#robotics Logs

Apr 20 2012

10:50 tsmeow-afk hopes no one ooc's
10:51 LoRez hah
10:54 katsmeow-afk you know what's ironic about the use of skirts with pistons tho? the rod swing with a crank precludes the skirt existing exactly where the wear is the worst, you can verify this in any 200k mile car engine: the cylinder goes oval where the skirt is minimal
10:57 katsmeow-afk it's due to the side vector when the crank throw is 90 to the cylinder axis, a situation you won't have so much with the "short throw" high pressure swashplate
10:58 katsmeow-afk a really high angles swashplate tho, you could have clearance issues
10:58 Skwint it's getting harder to take you ooc kats
10:58 katsmeow-afk <whew>
10:59 katsmeow-afk if engine makers used crosshead guides, engine blocks would last forever
11:02 katsmeow-afk in your small pump, it's too complex, and i don't think it's necessary
11:45 katsmeow-afk aha ha! i figure out how to make the mechanical injector pump deliver an dynamically electrically programmed variable amount of fuel to the cylinders, as needed
11:46 furrywolf well, I wasn't thinking of a complete separate crosshead assembley, but rather a bit of a kluge... just a bearing inside the cylinder. have the connecting rod, piston tang (skirtless piston), and the large bearing all coaxial (one would need a hollow shaft), then the bearing can take the side loading and roll on the inside of the cylinder walls.
11:47 katsmeow-afk but the cylinder is round, a round OD on a bearing is going to rub worse than the piston skirt ( which is a kind of sleeve bearing)
11:48 katsmeow-afk plus any weight added to the rod-piston assy is going to increase wear on the rod-swashplate connection
11:48 furrywolf I was just going to use a regular cylindrical bearing, and let it run on the outer corners, figuring the loading wouldn't be enough to make it wear unacceptably.
11:48 katsmeow-afk corners?
11:49 furrywolf umm...
11:49 furrywolf (=)
11:49 furrywolf the bearing can roll on the edges inside the cylinder
11:49 katsmeow-afk fixed bearing on the block, guiding the piston rod, separate from the rod tot he plate?
11:49 furrywolf don't make me draw a diagram. I very very much suck at drawing. :)
11:50 katsmeow-afk there ARE no edges in a cylinder, unless you mean the rod edges
11:51 furrywolf I mean the edges of the bearing.
11:51 furrywolf bleh, let me attempt (and fail) at drawing.
11:51 katsmeow-afk lemme point ut too, in an oiled pair of surfaces, there is no metal-metal contact, and friction is only matter of oil viscosity, whereas with a ball or rolelr bearing there is goign to be pont (or line) contact metal
11:51 furrywolf that's true at normal pump speeds, but at this low of rpm with the forces involved, an oil film will dissipate without extreme pressure additives.
11:52 katsmeow-afk i'd go with a skirt, groove oil control galleries in it, and pump 20wt oil into them
11:53 katsmeow-afk the skirt won't have any extreme pressure, that's part of my point, you don't have the side thrust associated with a crankshaft engne/pump
11:53 furrywolf I really don't want a pressurized oil system. I know you really like that idea, but... heh.
11:54 katsmeow-afk with the piston always going str8 down, there's virtually zero radial pressure on it
11:54 furrywolf nope, I just failed at drawing. not getting a diagram.
11:54 katsmeow-afk tell you how i'd do they bearing isea: use square rods, mount ehr bearing on the block, not on the rods
11:55 katsmeow-afk my fingers are still sore
11:55 katsmeow-afk the worse the traffic, the harder i grasp the sterring wheel,a nd the worse my paws are
11:56 katsmeow-afk i fear you will make the piston-to-plate area too complex, too heavy, too pricey, and take up too much space
11:57 katsmeow-afk plus, you are adding a place where slop can occur in a moving system, which will decrease stroke and therefor oil pressure if the slop increses
11:58 furrywolf I'm thinking about a regular ball bearing, coaxial with the wrist pin, inside the cylinder. From outer layer to inner layer, you'd have the crosshead bearing, a hollow tube/pin, the bottom of the piston (split in two halves) on this hollow pin, the wrist pin bearing (inside the hollow pin), the wrist pin, then the connecting rod, also split, on the wrist pin.
11:58 furrywolf I think I've figured out how to simplify the plate/rod mounting.
11:58 katsmeow-afk so the bearing does no spin, it acts as a separate skirt?
11:58 furrywolf unless I'm completely missing something, if I use a CV joint instead of a U-joint to drive the plate, and have the height of the mount right, the rod/plate interface is only three bearings, and only one bearing in the piston.
11:59 furrywolf the bearing does spin, and rolls along the inside of the cylinder. there will be a slight bit of slop, yes, but I don't think it will be an issue if it's sized correctly.
11:59 katsmeow-afk you missed something: the cv joint is not designed to take significant axial thrust in eaither direction
12:00 furrywolf the rotating part of the plate will be supported by a thrust bearing on a yoke that sets its angle and takes the axial force.
12:00 furrywolf the rotating part of the plate can't be used to set the angle; it has to ride on something else.
12:00 katsmeow-afk right, in all swashplate designes, there must be a thrust bearing somewhere
12:01 furrywolf I hate not being able to draw. I can see this whole assembly moving inside my head, with absolutely no chance of getting a decent approximation of it on paper.
12:01 katsmeow-afk i am familiar with swashplate deigns, moving mental pictures (sometimes with sound) is what keeps me up all nite
12:02 katsmeow-afk you can avoid the yoke you mentioned if you use an offset driver syystem
12:03 katsmeow-afk the yoke becomes part of the input shaft system
12:03 katsmeow-afk i cannot draw it, too complicated
12:05 furrywolf sigh, failed at drawing, again.
12:06 katsmeow-afk many different swashplate drawings to pick from : http://hydraulicspneumatics.com/other-technologies/chapter-8-air-and-hydraulic-pumps-part-2
12:07 furrywolf if you have the swash plate, with a CV joint on the shaft, there's a point on the shaft about which it can pivot, which I'll say is in the center of the plate. if you mount a pair of bearings in the plate, around a line through this point, the put a second bearing in a yoke in the middle of these bearings, with the connecting rod on it, the rod will only move in one axis.
12:07 furrywolf blah, and that description sucked as much as the drawing.
12:09 katsmeow-afk lol : "Remember too that the only time a fixed-volume pump will be at full flow and full pressure is when the device it is driving has stalled." <<== if the hydraulic motor stalled, it's flow goes to zero!
12:09 furrywolf yay accurate information? :P
12:10 katsmeow-afk yeas, the internets is so noisey
12:13 katsmeow-afk http://www.mathworks.com/help/toolbox/physmod/hydro/ref/swashplate.html
12:17 furrywolf blah. the best way to remember how badly you suck at something is to actually try.
12:25 furrywolf http://fw.bushytails.net/pump/swashplate-rod-01.jpg my failing at drawing. I was worried the rod-plate interface would be excessively large and complex too, but unless I'm missing an exis of movement, that seems quite doable.
12:26 furrywolf the bearings on the ends would need to be able to handle some thrust, but not too much, as long as the swash angle was reasonable.
12:28 furrywolf with a CV joint in the middle, the piston end only needs a simple wrist pin, unless, again, I missed a degree of movement somewhere.
12:28 furrywolf as long as the bearings on the swashplate are on the same axis that it pivots about.
12:29 furrywolf s/exis/axis
12:29 katsmeow-afk i sorta figured you'd go with something like that, because its easier to make than a square cross universal
12:29 katsmeow-afk but the rectable and square cross have thrust capability, the CV doesn't
12:30 furrywolf if I used a regular universal to drive the washplate, I'd need a second universal on the piston end. but if I use a CV, the only movement of the rod relative to the piston will be the radial distance from the shaft, and only need a wrist pin, not a ball joint.
12:30 katsmeow-afk you can even make the rectangular cross in two halves, making building/repair easier
12:30 furrywolf right. behind the plate is a thrust bearing, supported on a yoke.
12:32 furrywolf you don't want the thrust transmitted to the shaft - that wouldn't work. the thrust needs to go to something that's angled, so it moves the plate.
12:33 tsmeow-afk
12:34 furrywolf the big realization, for me at least, was that as long as I used a cv joint instead of a universal joint, and had the rod pivot on the same axis as the cv joint, I wouldn't need another ball joint at the piston end.
12:34 furrywolf assuming I didn't miss something, of course.
12:34 tsmeow-afk nods a
12:35 furrywolf and this means I can build it with all ball bearings using an acceptable amount of money and space, unlike two complete universal joints, one that has to fit inside the piston. :)
12:35 katsmeow-afk if it ever gets slop in it tho, you'll have movement that breaks your arguement against the ball joint
12:36 furrywolf at 30-120 rpm, submersed in hydraulic fluid, I'm figuring it has a lot of time before anything gets sloppy...
12:36 furrywolf consider the cv joint on a car takes 500ftlbs at 3000rpm for thousands of hours...
12:38 furrywolf cv joints are complex to machine, but otherwise simple and cheap... it needs six balls. not six complete bearings, just six balls.
12:38 katsmeow-afk 3000rpm?!
12:38 furrywolf well, that's racing speed. heh.
12:38 katsmeow-afk yeas
12:38 furrywolf 1000 cruising.
12:39 furrywolf but they don't break, even at racing speed. :)
12:39 katsmeow-afk depending on the tire, 1000 is still high for commn driving
12:39 furrywolf hrmm... not really. 3000rpm in unity gear with a 3:1 axle ratio...
12:39 katsmeow-afk 1) the cv won't take thrust, in either directon, 2) the thust bearing takes thrust in only one direction
12:40 furrywolf yes. I'm not planning on any thrust on the CV. thrust on the cv would be bad, as it'd be wasted force, not contributing to anything.
12:40 katsmeow-afk cv is bolted directly to the shaft holding the tire, except in forklifts and HMMVs
12:40 furrywolf they're usually press fit, not bolted...
12:41 furrywolf hrmm, I guess I need a cage too, to keep the balls in place.
12:41 katsmeow-afk press, held by nut on end of shaft, and pinned, been there, done that
12:42 furrywolf the nut and pin is holding the wheel to the shaft, not holding the cv to the shaft. the cv cup is on the other end... haven't looked at one closely, might even be forged as part of the shaft.
12:42 furrywolf the inner part of the cv is press fit on the axle half
12:42 katsmeow-afk DesignerThinking.com\carhub.html
12:44 furrywolf server not found
12:44 katsmeow-afk http://designerthinking.com/carhub.html
12:44 furrywolf oh, you used a backwards slash, and it's taking it to be part of the hostname
12:44 katsmeow-afk it's there, i just checked
12:44 katsmeow-afk then don't use nix :-P
12:45 furrywolf yes. I know how an axle goes together. :P
12:45 katsmeow-afk i just demoed i know about car parts too
12:45 furrywolf I've done probably... *thinks*... 20 axle shafts.
12:46 Tom_itx katsmeow-afk
12:46 furrywolf she's always AFK. :P
12:46 Tom_itx how much is a press fit on a 1/4" shaft?
12:46 Tom_itx .002"?
12:47 katsmeow-afk on a grade 8? prolly the break limit of the aluminum hub
12:47 katsmeow-afk depends on how you mount it, the other interferrence params, splines, keys, etc
12:47 tsmeow-afk is leaving that to
12:47 Tom_itx i wonder how much undersize i should cut the hole so i can heat it up and slide it over the shaft
12:48 Tom_itx simple press fit i think
12:48 furrywolf I've never been good with press fits. they're too tight, I take a tiny bit more off, then it's too loose.
12:48 katsmeow-afk i would cold press it, on such a small piece it will cool before you get it pressed
12:48 Tom_itx i think i can do it
12:48 katsmeow-afk yeas, again on small piece, you'd need a reamer to take off on .0001's
12:48 Tom_itx freeze the shaft and heat the aluminum
12:49 katsmeow-afk easier to do a 1/32 dimple and pointy-end grub screw
12:49 Tom_itx and put the shaft in a jig so the hub doesn't slide on too far
12:50 Tom_itx i can thread thru the hub with a grub screw like i did on those brass ones too
12:50 katsmeow-afk yeas
12:50 Tom_itx at 10,000 rpm it might shake though
12:50 katsmeow-afk what i *hope* for this thing doing it making torque, a good bit of it, but not much rpm
12:51 Tom_itx what you described may strip the 4-40 screws holding the fr4 to the hub
12:51 katsmeow-afk even 20rpm would be a limit, tho surges as high as whatever it takes,, this is a prototype to identify weaknesses and possibilities
12:51 katsmeow-afk i know
12:52 katsmeow-afk replacing the 4-40 with 1/16 roll pins would be the next step
12:52 katsmeow-afk if those shear, i win :-)
12:54 Tom_itx they won't shear
12:54 Tom_itx it would screw up the aluminum hub or the fr4 first
12:54 katsmeow-afk yeas, *if* i can get the torque
12:54 Tom_itx i doubt i get to it this weekend though
12:55 katsmeow-afk *if* i can, it might make you and rue and zhanx as happy as it will make me
12:56 Tom_itx once i get ready, i'll need to get those hole dimensions again
12:57 katsmeow-afk k
12:58 katsmeow-afk the exact disc OD depends on how you can place the 0.235 holes for an even count around the preimeter
12:59 katsmeow-afk i am not terribly concerened about the exact OD at this time, it's a prototype
12:59 katsmeow-afk it's more important that there be no breakthru , breakage, breakout, shattering on the fr4 tween the holes
13:00 katsmeow-afk hence also, the stability and accuracy of a mill , as opposed to the drill press
13:01 katsmeow-afk if the specs don't change here, the 0.235 holes allows me a .0005 pressfit
13:02 Tom_itx is the .235 critical?
13:02 Tom_itx 15/64 is .2344
13:02 katsmeow-afk it's critical to a certain point, why?
13:02 katsmeow-afk o
13:02 Tom_itx .2362 is 6mm
13:02 katsmeow-afk what about i metric bits?
13:02 katsmeow-afk yeas
13:03 Tom_itx i don't have any
13:03 katsmeow-afk 6mm is too big
13:03 Tom_itx i may have a 6mm
13:03 katsmeow-afk 6mm is .0005 too big
13:03 katsmeow-afk bit i can glue it,, "it's a prototype"
13:03 katsmeow-afk i'll accept 6mm
13:03 Tom_itx i'm not sure if my stuff is .0005 accurate anyway
13:04 Tom_itx especially thru fr4 since it's so abrasive
13:04 katsmeow-afk adjust runout = pick different bit size
13:06 katsmeow-afk considering the hole insides will be porus and fiberous, cyanoacrilic glue will adhere well, so 6mm is acceptable, maybe a wallowing 15/64, i dunno your mill's tir
13:07 katsmeow-afk point is, .250 is unacceptable
13:08 katsmeow-afk it you deliver holes a tiney undersized, i can enlarge them .0005 ata time
13:08 katsmeow-afk if too large, , well, glue has it's limits, and exact placement becomes a problem
15:25 katsmeow-afk humans make no sense : "That's a nice section of the river. I don't know what fun they get out of that."
15:26 ace4016 ?
15:27 XSerpentX skinny dipping, no doubt
15:27 katsmeow-afk a $5 million DOG ?!?! : An Illinois man who claims tainted dog treats killed his Pomeranian has sued Nestle Purina and Wal-Mart for $5 million in federal court
15:31 ace4016 i've been hearing about tainted chinese dog treats killing dogs recently...
15:32 katsmeow-afk it's a dog, people give them away, they seldom pay $5 million for killing *HUMANS*, especially in Afghanistan, Iraq, wher thegoing pay is under $30k
15:33 ace4016 if it was a civilian that was killed by a tainted snack, i'm sure more than $5mil would have been asked. most citizens don't "care" about their soliders and some feel no remorse
15:34 katsmeow-afk no, a civilian killed by usa soldier, like recently
15:34 ace4016 ah
15:34 katsmeow-afk no matter, in all the accidental civilina killings, or on purpose, none got more than $30k, afaik, and some less than $8k
15:35 ace4016 the gubermints set that up to avoid law suits, especially since sueing the federla government is like taking a loan out that you're going to pay back
15:36 tsmeow-afk
15:38 ace4016 hrm
15:39 rrywolf stretches out and y
15:40 tsmeow-afk notices the fur in a completely noncommittal and innocent
15:43 furrywolf I'm thinking tripod joint instead of rzeppa... let someone else make the high-precision stuff.
15:43 XSerpentX do it
15:43 katsmeow-afk can you get one as small as you need?
15:44 katsmeow-afk i kow you can get square universales, they are used in some steering columns
15:45 katsmeow-afk barely 2" square
15:45 furrywolf no; I'd be building all components except for ball/roller bearings.
15:46 furrywolf a tripod joint uses three ball bearing units instead of six loose balls... by using standard bearings, it keeps me from having to machine to the precision needed to make my own cv joint.
16:14 katsmeow-afk ouch : http://marinediesels.info/Horror%20Stories/damaged_gear_train.htm
16:15 Steffann nice nice
16:16 furrywolf bah. grind the burrs smooth, keep it in service. there's plenty of tooth area left.
16:16 Steffann katsmeow-afk that website is OLD
16:16 furrywolf the actual damage is, what, 5% of the tooth edge?
16:17 Steffann They disable everything they could
16:17 Steffann right mouse button, you can't select the text ect.
16:17 Steffann etc.
16:18 tsmeow-afk can : marinediesels.co.uk deals with the construction, operation, running and maintenance of large slow speed two stroke crosshead diesel engines such as Sulzer, MAN B&W and Mitsubishi, and medium speed four stroke trunk piston diesel engines such as Wartsila, Pielstick, Sulzer, MAK etc., as found in the majority of ocean going merchant vess
18:43 katsmeow-afk ~"Writers working under tight restrictions produce optimum novel material - like, kinda just guessing for example, epigrams employing deadpanningly cunning backward alphabetization.~"
18:54 Skwint extending xkcd?
18:55 katsmeow-afk yeas, hence the almost quotes ~" ____ ~"
18:55 Skwint should be possible to do an entire alphabet, each letter once, given time :)
18:55 katsmeow-afk it is, done it
18:55 Skwint though starting zyxwv will be painful