Torsion spring design manual , engineering reference

[ratbldr427]

I'm in the process of modifing some equipment to use some wound wire torsion springs instead of the flat torsion bars currently used. I understand the basics but need something explaining life cycle , fatigue, end mounting design and material.
What I am trying to come up with is something that will last 20million cycles in 2-3 years. They will fit over a 20mm tube with a total 80 in lbs torque over 40deg. I ordered some cheap McMaster Carr ones for testing but need more engineering data. I may have to make a test rig but don't want to test every combo available , reinvent the wheel so to speak.
The currant bars break at 1-1/2 years ,15 mil cycles. Engineering problem that can't be fixed. They last just past the warranty and the complete assy is 24k ea, 14 used in each machine .

 

[Dan from Oakland]

Call Century Spring or Associated Spring and ask to speak to an applications engineer.

 

[Freedommachine]

I have made hundreds of double torsion springs from 1080 music wire. The highest cycle count would be 3000ish though, nowhere near your requirement.

I agree with Dan, this is a problem for someone with a PhD in springs.

 

[Erich]

Second vote for let the application engineer at one of the spring houses at it.
Steel has an inifinite fatigue life, IF you keep the stresses low enough.
Low stress means more material volume. So, large wire diameter, lots of turns to get the spring rate down to the level you want. Basically, the space you have available will drive whether or not you can meet your design objective.

 

[gbent]

I come up with 19 hz with 100% duty cycle. I'm guessing you are getting some heating in the spring. Have you looked at gas springs?

 

[Erich]

Re work your numbers I get 0.21 cycles/sec.

 

[ratbldr427]

I come up with 19 hz with 100% duty cycle. I'm guessing you are getting some heating in the spring. Have you looked at gas springs? Thanks for all the replys everyone. gbent can't use gas springs. Cycle time at 17k sheets per hr is 1.5secs, 14 bars and each cycles twice per sheet.

The t/bars insert into those socket ends and they simply wear until they break off. They are installed with moly lube but no way to relube with out removing and disassembly. The picture shows a new end ,worn end and the early style wide bar that I have only seen one or two break in 20 years. The other wear point is the tube end but I solved that by flame spray a hard surface, 60RC on.
I bought tubing and form drills to make new tubes but since the end bushing wear is the only thing that wears that is fixed with the flame spray.

The complete bar on the bench has enough space to install 4 torsion springs around the 20mm tube externally. The t/bar is inside the tube, 2 bars that anchor to the bar in the middle and anchor to the frame at the ends. Due to the speeds this press runs the mfg made those bars as light as possible, reduced the tube dia. so there is no way to install a wider bar as you can see by the sockets. So that is why I am considering external torsion springs. The unit load on the springs is very low( about 18"lb ea x 4 springs) but the problem is the same as with the flat bars in that the end anchor if not done correctly then they would simply wear out from constant rubbing. But at least I can clamp them and maybe that will stop them from wearing out.

The mfg will not sell the t/bars separately, only a $ 6k kit that includes the tubes, bars and anchors and chain link pins(4 req'd per bar and are avail separately @ $900.ea!) The t/bars are the same material as what is used in the impression and transfer cyls. They are 60" long and are a stack of 4 ea which I can buy for 300 or so and cut in half to get 8 of these and machine the ends. I also make the sockets in my spare time out of scrap or O2 drill rod.
Since we got this press I have saved the co about $400k on the bars I have rebuilt.

 

[BugRobotics]

I've worked with Lee Spring quite a bit in the past with custom spring orders. Knowledgeable crew over there.

 

[Bill D]

You are saying it is not spring failure but rather lubrication failure causing the spring end to grind itself away and break off.
Sounds like you need to add a tube so the sockets can be greased from outside the machine.
Bill D

 

[ratbldr427]

You are saying it is not spring failure but rather lubrication failure causing the spring end to grind itself away and break off.
Sounds like you need to add a tube so the sockets can be greased from outside the machine.
Bill D

A little more than just lube. The narrow bar has a very high edge load and the only way to stop the erosion would be to clamp the ends so they don't rub but I don't see any way to do that. It is an engineering flaw and if you look at the pictures there simply is no room to do any thing and constant lube would be a non starter( don't think it would make any difference anyway) because the mess it would make on the printed sheets. It is not a problem on the early style wide bars in the picture and they are not lubed,I have only seen maybe 2 broke in 20years. The center anchor on the early bars has broken a few times because of a crappy weld.

We are getting another new press and I think they are going to have the early style bars installed. It is not a simple swap or we would have done that already.

 

[Bill D]

I did not mean constant lube. Just a way to add grease from time to time. But I supose the old grease has to go somewhere and cause a mess there.
Bill D

 

[Garwood]

Ok, so if the old bars had more contact area making for lower/manageable point loads, can you make the end cap/slotted thing longer thereby increasing contact area?

What about all those magic anti-wear coatings out there?

Make the ends out of 8620 and carburize it to 65rc then coat the spring ends in one of those crazy mega-hard surface treatments.

 

[ratbldr427]

Ok, so if the old bars had more contact area making for lower/manageable point loads, can you make the end cap/slotted thing longer thereby increasing contact area?

What about all those magic anti-wear coatings out there?

Make the ends out of 8620 and carburize it to 65rc then coat the spring ends in one of those crazy mega-hard surface treatments.

Garwood, the tube id is 14mm,the bars are 3mmx10mm and the connecting wall of the slot is minimal . The larger bars are open slots but every thing is simply beefier. Clamping would help but at least one end has to float to allow for length change. Again just no room, any increase in one area takes from another.
I did make some end anchors out of 16mm 12.9 Allen head bolts. Don't know how hard they are but it was a bitch milling the 3mm x 10mm slot in them.

Torsion springs solves the length change due the nature of the design so clamping the anchor ends would solve the rub wear.
Lube is a waste because the bars are preloaded, there is never a relax for clearance to allow it to recirculate. Dosn't work like a rotator. After a few minutes running the lube is displaced between the wear surfaces and even though there is lube all around the bar none is between the wear surfaces.
If I had room a Nyloil insert might solve the problem.
The spray powder I used on the tubes is 60RC but I think the heat would ruin the bars although there many different ones out there.
I think the torsion springs are the best bet all things considered and if that dosn't work then I just go back to working on my pet project , a flying pig.