't the best idea to seal the barrel, in high flow situations exacerbated by the hot gasses it might blow out with every shot.
You'll get many different answers. Some people don't mind working on the edge of the material specifications and operate on the assumption that something will probably give, while others won't touch anything without a 3 or 4 times safety margin.Also, what factor of safety would you aim for?
Yours truly certainly falls into the former category at times, with such brilliant ideas as using a thin alloy tube from a printer cartridge as a hybrid chamber with predictable results certainly not something I'd recommend to anyone else.
One thing to consider about your design, maybe an o-ring isn't the best idea to seal the barrel, in high flow situations exacerbated by the hot gasses it might blow out with every shot.
Last edited by jackssmirkingrevenge on Wed Jun 19, 2019 2:59 am, edited 1 time in total.
hectmarr wrote:You have to make many weapons, because this field is long and short life
- Moonbogg
- Staff Sergeant 3
- Posts: 1733
- Joined: Mon Oct 13, 2008 10:20 pm
- Location: SoCal
- Has thanked: 158 times
- Been thanked: 100 times
I would make the bumper fill out the volume more like I show with my incredible paint skills. I'd use a medium hardness bumper and see how it does. Another concern I have is the front of your piston looks like it's going to slide inside the steel pipe and rub against the walls. Maybe not a big deal, but I'd try to isolate the aluminum from ever touching the pipe, at least for the wear surface that slides back and forth. You can machine a step and use a wear ring. They are used in pneumatic and hydraulic cylinders and they install on the piston. They serve as the wear surface. The piston will slide over a large opening, so I'm not sure if the wear ring will spring outward and try to fall off at that point though. Just something to think about.
I tend to overthink things though and I find myself hyper analyzing the small details, although I'm glad I did that for my last project because it actually works, so there's that to consider. I'd never let metal slide along metal.
Also, consider the possibility that if you don't vent the pilot fast enough and if the air compresses in the pilot upon firing, you will experience piston bounce. It will slam back into the port opening and you have no front bumper or sealing face to cushion the blow. The benefit of your front o-ring design is it should prevent leakage before the opening pressure is reached, but the downside is you have nothing to cushion the front of the piston, so you must ensure it opens and stays open without bouncing back.
Also, JSR pointed out that the o-ring might fly off. I don't know if it will or not. I didn't know if my sealing face was going to stay on or not either. Luckily it works, but every cannon of this type is a bit of an experiment really since not many people have built them. Hopefully your design makes a contribution to our body of knowledge
Regarding the solidworks thing, you just have to pick a fixture point and deal with the unrealistic condition that they introduce. Use creative ways to try and make the simulation as realistic as possible. Watch the animation that shows the part deforming to make sure it deforms as expected. It's a limitation, but without spending $12,000 on the full simulation software, us plebs are stuck with simulation Xpress. Make sure your materials are selected and have the correct yield strength and other material properties are correct.
The higher the safety factor, the better. If it can be made stronger for a little extra cost, I'd do it. Basic consumer goods are often designed with a 2:1 safety factor, but where risk to people's safety are concerned, often times the safety factor can be as high as 10:1. Based on my best efforts to simulate my design, my safety factor is about 3.5 in the worst case and at the weakest point, but I honestly don't know how good that is. It's just my best effort and since I'm not an engineer with formal training and experience, I get what I get and I take my own risks.
I hydro tested my chamber to just past 450psi and the max pressure should be about 330psi. That doesn't mean it's safe. It just means it didn't break during that pressure test. I decided to take the risk and use the cannon. There's always a risk with these. At least you aren't using PVC. We've seen what often happens with hybrid PVC cannons.
I tend to overthink things though and I find myself hyper analyzing the small details, although I'm glad I did that for my last project because it actually works, so there's that to consider. I'd never let metal slide along metal.
Also, consider the possibility that if you don't vent the pilot fast enough and if the air compresses in the pilot upon firing, you will experience piston bounce. It will slam back into the port opening and you have no front bumper or sealing face to cushion the blow. The benefit of your front o-ring design is it should prevent leakage before the opening pressure is reached, but the downside is you have nothing to cushion the front of the piston, so you must ensure it opens and stays open without bouncing back.
Also, JSR pointed out that the o-ring might fly off. I don't know if it will or not. I didn't know if my sealing face was going to stay on or not either. Luckily it works, but every cannon of this type is a bit of an experiment really since not many people have built them. Hopefully your design makes a contribution to our body of knowledge
Regarding the solidworks thing, you just have to pick a fixture point and deal with the unrealistic condition that they introduce. Use creative ways to try and make the simulation as realistic as possible. Watch the animation that shows the part deforming to make sure it deforms as expected. It's a limitation, but without spending $12,000 on the full simulation software, us plebs are stuck with simulation Xpress. Make sure your materials are selected and have the correct yield strength and other material properties are correct.
The higher the safety factor, the better. If it can be made stronger for a little extra cost, I'd do it. Basic consumer goods are often designed with a 2:1 safety factor, but where risk to people's safety are concerned, often times the safety factor can be as high as 10:1. Based on my best efforts to simulate my design, my safety factor is about 3.5 in the worst case and at the weakest point, but I honestly don't know how good that is. It's just my best effort and since I'm not an engineer with formal training and experience, I get what I get and I take my own risks.
I hydro tested my chamber to just past 450psi and the max pressure should be about 330psi. That doesn't mean it's safe. It just means it didn't break during that pressure test. I decided to take the risk and use the cannon. There's always a risk with these. At least you aren't using PVC. We've seen what often happens with hybrid PVC cannons.
-
- Private 2
- Posts: 20
- Joined: Sun Jun 09, 2019 5:29 pm
- Location: Bristol
- Has thanked: 2 times
- Been thanked: 2 times
Thanks for your answer’s guys.
I changed the sealing method at the front (left on image) of the piston to a face seal – current seal thickness is 8mm and just some rubber sheet that I found lying around at work. It contacts a “bush” that will be pressed into and welded to the steel pipe with an area just shy of 740mm2– hopefully this will work.
I removed the reduced diameter section at the rear (right on image) of the piston to increase the contact area with the bumper (2020mm2), whilst also increasing the spacing of the 2 o-rings
The bumper has also been increased to completely fill area behind the position whilst limiting the total travel to 15mm (30% of barrel ID) – I’m hopefully that this reduced throw will decrease the stress on the rear of the cannon. The small piston at the rear that seals the pilot area requires 11.4mm of travel to expose the o-ring and vent the volume. Because this piston is just threaded on the rod, I can adjust this if I get any issues with piston bounce.
All the changes made to the piston have equated to a SF4.5 and the images from SolidWorks showing von Mises and displacement are below:
So, thanks for now, until I change it again, just because.
Regards
Stuart
I can see the appeal, from a weight perspective, in getting as close to a safety factor (SF) of 1, but having some extra in the bank allows me to up the pressure in the future, if I ever feel inclined. Based on that, I’m aiming to get as close to SF4 as I can - the guy in red on my left shoulder doesn’t really care about this, but it’ll keep the guy in white on my right shoulder a little happier!Some people don't mind working on the edge of the material specifications and operate on the assumption that something will probably give, while others won't touch anything without a 3 or 4 times safety margin.
One thing to consider about your design, maybe an o-ring isn't the best idea to seal the barrel, in high flow situations exacerbated by the hot gasses it might blow out with every shot.
I had my doubts about how successful an o-ring seal was going to be. Actually, it was one the main reasons for creating this thread, as it’s not a design that I’ve seen. So as a result, I have altered the design (yet again) to address some your concerns.Also, JSR pointed out that the o-ring might fly off. I don't know if it will or not. I didn't know if my sealing face was going to stay on or not either.
I changed the sealing method at the front (left on image) of the piston to a face seal – current seal thickness is 8mm and just some rubber sheet that I found lying around at work. It contacts a “bush” that will be pressed into and welded to the steel pipe with an area just shy of 740mm2– hopefully this will work.
I removed the reduced diameter section at the rear (right on image) of the piston to increase the contact area with the bumper (2020mm2), whilst also increasing the spacing of the 2 o-rings
The bumper has also been increased to completely fill area behind the position whilst limiting the total travel to 15mm (30% of barrel ID) – I’m hopefully that this reduced throw will decrease the stress on the rear of the cannon. The small piston at the rear that seals the pilot area requires 11.4mm of travel to expose the o-ring and vent the volume. Because this piston is just threaded on the rod, I can adjust this if I get any issues with piston bounce.
All the changes made to the piston have equated to a SF4.5 and the images from SolidWorks showing von Mises and displacement are below:
So, thanks for now, until I change it again, just because.
Regards
Stuart
- Moonbogg
- Staff Sergeant 3
- Posts: 1733
- Joined: Mon Oct 13, 2008 10:20 pm
- Location: SoCal
- Has thanked: 158 times
- Been thanked: 100 times
That's looking pretty good. I'd reduce the OD of the bumper by just a little to allow for it to crush and compress. Right now it's filling the volume before compression. I'd make the sealing face 50 to 100% thicker and get that washer to be nearly as large as the ID of the barrel port. It will prevent the sealing face from getting ripped off. I'd make that washer as thick and as strong as you can manage and I'd even use a larger bolt to secure it. If you are tapping into aluminum make sure to use as large of threads as you can and make sure they are coarse threads. I'd suggest using anti-seize compound on all threads into aluminum as well. If it galls, you're screwed. It's looking pretty good.
One last thing. If it were me, and again I'm picky and paranoid, but I'd try to increase the distance between the two o-rings on the piston to make it as stable as possible during movement. There is something called the sticky-drawer effect. It's basically when a hockey puck shaped cylinder tilts inside a tube and gets stuck. Yours looks like it shouldn't happen, but if it was my cannon I'd just try to add a little more space for as much stability as possible. If it cocks to the side, it will be aluminum rubbing on steel.
My piston's o-rings aren't very far apart either, but yours has a heavy front that cantilevers quite a bit and I'm imagining it possibly trying to tilt the piston. I don't know, I'm paranoid like that. I'd make it as stable as you can without increasing the weight too much.
One last thing. If it were me, and again I'm picky and paranoid, but I'd try to increase the distance between the two o-rings on the piston to make it as stable as possible during movement. There is something called the sticky-drawer effect. It's basically when a hockey puck shaped cylinder tilts inside a tube and gets stuck. Yours looks like it shouldn't happen, but if it was my cannon I'd just try to add a little more space for as much stability as possible. If it cocks to the side, it will be aluminum rubbing on steel.
My piston's o-rings aren't very far apart either, but yours has a heavy front that cantilevers quite a bit and I'm imagining it possibly trying to tilt the piston. I don't know, I'm paranoid like that. I'd make it as stable as you can without increasing the weight too much.
-
- Private 2
- Posts: 20
- Joined: Sun Jun 09, 2019 5:29 pm
- Location: Bristol
- Has thanked: 2 times
- Been thanked: 2 times
Thanks Moonbogg.
Updates to the following:
- Bumper OD reduced. There is now radial clearance of 1.5mm (Annoyingly, I didn't mean to completely fill the gap!)
- Thickness of front seal increased form 8mm to 15mm (I'll need to find something suitable to make this from)
- Screw securing front seal has increased from M6 to M10 (same size the threaded rod) and is now countersunk
- Distance between o-rings centres has increased slightly. Now 21mm.
As always, here's the pictures:
I re-ran the SolidWorks simulation as before. SF = 4.11
Regards
Stuart
Updates to the following:
- Bumper OD reduced. There is now radial clearance of 1.5mm (Annoyingly, I didn't mean to completely fill the gap!)
- Thickness of front seal increased form 8mm to 15mm (I'll need to find something suitable to make this from)
- Screw securing front seal has increased from M6 to M10 (same size the threaded rod) and is now countersunk
- Distance between o-rings centres has increased slightly. Now 21mm.
As always, here's the pictures:
I re-ran the SolidWorks simulation as before. SF = 4.11
Regards
Stuart
- Moonbogg
- Staff Sergeant 3
- Posts: 1733
- Joined: Mon Oct 13, 2008 10:20 pm
- Location: SoCal
- Has thanked: 158 times
- Been thanked: 100 times
Looking good! Look forward to seeing it in action! Consider trying out some "X-rings" instead of o-rings. They seal better at low pressure and offer less friction for dynamic applications like yours. I use them and I can push the piston back with a single finger.
- jackssmirkingrevenge
- Five Star General
- Posts: 26203
- Joined: Thu Mar 15, 2007 11:28 pm
- Has thanked: 569 times
- Been thanked: 343 times
Agreed. What's the general plan for this? Fixed test-bed? Hand-held launcher?Moonbogg wrote:Looking good!
What about the caliber and what do you plan on launching?
hectmarr wrote:You have to make many weapons, because this field is long and short life
- Moonbogg
- Staff Sergeant 3
- Posts: 1733
- Joined: Mon Oct 13, 2008 10:20 pm
- Location: SoCal
- Has thanked: 158 times
- Been thanked: 100 times
I'll just say I don't know how the flat head screw will do for securing the rubber. I noticed it right away but didn't want to appear as being too picky since it's not my cannon after all, but it's best to try and help I think.
It might screw in far and force the rubber to move outward on the angled surface of the screw head. If you stop at an appropriate amount of screwing, then the screw is just floating in the threads with nothing secure to mate with. The rubber might not be totally secure at that point either. Just a thought. Maybe not a big deal to everyone under the sun, but it bugs me. I avoided that issue with mine. I personally like the washer method better, preferably with a spacer in the center of the rubber hole so the screw has a solid spacer tube to mate up against solidly while compressing the rubber perhaps 0.25mm. So if the rubber is 50mm thick, the washer screws in to allow only 49.75mm of a space so the rubber is slightly compressed and secure. Just my take. That's how I did mine and it works. Not sure how others have done theirs.
It might screw in far and force the rubber to move outward on the angled surface of the screw head. If you stop at an appropriate amount of screwing, then the screw is just floating in the threads with nothing secure to mate with. The rubber might not be totally secure at that point either. Just a thought. Maybe not a big deal to everyone under the sun, but it bugs me. I avoided that issue with mine. I personally like the washer method better, preferably with a spacer in the center of the rubber hole so the screw has a solid spacer tube to mate up against solidly while compressing the rubber perhaps 0.25mm. So if the rubber is 50mm thick, the washer screws in to allow only 49.75mm of a space so the rubber is slightly compressed and secure. Just my take. That's how I did mine and it works. Not sure how others have done theirs.