Fixing Tooth Brake to Shaft

[mrpi]

Hello,

I'm working on a design that uses a worm-type gear reduction unit. The output "shaft" of the gear reducer is in fact a 25mm hollow shaft provided with am 8mm keyway.

I use a 25mm keyed shaft placed through the worm reducer hollow shaft to transmit the output torque.

A little about the mechanism:

One end of the 25mm keyed shaft is coupled to a ball-screw actuator with a keyed clamp shaft coupler.

The other end of the 25mm keyed shaft is provided with an overhung shaft brake.

The shaft brake is a spring applied tooted-type brake. The stationary "stator" part of the brake is bolted to the output flange of the worm reducer and the 25mm keyed shaft passes through the center of it. Internal springs push a ring of brass teeth axially against the rotating half of the brake.

The rotating "rotor" half of the tooth brake is a steel ring of teeth that is axially fixed to the end of the 25mm keyed shaft with an aluminum plate.

When power is applied, an electromagnet pulls the brass stator teeth back away from the rotor teeth, and the brake is disengaged. When power is removed, the internal springs push the stator teeth against the rotor teeth and the brake is engaged. Rated torque of the brake is on the order of 100 Nm.

Currently, I couple the rotor to the 25mm shaft with 6 axial pins arranged in a "bolt-circle" like bullets in a revolver. A central M8 bolt holds the rotor plate axially against the shaft. These pins are 3/16" diameter. The shaft is 303 stainless and the rotor plate is 6061-T6 aluminum. We have had problems with the holes in the shaft being drilled too deep, and there are instances of an assembly have very little pin engagement into the aluminum plate. This reduction in bearing area caused the pins to deform the holes in the aluminum plate, and loosen the central bolt. There's also the difficulty of ensuring all 6 pins fit into all 6 holes without excessive clearance.

I would like to increase the robustness and manufacturability of this interface. Axial space is quite limited "overhead" of the brake.

The 25mm keyed shaft is held to f6 tolerance to provide a slip fit into the worm reducer hollow shaft. Concentricity between worm reducer output axis and brake axis is critical to achieve full holding power of the brake.

The brake stator is supplied with a 28mm bore of H7 tolerance, so there is not much room to increase the shaft diameter (25mm currently, <28mm max).

I would like to move to a keyed or splined design instead of the 6 pins. At best I would have around 17mm of axial engagement length to couple the rotor plate and the output shaft.

I would also like to move to a shaft and rotor plate with higher strength, like 17-4 H1150 instead of the current 303ss.

A splined coupling I'm worried will have high manufacturing/development costs and will not provide as good concentricty as a keyed coupling.

Any suggestions would be greatly appreciated.

(I've attached a crude diagram showing general layout of components)

Beat to fit, paint to match.

 

[desertfox]

Hi mrpi

Firstly just a few observations the six pins in the 25mm shaft are to close to the centre of rotation and are taking
quite a load, for example you don't say what pitch circle diameter the pins are on but the radius has to be less than 12.5mm and there roughly 4.7mm in diameter and they need to miss the m8 bolt in centre sheesh! lets say there on a pitch circle of 16mm dia so a radius of 8mm, now at 100Nm brake torque then assuming those 6 pins take equal share of the load which from your post they clearly don't, each pin would see 2083N which is just under a 0.25Tonne each.
So even at the outer diameter of the 25mm shaft (12.5mm raduis) assuming you got rid of the pins your looking at a force generated by the torque of about 8KN.
so I don't blame you for wanting rid of the pin design especially as you would never get six pins working all together anyway.
What centralises the rotating part of the brake or are you just relying in the central m8 bolt and pins? because you make no mention of a guiding sleeve or anything that holds the rotating brake concentrically to the other half although you say concentricity is critical.
Your material choices are stainless steel and aluminium I presume that its the enviroment that the drives working in that dictates those materials?
Another main consideration of course is the heat the brake as to dissipate during operation you make no mention of that yet your material strengths during braking could be significantly reduced and again lead to the failures your experiencing.
can you fill us in with a bit more detail, if your having failures in the field then you need to fully understand the cause before you can correctly address the problem.

desertfox

 

[mrpi]

The pins are on a 17.8mm circle.

The 100Nm load is worst case expected only several times during the life of the unit as it is a transport-shock-type loading. This is also a value provided by the customer and I believe it has been padded by several applications of safety factors as it was handed down through the design departments.

Nominal working load peaks at 33Nm and is operated at 5% duty cycle.

The brake stator and rotor are both located into place using alignment jigs and test indicators are used to verify concentricity.

The central M8 bolt is there to hold the rotor plate up against the shaft. The pins are to transmit torque, not provide the concentric alignment. If the rotor plate were a self-aligning design (spline, keyed shaft, etc) it would simplify the assembly process by eliminating the jigs and test indicator setup process.

The brake is only powered during the 5% operation (~2 min), and consumes less than 5W so heating is negligible.

Max operating environment is 52°C ambient air temp.

Materials were chosen for corrosion resistance, weight, and ease of manufacturability.

We have seen 2 units fail in the field. I believe it was a combination of insufficient pin engagement length and excessive pin hole clearance in the aluminum plate. Its possible that the assembly tech reamed the holes in the aluminum plate using a reamer chucked into a hand drill because he felt the fit was too tight. I could see this producing a not-so-ideal hole profile. I don't know how many units were reamed in this manner....

I have not had the opportunity to examine the failed units. The field service tech noted that the central M8 bolt was loose and then noted the damage to the pin holes in the aluminum plate. I (the engineer) was not notified of the damage or the field repair until after the fact. Otherwise I would have had him measure the amount of pin engagement and make other observations of the failure.

I am unsure if movement allowed by excessive pin hole clearance broke the Loctite thread retainer on the M8 bolt, which then backed out and allowed further movement between the rotor plate and the shaft: or insufficient pin engagement allowed yielding of the plate which allowed the movement which broke the Loctite thread retainer.

Short of it is, I would like to move away from the pin design.





Beat to fit, paint to match.

 

[desertfox]

Hi mrpi

How many units do you have in the field and how long have the been in service or is this a new product?
Using a single bolt to hold a plate to a rotating shaft is not the best idea, I can see it coming loose and undone each time the brake operates, the only thing possibly keeping it there any length of time is the pins.
After you align the rotating brake plate with the stator part via jigs what keeps it concentric after the jigs have been removed?
Put a key in the rotating brake plate and fit a keep plate behind the rotating brake disc to stop axial movement and allow the keyway to take the torque.
Also it sounds to me as though it time you operate the brake that the clearence in the pin holes allows the plate to rotate thereby unscrewing the M8 bolt.
Have you any calculations on how tight the m8 bolt should be or on what stresses the pins/ plates are seeing.
Also if the pins are sloppy this also means the plate and pins are seeing an impact load on every operation.

desertfox

desertfox

 

[mrpi]

I used the pin/central bolt arrangement as this is what I have seen used for flywheel/crankshaft mating on VW engines, and it seemed like a good idea at the time.

This a new product for us, and we have ~16 units in the field that have been in service for 8 months or so undergoing testing. Its possible that earlier firmware that was being tested may have powered the motor through the brake or engaged the brake while the motor was still spinning. Also, since the actuator is holding up a load, cyclic loading is only seen when the software allows the motor to power through the brake. Normally the load is only seen in one "direction".

The software normally ensures that the motor has completely stopped before engaging the brake to eliminate shock load on the brake. The actuator is servoed into position, if that makes sense.

The brake plate is secured with three M8 bolts and friction holds alignment. Only the central bolt that holds the plate to the shaft has seen a problem loosening on the two units.

The rest of the units have operated without problems.

Hand calcs (back-of-a-napkin-type) showed the max loading to be below that which would produce yielding in either the shaft or plate. FEA models show the stresses to be even lower than what the hand calcs indicated (at full pin engagement, which may not have been provided).

If the central bolt were over-torqued during assembly to the point of yielding the female threads in the shaft, would this compromise the ability of the Loctite to retain the bolt?




Beat to fit, paint to match.

 

[desertfox]

Hi mrpi

Okay so the brake is really a holding brake not really used for dynamic braking.
Now I am slightly confused I thought there was only one m8 bolt and six pins holding the rotating brake disc now your saying theres three did I miss something in the first post?
Anyway I would say your alignment would be better with a keyed shaft for the rotating part as compared with friction so I would go for it.
Regarding the loosening of the central bolt I would say its more likely that its not been tightened enough as opposed to the female threads yielding due to overtightening but thats my opinion without knowing or seeing the parts I cannot be sure.

desertfox

 

[unclesyd]

I once saw a setup similar to yours that used a square end drive shaft with an external retaining ring as an end stop.

there are some other possibilities but will require a little more investment.

 

[williedawg]

If the pin method is the only option, it may have worked, but the pins need to be a press-fit in the aluminum, and slip-fit in the 25mm shaft, but this requires precision locating.

yes, there are other possibilities.... the keyed shaft makes more sense; you could even put in 2 keys at 180deg if you're concerned about the aluminum plate .
Typically, tho', you'd want the brake plate bottomed out against a shoulder on the shaft...which I think can still be done....maybe 20mm dia step and make it less in length than the thickness of the al plate, so that the 8mm bolt holds it tight.

Should the aluminum still cause problems, make a steel hub and bolt or rivet it to the aluminum plate; thus transferring the torque load from steel to steel (using the keyed method)

This picture is of a cable winch mechanism that recently failed. They had used a "flatted" shaft and corresponding hole in the large gear. The hole just mushroomed out so much it wouldn't turn the cable drum. If I decide to fix this winch, I'll make a hub with the correct profile and rivet it to the bored-out gear.

 

[mrpi]

desertfox, sorry about the spotty description. The brake rotor is a steel ring of teeth that is bolted to an aluminum plate with three M8 bolts, the aluminum plate (rotor plate) is in turn bolted to the 25mm shaft with one M8 bolt over the 6 pins.

The 25mm shaft is retained into the gearbox with Loctite sleeve retainer to fix it axially and its specifically set to each brake. So replacement of the 25mm shaft would be ideal, but isn't possible/practical for a field repair.

I have already made steel plates to replace the aluminum ones that will fit longer pins.

unclesyd, I'm looking into a polygon-type shaft end as an alternative to the keyed shaft. I'm still waiting to hear back from General Polygon regarding concentricity and torque capacity of their profiles.

williedawg, the keyed shaft was my first reaction, but there just isn't much room to fit a hub over the key(s) and I'm worried that going down in diameter will make the shaft the weak link. Absolute max diameter is 28mm, so once you get an M6 key in there, that's 3.2mm for keyway depth and at least 2mm over the keyway so now the max shaft diameter is down to ~18mm. I suppose I could have the steel hub wire-EDM cut with integral keys.

Thanks for the replies.


Beat to fit, paint to match.

 

[unclesyd]

The use of the a polygon was what I was referring to in my post.
Here is another supplier of Polygons, Stoffel, that you may want to check out.

I wouldn't worry about the Polygon carrying the load. In all our applications I recall concentricity being a problem.

http://stoffelpolygon.thomaswebs.net/home.htm

 

[desertfox]

Hi mrpi

Thanks for clearing that up so If I understand correctly now then the aluminium plate that is fastened to the shaft relies on its alignment with the fixed part on those six pins and central bolt after assembly.
Your pin engagement length issue sounds to me as though its a tolerence issue on the manufactured parts, have you done a tolerence study? Have you done any calculations to confirm a keyed shaft is to weak? are the brake parts ie aluminium plate/ steel brake able to be modified are or the bought out items?
Not sure how cheap a hexagonal shaft will be, what I do know is that the fit as to be pretty good otherwise you end up driving on the edges of the hexagon and from my experience another bearing failure.
You can roughly get an idea of the torsional stress on a hexagonal shaft by calculating the shear stress on a round diameter equivalent to the across flats of the hexagon.

regards

desertfox

 

[mrpi]

desert fox: The aluminum plate is held to the shaft and is located by the pins and bolt. The tooth ring rotor is fixed to the aluminum plate by 3 bolts, and relies on these three bolts for alignment.

I have not performed any tolerance study.

Both the aluminum plate and steel shaft were custom made, so any modification is possible.

The calculation I did showed the max fittable key length to have a marginal safety factor. There just isn't much room in that area.

unclesyd: thanks for the link. I had never heard of polygon couplers until recently.

Because this is such a critical component, I think some investment in having a spline or polygon coupler is still a good value.

Beat to fit, paint to match.

 

[desertfox]

hi mrpi

Well either I am misunderstanding or you missed my point:-
the three bolts fasten the toothed ring to the aluminium plate which I understand however the plate is then mounted to the shaft by the pins and central bolt, now the stator part of the brake is already mounted over he shaft and the only thing I can see that positions the aluminium plate with the toothed gear relative to that stator is the central bolt and pins or did I miss something, I can see that three bolts position the toothed ring in the plate I cannot see how those three bolts help to align aluminium plate and the stator part of the brake.

desertfox

 

[desertfox]

Hi mrpi

If the brake parts are custom made why not increase the shaft length at the rear and allow it to pass through the aluminium plate, if required you could increase the diameter of the shaft and the hole in the aluminium plate locally so you could fit a key. Okay a stepped shaft is more expensive to produce but is it any more expensive then going for a spline?

desertfox

 

[mrpi]

desertfox,

I think I see the confusion here: The aluminum plate is "fixed" to the shaft, the outer three bolts "float" the toothed rotor for concentric alignment/adjustment after assembly.

The ultimate solution would be to just make the shaft and aluminum plate assembly all out of one piece of steel. Eliminate the joint in question all together.

I was hoping to keep the shaft constant diameter, and I haven't determined if having a polygon/spline machined onto the end of the shaft would save anything over a stepped shaft.

The central bolt only has ~.150" over its head before it hits other components. So I dont have much room for more engagement length.

This raises another question about reliability. Right now, there's only one fastener holding the whole thing together (the one central bolt). If I were to locally increase the shaft diam as you suggested (put a bolt flange on the end) I could fit more bolts and/or pins on into the increased diameter. This might be a better alternative than the spline.

Hmmm, I'll see what I can fit. Thanks for talking me through these ideas.




Beat to fit, paint to match.

 

[desertfox]

hi mrpi

Okay I understand now the bolts float and alignment of the two brake halfs is acheived after aluminium plate is assembled to the shaft,now what happens to the alignment when the central bolt on the shaft comes loose and/or the pins are a sloppy fit in the holes, presumeably the alumnium plate will drop slightly due to clearence in the hole for the bolt?
The downside of having a stepped shaft is that you can only assemble it from one direction but the advantage you might gain should outweigh this.

desertfox

 

[mrpi]

Exactly, when the central bolt comes loose, the plate drops off the end of the shaft and the brake no longer works. That's the problem I'm worried about. The central bolt actually has to back out quite a bit before the brake teeth no longer engage, the stator tooth ring is spring pre-loaded against the rotor.

The fit of the pins keeps the plate from "dropping" radially away from the shaft even when they wallowed out the aluminum plate.

The brass tooth ring on the stator has a small amount of movement radially and it will self-center to a small degree. So its only the last ~25% of full-load capacity that is compromised by concentric run out. If the brake is a little bit out of adjustment you wont notice it until you try to put full torque on it.


Beat to fit, paint to match.

 

[unclesyd]

Don't machine your shaft, roll it. I don't know if rob can roll just the end. like an axle or not. I put this i s they can broach anything they roll.
The other urls are for people that can make end splines like axles.

http://www.grobinc.com/coldrolled/

http://www.coldformingtechnology.com/index.htm

http://www.michiganautomaticturning.com/

if you do a lot of machine design work you should get a copy of the Stock Drive Catalogue.

http://sdp-si.com/Catalogs.htm

 

[desertfox]

Hi mrpi

Ah so now we agree the aluminium plate at the end of the day relies on the central bolt and pins on the shaft for its alignment, once anything is loose or worn alignment starts to go albeit thats its only noticable on full load.
Good luck and keep us posted.

desertfox

 

[macthenife]

Replace the ring of 3/16 dia pins with bolts.

The flange and wheel hub of a motor vehicle are clamped in this way.
The wheel studs combine the functions of location,clamping and torque transmission(without the need for accuracy).