|
|
Post by jetjeff on Oct 25, 2015 15:26:15 GMT -5
Hello,
I machined a shaft using chrome moly tubing for the center section. The ends (threaded portion) are a press fit. Is press fitting enough to secure the ends? My only options at this point it weld it, or install roll pins or solid pins. Williams International does weld the ends to the hollow center tube. The Heineken mini keg is not the actual outer case, just a decorative sleeve.
|
|
|
|
Post by jetjeff on Oct 25, 2015 15:27:40 GMT -5
|
|
|
|
Post by smithy1 on Oct 25, 2015 17:26:38 GMT -5
What size is your turbine?? I assume you're talking about the shaft "tunnel"?? If so I doesn't need to be steel, most of my micro turbine shaft tunnels are aluminium and are bolted or screwed to each end. Do you have any photos of the offending item?
Cheers,
Smithy.
|
|
|
|
Post by jetjeff on Oct 25, 2015 17:45:49 GMT -5
|
|
|
|
Post by jetjeff on Oct 25, 2015 17:47:17 GMT -5
It's not the tunnel, it's the actual shaft.
Jeff
|
|
|
|
Post by smithy1 on Oct 26, 2015 21:41:08 GMT -5
It's not the tunnel, it the actual shaft. Jeff Fair enough...When you said "Tubing" you had me a bit confused...I wouldn't be using tubing for the main shaft either. Depending on the rotational direction of your impeller and turbine, for CW rotating impeller, as viewed from the front, you should generally run "left hand" threads on both ends of the shaft. Welding or "Pinning" of the shaft is not necessary, some of the micro turbine manufacturers use a small interference fit but normally they're a nice snug sliding fit, the "self tightening" nature of the nut is enough. Cheers, Smithy. |
|
|
|
Post by finiteparts on Oct 26, 2015 23:08:43 GMT -5
Jeff,
First things....that is a really nice shaft and I am curious why you decided to go this route verses just using a solid rod?
I think the chances of welding it without distorting it is pretty slim. I would bet that Williams uses electron beam or laser welding to control the heat affected zone and grain issues. Secondly, they probably have an in depth heat treatment process for getting the proper grain structure back after the weld has been completed. Fundamentally, you would want to avoid welds on high strength steel shafts due to stress concentration...especially near the ends were you are more likely to get bending (overhung rotors).
The use of tubing is a good idea if you are really optimizing for weight vs stiffness, but a big issue is that usually thin wall tubing is not held to exacting wall thicknesses. It can vary a few thousandths here and there, which makes it tough for high speed applications that require very good balancing.
So on the turbine end, fundamentally there are a few things to check. First, the thermal expansion from the "hot" turbine will help keep that end fixed, at least until it yields for the first time. If the tubing gets to a temperature where the yield stress drops below whatever stress is imposed on the joint due to rotation and/or bending (due to the overhung rotor arrangement), the tubing will yield. Then when you cool it down, you might not have an interference fit any more...or worse, you loose your tight fit while it is rotating and sling out a rotor. Slip fits are never a good idea if you are not totally sure of the thermal condition that the joint will be operating in...you have to have them controlled because the thermal stresses are so large.
I am not saying that the arrangement won't work, but you might want to do some design work to feel comfortable that all your hard work won't turn into a high speed spark generator. You need to understand the stress state the you put the joint in, due only to the press fit. Then you need to check the thermal expansion of each part in the joint. If the inner part is a different material than the tubing, there could be a difference in thermal expansion coefficients that could lead to very large stresses in the joint. If everything looks good for the running state, you need to check the soak-back condition at shutdown and make sure that you don't yield the tubing due to a high thermal soak into the inner part and not the tubing.
The compressor end needs a similar design look, but the temperatures are lower and so if you can make the turbine end joint work, then the compressor end should work too.
In general, I would say for the homebuilder that they would have a more dimensionally controlled shaft if they machined it out of a solid bar. This way, you know that all your centerlines stay true to each other and there are no "weak" spots in the shaft in case you go into a condition were you might run through some higher bending loads.
Good luck!
Chris
|
|
|
|
Post by jetjeff on Oct 27, 2015 0:58:42 GMT -5
Hmmm,,,interesting points to consider. The wall thickness of the moly tubing I used is .095, the Williams shaft is .055. I used tubing for it's weight savings. I agree welding would distort the trueness of the shaft, but somehow Williams was able to overcome this.  |
|
|
|
Post by racket on Oct 27, 2015 15:07:57 GMT -5
Hi Jef
Could you please supply some dimensions for the shaft , also the actual inference forces you used to assemble it .
Cheers
John
|
|
|
|
Post by jetjeff on Oct 28, 2015 3:25:21 GMT -5
John,
Length overall is 9 1/16". Tubing diameter is 5/8". Turbine end shaft is .312". Compressor end shaft is .250", and the bearing journals are 10mm. Interference fit was .003". The compressor (7.3 Powerstroke diesel) is somewhat unusual as the compressor is threaded at the end 1/4-28 thread (no nut needed at the end). Tubing wall thickness is .095". A dial indicator shows about .001" from true.
Tubing was used to reduce weight and decrease spooling time.
Jeff
|
|
|
|
Post by racket on Oct 28, 2015 15:45:24 GMT -5
Hi Jeff
How far into the tubing are the "ends" pressed ,.............whats the axial length of the inference fits ??
The longer the fit, the more force you needed to apply to press them in, and the greater the chance it will be sufficient to hold things in place .
Cheers
John
|
|
|
|
Post by jetjeff on Oct 28, 2015 16:37:50 GMT -5
3/4"
|
|
|
|
Post by racket on Oct 28, 2015 18:29:41 GMT -5
More than enough , if the tube is 0.625" OD with 0.095" WT , that gives a tube ID of ~0.435 , so ~1.7 times bore dia for the inference length , it would have taken quite some force to press in .
Assuming an air flow of 0.5 lbs/sec at a P2 of 30psi at 120,000 rpm , so ~50 HP to drive the comp , a torque of ~2 ft/lbs .............bugger all :-)
As long as the shaft is well balanced , I'd be happy to use it .
Cheers
John
|
|
|
|
Post by jetjeff on Oct 29, 2015 0:06:43 GMT -5
John / Chris,
Thanks for your input, I appreciate it!!
I wonder if filling the tube partially with Sodium would have aided heat transfer, the way high performance car exhaust valves are?
Jeff
|
|
