Talk:Thompson coupling

Sliding surfaces

Good point User:Van helsing - cardan joints don't have sliding surfaces, but CV's do - the balls slide against the inner race when the input and output shafts are at different angles. It is removing this source of friction that is claimed to be a benefit, and is one of the critical aspects of the invention. This source of friction has been the subject of significant research, particularly in the development of grease for CV's. I reverted back some material, but left out the sliding surfaces for cardan joints. I reverted back the aspects of axial and radial loads because these are in fact significant improvements over the traditional CV joint. CV joints cannot cope with significant axial loading with rotation.

Your point about sales hype is a good one, but the features of the joint are unambiguous, and are the reason the joint has been developed. To leave these aspects out would be perverse. GrahamP 10:09, 12 February 2007 (UTC)

Thanks for your post. The sliding surface thing however makes me a bit curious: "the balls slide against the inner race when the input and output shafts are at different angles". I’m actually quite convinced that the balls in a Constant-velocity joint are supposed to roll in their grooves (like in a ball bearing). Apart from moving grease around, I don’t see a cause for (sliding) friction here (but always willing to learn of course). I agree to your axial and radial load statement; that is where the advantage lays when you need a constant velocity joint. --Van helsing 13:20, 13 February 2007 (UTC)

Thanks for the reply. In a standard roller bearing, the groove runs circumferentially, with the movement of the ball in the same direction as the groove, but in the CV joint, the grooves run axially, with the torque acting on the CV joint causing the balls to force on the sides of the grooves. The balls slide up and down the grooves only with differential angles between input and output shafts. This is where the friction comes into play. In theory (I think), if there was no differential angle, there would be (almost) no friction, but in that case, you wouldn't need to use a CV joint. GrahamP 21:53, 13 February 2007 (UTC)

Still, that would make it a rolling friction, not a sliding one (bit like a axial loaded standard ball bearing). Anyway, wrong or right, the claims are sourced, therefore okay to state it like that. --Van helsing 22:12, 13 February 2007 (UTC)

Hype

This is nothing more than a slight re-packaging of a double cardan joint. They have been used for many years in agricultural machinery. oh, and Panhard Dyna Z Greglocock 06:09, 17 June 2007 (UTC)

After stumbling across http://cvcoupling.com/index.php?option=com_content&task=view&id=4&Itemid=52 my hype detector is on overdrive. Check this out:

Like the Bendix-Weiss joint the balls in the Rzeppa joint skid in the grooves when the joint is operated at an angle and even with modern lubricants the wear rate and friction is high, such that the life of these joints is reckoned in minutes when used at any appreciable angle and torque setting.

Minutes, even. I reckon the age of the universe at 7.2e15 minutes and my faith in this article at 0.01 minutes. Ignoring the hype, it does state that great circle geometry has something to do with the fact that the balls don't slide, and that engineering great circle geometry posed challenges, which are left unspecified. If the reasoning behind either claim was known, it would improve the article. MaxEnt 03:07, 23 September 2007 (UTC)

Speedy deletion/advertising

In reference to Greglocock and MaxEnt comments above, they have failed to understand the operation of Thompson Coupling. I needed to post some details of the design earlier in 2007 in response to Van helsing's comments. Note that I have no particular interest in the coupling other than being an engineer with an appreciation of the novelty of the design in the context of traditional cardan joints, double cardan joints and CV joints, which are in widespread usage. The coupling is not a repackaged double cardan joint, which as Greglocock correctly points out, has been in widespread usage for many decades. To obtain constant velocity, the double cardan relies on the same axial angle between input and output shafts, which is one of the limitations that the Thompson Coupling overcomes. Similarly for CV joints, the limitation is the sliding (rather than rolling), creating friction. My earlier comments regarding the operation, together with the inventor's website should provide sufficient information for those interested. From an engineering perspective, I don't know how far I need to go into providing a under-graduate tutorial on universal joints in wiki talk pages, and I don't believe the ridicule by MaxEnt is helpful. If users dispute the novelty of the coupling, then it would be helpful to provide evidence contradicting the claims. I would recommend that any users wanting to contribute to the article familiarises themselves adequately with the operation and geometry of all types of universal joints before contributing, and possibly ridiculing those who do. The article does not contain sales hype, and is a legitimate article about an item from a commercial entity. GrahamP 01:07, 29 October 2007 (UTC)

So... how much power do you think is absorbed in a typical CV joint's (quasi mythical) sliding surfaces? I have seen reports on the efficiency of /real/ CV joints measured in /real/ operating conditions (as you can imagine, vehicle manufacturers are very interested in this), and even if that tiny amount of power was saved in this design, the difference would be minute in most applications. So far as its kinematics compared with a double cardan, are you sure? Have you actually checked it out? I'd agree that at high speeds, and high articulation angles, there are some second order effects in a DC, which is why they are not used so much under those conditions, and it is possible that a reconfiguration such as this may help in those circumstances. But the whole article reeks of hype. Greglocock 02:24, 30 October 2007 (UTC)

You've hit the nail on the head Greglocock - vehicle manufacturers are VERY interested in driveline losses. As you would not doubt be aware, power losses in a car with a manual gearbox are usually quoted at around 25-30%, most of which is consumed by your quasi mythical sliding surfaces of gear surfaces and bearing surfaces, which the second law of thermodynamics tells us ends up eventually as heat. I assume you've taken the sump plug off a gearbox or differential after it has been running for while - despite the best oil technology, sliding surfaces still cause heat. Friction losses from bearing shells in engines are also significant. If you don't believe me that CV joints have sliding surfaces, I invite you to purchase a second hand one from your local car wrecking yard and open it up. By memory, CV losses (brand new) are of the order of a few percent, and higher when the CV ages. The losses are significant for two reasons - firstly they contribute to a reduction in fuel economy (if we are talking about vehicles) and contribute heat, which needs to be disipated causing premature wear.Here's a good link I found describing CV joint efficiency [1]. Regarding double cardans, a differential angle between input and output results in an instantaneous velocity difference. For applications in which the angles are fixed, it can be set up to obtain near-constant velocity for design conditions - for applications, such as machinery with fixed or narrow angle differences, the double cardan may be the preferred option. The Thompson retains constant velocity over a range of angles. I invite you to do a search for patents with the term "constant velocity" and "universal joint" - there is a plethora of inventions to overcome the problems you claim either don't exist or have been solved decades ago. The Thompson Coupling is merely an incremental improvement for certain applications of universal joint technology, with the main points being that it retains true constant velocity for a range of angles, does it without axial sliding surfaces, and maintains axial strength. If you can find another commercial joint that does this, then start a new wiki page and I'll read it!GrahamP 01:07, 1 November 2007 (UTC)

That's a good link. Now, why do you think they use such an odd way of measuring the inefficiency? Surely it would be easier to measure the torque in, and the torque out, and the speed of each shaft? That, incidentally, is a /big/ clue. Greglocock 05:15, 1 November 2007 (UTC)

Thanks for your reply. I'd agree with you that the torque input/output method is equally valid. As I you probably know, the obvious problem with it, however, is that at low torque differentials, you will need highly accurate measurement of the absolute level of torque to obtain a statistically significant result. Say the loss is 2%, and the accuracy of the each motor/drive is 1%, then it's simply not going to be good enough - you'll end up with very low confidence levels, and invalid results. The benefit of measuring heat is that you're directly measuring the differential (ie: the loss). If you can measure at 10%, then your results will be only slightly worse than 10% accurate. By measuring the energy loss, you can work backwards from joules, watts, then torque to get a result. The other benefit is that you can get a time averaged level of power loss - say run the equipment for an hour. Anyway, easier is in the eyes of the beholder - a mechanical engineer or electrical engineer might like to work with motors but a physicist might be equally comfortable working with heat. Personally, if was being paid to do the test, I'd prefer to measure heat directly as I know that I'm directly measuring the parameter of interest. Providing I can measure one of the torques to within a few percent, I'll get a good quality result - the rotational velocity can be measured highly accurately. As a starting point, I'd assume that we're dealing with 2 to 5 % loss at an angle of 15 degrees . Hope this clears your points up. GrahamP 07:10, 1 November 2007 (UTC)

Now we get to the nub of it. Your 5% estimate is way high , even so a saving 8% of 5% is down in the weeds (0.4%), if you compare that figure with the range of your estimate (2-5%) then I think that supports the argument that this joint is being hyped mercilessly. We see the same thing with novel engine designs - for some reason the Australian media thinks that every backyard inventor can do what no large company has managed. The reason that temperature methods are used to measure losses in driveline systems is because the losses are so small, and an 8% saving of something small is not very much. As you can see measuring torques to an accuracy of a few % is no good enough in this context.Greglocock 22:45, 1 November 2007 (UTC)

Exactly! Personally, Greglocock, I couldn't care less whether the loss in a specific application is 0.001% or 50% or whether anybody uses them - the point is that they are an incremental improvement on existing designs and may have application in a range of areas. You may suggest that they will not be a commercial success, and that's fine! You might suggest that they have been overblown in the media, that's fine! You're not going to get an argument from me about overblown hype in the media. However, my view is that the wiki article provides a brief, factually correct overview of the coupling with a number of links for those interested to follow. Some will agree with you about their novelty, but most qualified engineers will not. For a small list of people who think they have potential, I refer you to Auto Engineer, April 2007; Factory Equipment News April, 2007; Mining Monthly Dec 2006; Australasian Science, Engineering World etc.... The Thompson Coupling isn't going to stop wars, end poverty and solve climate change, but nonetheless it is an interesting and novel solution to torque transmission that may have commercial success, and in my view, easily worthy of a wiki page. If you can find anything that is verifiably inaccurate in the article, or in my posts, I urge you to change/discuss it. GrahamP 23:41, 1 November 2007 (UTC)

Weird claim

"Unlike double cardan joints, the Thompson Coupling does not require equal angles between input and output shafts to maintain constant velocity. "

Not too sure what is being claimed here - a double cardan is a thing unto itself, it has an input shft and an output shft and no other point of reference. Therefore the term "equal angles" is meaningless. I am not 100% confident of this but am, say 90% confident. It looks to me as though someone is (probably deliberately) confusing this with a pair of Hookes joints in the usual setup.

Greglocock (talk) 22:45, 27 November 2008 (UTC)

Thanks for your comments Greglocock. I changed the wording to clarify the issue of angles, and also intermediate section. I take your point about the double cardan being a "thing unto itself". For the point of the page, the intermediate shaft is simply the centre section of the double cardan - it experiences the same instantaneous velocity difference as if there was a shaft connecting a pair of cardans - the sentences are trying to highlight that the intermediate piece (or shaft) always has a different instantaneous velocity, except when the input shaft, output shaft and joint are aligned along the same axis. The references are good - I think they should stay. GrahamP (talk) 21:04, 30 November 2008 (UTC)

No problem with references, big problem with the claim. Your revised wording is better but it still implies that a conventional DCJ is in some way difficult to set up to get 'equal angles', whereas from practical experience I can say that a DCJ based shaft is a factor of 10 easier to set up than a single Hookes joint based shaft, in both high torque and high articulation compound angle cases. The angles sort themselves out if you let the shaft lines intersect, not exactly a big ask. Greglocock (talk) 21:50, 30 November 2008 (UTC)

The latest revision 13:13 1 Dec is still no better overall. The sentence in question is meaningless, since equal angles are meaningless in terms of a joint with intersecting shafts. I am strongly tempted to just delete it, it adds nothing to the article, and repeated revisions have failed to improve it in respect of the main problem, which is that a double cardan jont has equal angles between the input and output shafts if they intersect, which is how normal people set them up. If the next revision doesn't fix it then I'm going to have at it. Greglocock (talk) 02:29, 1 December 2008 (UTC)

If the wording is unclear, I'm happy to change - can you recommend better wording? I know you're coming from an automotive standpoint, but the article isn't solely about drive shafts in cars. I agree that in some situations, the angles will be even, but not always - remember this article isn't about 4WD drivelines, and the article isn't supposed to be a beat-up on DC's! P.S Have you read the references thoroughly yet? GrahamP (talk) 02:51, 1 December 2008 (UTC)

I suggest deleting the entire sentence. Are you connected with Thompson Couplings at all? Greglocock (talk) 03:22, 1 December 2008 (UTC)

No, I'm not connected with Thompson Couplings. The sentence is factual - references provided meet reliable source guideline, and meets NPOV. Add, modify if you can improve, but provide reliable sources and maintain Wikipedia:Etiquette .GrahamP (talk) 09:57, 1 December 2008 (UTC)

OK, so which source says "Unlike double cardan joints, which provide near-constant-velocity [3], the Thompson Coupling does not require equal angles between the input shaft axis and the joint, and output shaft axis and the joint, to maintain constant velocity.". ? Doesn't seem to be Sutherland as he explicitly states the forces in a DC joint are different to those mentioned in his article. Doesn't seem to be Sheu he is talking about DC shafts not DC joints. Sopannen is also discussing DC shafts, although he confuses them with joints. So, which ref makes the claim and what does it actually say? My suspicion is that in every case it is someone confusing DC shafts and DC joints, most large DC joints use an internal centreing mechanism to prevent misalignment, and with small ones you merely make sure the shafts intersect. (Incidentally I think it was a perfectly legitimate question, asked politely)Greglocock (talk) 03:13, 2 December 2008 (UTC)

I recommend re-reading Sopanen, page 2, last paragraph, and page 3, first paragraph, particularly "They found that the relative phase angle is the most important variable affecting to the input-output displacement relationship of the double cardan shaft", and "Fischer and Paul concluded that the fluctuations in input and output displacements vanish when the joint angles of two joints are equal and the relative phase angle is set to zero. When the two joint angles are unequal some fluctuation will exist. However, the fluctuation is minimized when the relative phase angle is zero".
If you think the technical papers are wrong, fair enough - Wikipedia is not the forum to debate whether published technical papers are wrong, but Wikipedia requires references to meet reliable source guidelines.
I think it's fair enough if you want to include an additional sentence that says something about the fact that for many automotive applications, like 4WD's, the joint angles will be the same, which maintains the drive at near-constant-velocity. GrahamP (talk) 19:26, 2 December 2008 (UTC)

resetindent- So you agree that Sopanen is talking about double cardan /shafts/ not DC joints? If so then he cannot be used a reference for a discussion of joints, as Sutherland points out, the forces are different in joints (not something I actually agree with in the final analysis, but certainly the releative magnitude of the different forces is different). You need a reference that discusses the effect of mislaigned shafts on DC /joints/, you won't find one in any sensible literature because the solution is well known, as it has been for many years, - a centreing ball is standard fitment where the alignment cannot be guaranteed. We don't need additional sentences, we need to remove sentences that make bogus claims. Greglocock (talk) 21:35, 2 December 2008 (UTC)

Would you agree that double cardan can refer to one of three possibilities?
1) A pair of cardans connected by an intermediate shaft, without a centering mechanism.
2) A DC joint which has a centering mechanism.
3) A DC joint without a centering mechanism.
Would you agree that the reason for the centering mechanism is to maintain an equal angle between each of the input/output shafts and intermediate shaft/yolk in order to assume the optimum near-constant-velocity parameters (which I assume is the debate we are currently having)?
Would you agree that the reason you set up the joint to design specs is to acheive the same result - namely equal angles between in/out and yolk to ensure near-constant-velocity?
Are you assuming that all the discussions in the article and this discussion page so far are only referring to possibility #2?
As an aside, I had a quick look at the US patent website, which reveals many types of double cardan mechanisms, many of which discuss an improved centering mechanism to improve on the standard products already in the market. My view is that double cardan does not assume ipso facto a centering mechanism, although, as I gather you are arguing, in practice, most DC joints do incorporate a centering mechanism, to ensure optimum

operation. GrahamP (talk) 05:41, 5 December 2008 (UTC)

I haven't the faintest idea whether 'most' DC joints include a centreing mechanism. Anyway, I gather you now agree that those references are not discussing DC joints. Greglocock (talk) 00:20, 6 December 2008 (UTC)

Are we still debating the validity of the sentence that says "... does not require equal angles between the input shaft axis and the joint, and output shaft axis and the joint, to maintain constant velocity"?
The basic geometry of the double cardan is identical with or without the centering element, and is identical whether the intermediate coupling is a short shaft, long shaft or yolk (as in the case of the joint) - the references provide sufficient information to make this clear - no more references are required - the whole point of a centering element with a DC joint is to fix the differential angle problem.
Correct me if I'm wrong, but I gather you've have been assuming all along that the double cardan must include a centering element (while I have assumed the more general case), and this is why you've been challenging the validity of some aspects of the article?
What is your view of my previous post or are we close to reaching an agreement?GrahamP (talk) 19:02, 7 December 2008 (UTC)

No, the DC joints I deal with daily in my job do not have centreing elements. My objection to the sentence is that it exaggerates a problem that normal people solve in one of two ways, either by aligning the shafts so that they intersect, or failing that, fitting a centreing mechanism. So having invented a straw man problem, the Thompson Coupling is claimed to solve it. Adding references that confuse joints with shafts in this case is just wasting my time. Deleting the sentence in question is still my preferred course.Greglocock (talk) 21:57, 7 December 2008 (UTC)

If you have no objection to the technical accuracy of the article, but to an alleged WP:NPOV, then
1) Why didn't you say that to start with?????
2) Why not merge the article into universal joints, with a new section after double cardans, which already discusses the same issues as the problem sentence, and put the references into universal joints - maybe expand the entry to include other universal joints. Do you want to do this or do you want me to?
P.S: I can understand that your main interest is in cars, but I'm familiar with industrial automation and machinery where vibration and stress due to changing velocities of high inertia components is a genuine issue. You think that its a straw man for the automotive game - maybe you're correct - but there's plenty of industries and uses for uni-joints - ask Boeing [2]. GrahamP (talk) 09:25, 8 December 2008 (UTC)

What is the hard bit? The article makes a straw man claim about DCJs. The references aren't helpful. The sentence is not talking about UJs or proper CV joints, it is directly comparing DCJs and TCs. I shall add the cn tag back to the specific claim that none of your referneces make. Greglocock (talk) 09:41, 8 December 2008 (UTC)

My proposal is to get rid of the sentence you don't like and merge the main parts into UV's (the section on DC's already covers the characteristics of DC's). Do you have an alternative WP:NPOV proposal that we can both live with? GrahamP (talk) 02:52, 9 December 2008 (UTC)

What do you mean by UVs? guessing it is a typo for UJs, then that makes sense, I think.Greglocock (talk) 04:01, 9 December 2008 (UTC)

Yes, should have been UJ's - ok I'll do it in the next couple of days and leave it to you to comment.GrahamP (talk) 06:42, 9 December 2008 (UTC)