Might have blown this part, can I safely use it?

[verbotenwhisky]

Can't disagree with you.

That being said, I need to have a probe that I can replace the o-rings somehow. I had thought that using an alignment pin and a taper tool would work. Think it still will, if I get the dimensions correct. I'm trying to determine what they might be. I'm a bit reticent designing for high pressure air, so I'm trying to balance getting it done and having it safe.

I have done FEA before, so I have some basic ideas what to look out for, but have no formal mechanical training. That being said, I like having all my body parts, so I will always take the less risky approach. Intellectually it's an interesting problem - it's making me think, which is safer than just blindly doing something.

May have to create a simplified model and noodle on it to understand the basics. Once I do that, I may simulate it to see or validate what I thought. Can't sim the whole model, there's something in it that's preventing meshing. Probably need to simplify it.

It is likely within the safety margin but I wouldn't use it if it were my pink hide at steak...

We play with cross sectional values and material strengths all the time but all it takes at this size is a burr, all materials specs have a tolerance as do dimensional values, if they stack up in the wrong direction it can make for a very bad day.

 

[WobblyHand]

Unless I can prove it is safe, with an acceptable margin, it's getting destroyed. Since I don't know the material, I can't determine the margin. So I will destroy the piece. Stinks, but that's what I need to do.

This still leaves me with the issue of determining the safety of the same part I made from 1144, using these prints. It has a 2mm blind hole. I will try to analyze that part. It may not be safe either, but I used it twice. I'd like to determine if I can have any blind hole in the location shown. The stock part has no blind hole on the left. I'll ponder that a while.

 

[piper184]

Forgive me if I misunderstand:
The purpose of the blind hole is to facilitate a device to aid in the installation of the O-ring?

Perhaps it is time to back up and take a broader view of the end objective. There must be another solution.

 

[verbotenwhisky]

Unless I can prove it is safe, with an acceptable margin, it's getting destroyed. Since I don't know the material, I can't determine the margin. So I will destroy the piece. Stinks, but that's what I need to do.

This still leaves me with the issue of determining the safety of the same part I made from 1144, using these prints. It has a 2mm blind hole. I will try to analyze that part. It may not be safe either, but I used it twice. I'd like to determine if I can have any blind hole in the location shown. The stock part has no blind hole on the left. I'll ponder that a while.

Do you have access to SolidWorks? if you do it should have cosmos, you can model it assign materials and run FEA.

 

[Ischgl99]

I think your biggest concern with this is blowing the end off, you have to assume the 5,000 psi is pressing against the o-ring, which is pressing against the wall. So, sticking with metric, you have 875 N of force acting on that, using the dimensions for your designed 2mm hole gives you a stress of 85MPa in that section, and for the 2.5mm hole, 102 MPa. Yield stress of 304 stainless is 204 MPa, so you have about a 2x factor of safety, but that’s not taking into account any stress risers from sharp corners in the o-ring groove that could increase the stress on that section. 1144 has a yield of 620 MPa, so that would be much safer. I would find a way to eliminate the blind hole and not worry about it. If it’s just for getting the o-rings on, maybe adding a bit of taper on the end instead of rounding the corner. Would making the lefthand side longer work to integrate the taper?

 

[WobblyHand]

Do you have access to SolidWorks? if you do it should have cosmos, you can model it assign materials and run FEA.

No I don't. But I do have FreeCAD and have done FEA with it. Unfortunately, as is, the model is having difficulty meshing. Getting weird errors. May need to simplify the model so that it can sim the stuff that I am interested in, not the whole thing. I haven't done much of that before. I did slice the model in half, but there's features still in there that are driving the mesher mad.

 

[verbotenwhisky]

No I don't. But I do have FreeCAD and have done FEA with it. Unfortunately, as is, the model is having difficulty meshing. Getting weird errors. May need to simplify the model so that it can sim the stuff that I am interested in, not the whole thing. I haven't done much of that before. I did slice the model in half, but there's features still in there that are driving the mesher mad.

Yeah that is why I like Cosmos, it will mesh easy with Solidworks model. Nastran works pretty well also.

 

[WobblyHand]

I think your biggest concern with this is blowing the end off, you have to assume the 5,000 psi is pressing against the o-ring, which is pressing against the wall. So, sticking with metric, you have 875 N of force acting on that, using the dimensions for your designed 2mm hole gives you a stress of 85MPa in that section, and for the 2.5mm hole, 102 MPa. Yield stress of 304 stainless is 204 MPa, so you have about a 2x factor of safety, but that’s not taking into account any stress risers from sharp corners in the o-ring groove that could increase the stress on that section. 1144 has a yield of 620 MPa, so that would be much safer. I would find a way to eliminate the blind hole and not worry about it. If it’s just for getting the o-rings on, maybe adding a bit of taper on the end instead of rounding the corner. Would making the lefthand side longer work to integrate the taper?

Think the stock part with the 2.5mm hole is a lost cause unless I basically plug the hole again. Since I don't know the material, it's hard to know how to proceed, save for plugging perhaps.

The part I made with 1144 and the 2mm hole, I think is ok. I'd feel a lot better if I could sim it. I may take my model, make a copy of it, and cut off the right hand side. Basically just simulate the first 10mm of left hand side, while under appropriate pressure.

Below are some pictures and description of the 1144 probe I made.

Yes, the blind hole is simply to allow for a taper tool to register to the part, to install o-rings. Without the taper tool, it's a 5 ring circus trying to install the small, hard, lubricated o-rings. Stretching a lubed 1.5x4 NBR90 o-ring from 4mm to 7mm and getting it over the probe without the tool is not possible for me. I spent a day trying. With the tool, the o-rings were installed in about a minute. I could make the pin shorter, say 2-3mm, or thinner. I have some 1.6x8mm pins available.

Here's the taper tool. 4mm section, tapering to 7mm over 25mm, with 15mm ends. The ends were so I could grab it in a chuck at either end. I hadn't pressed in the 2mm pin yet.

Butted up to the probe, with the o-rings successfully installed. I put an o-ring on the 4mm end, then pushed it to the taper. Then I installed the drill chuck to hold the taper and press it against the probe, which is being held by an 8mm collet. Getting the o-rings up the taper was not that easy to do even with the tool, the taper was about 6.9 degrees, and those NBR90's fought me nearly all the way. Once at the grooves they just jumped in.

Registration pin picture with the blind hole in the probe.

That pin in the taper tool looks big in this picture. It's only 2mm, or 0.079" in diameter, and 4mm (0.1575") long. It could be smaller in diameter and shorter, but I feel something like it is necessary. I suppose one could install the pin in the probe, but having a pin stick out seems like it's waiting to be damaged. The taper tool would only be used for installation of o-rings. Plus installing the pin would require a weep hole to allow the compressed air to escape. I did drill that weep hole on the taper tool, but it's not visible in the picture. I wouldn't want to install a weep hole in the probe.

 

[WobblyHand]

Yeah that is why I like Cosmos, it will mesh easy with Solidworks model. Nastran works pretty well also.

FreeCAD comes with a built in free FEA tool. It's not that hard to use, if you can get the part to mesh properly. I'm a little bit further hampered since not all the tools seem to be easily installable on a Mac. I was able to install gmsh, but haven't figured out how to install netgen, which is an alternate mesher. Nonetheless, I have been able to do both FEA and CFD on my machine.

 

[verbotenwhisky]

FreeCAD comes with a built in free FEA tool. It's not that hard to use, if you can get the part to mesh properly. I'm a little bit further hampered since not all the tools seem to be easily installable on a Mac. I was able to install gmsh, but haven't figured out how to install netgen, which is an alternate mesher. Nonetheless, I have been able to do both FEA and CFD on my machine.

Cool.... I do very heavy weight stuff on this machine with extreme file sizes but it is a beast, its a laptop that is bigger than most desk tops.