Re: How is the belleville washer solution working?
|
Hi gang, Sorry ... I've been on a long vacation, which was an amazing trip through some national parks.? I'm in LA today and getting finally home tonight.?? I've read Anthony's noted and the others. ? To answer Brian's question: the two force directions (radial that the spring loaded drive addresses, and axial that the Belleville washer addresses) are perpendicular to each other, so they are independent. No need to worry about them interfering with each other. I also understand Anthony's concern about damage to the balls or races of the small R4 bearings, if we apply too high an axial force.? The bearings can be broken apart by impacts and I have also done this during disassembly of a bearing block.? Like I said...buy some spares!?? However, Magnus is Scandinavia proved that heating the blocks allows for much easier effort in yanking out the stock bearing, because the aluminum expands faster than the stainless steel of the bearing. Anyway, the fully compressed force of the R4 Belleville washer is only 7 lbs. The bearing itself must be polished down in OD to allow it to slip (when lubed) so the 7 lbs force can keep it forced toward the opposite bearing.?? Also... I keep my scope load as balanced as I can make it. It is never perfect. The only load on the worm when balanced is when it is accelerating (F=ma).? And as Anthony said that's when you are starting to slew or stopping a slew.? That is rare... and I therefore believe the worry over bearing damage is not a real concern.? ?Experience on my mounts with super heavy (like a 65 lbs f6 10 inch Newtonian) mounts shows that once set up with a Belleville washer on the old version OPW, it is running fine year after year.? Those are with R4ZZ ABEC7 bearings and a brass precision worm. Here are some other notes: 1. - which Belleville curve is best? That picture Anthony attached of the Belleville (that I drew...you are showing) is of an older sketch, and is not optimal ... I now recommend you use the opposite curve on the washer, so the outside of the bearing is pushed in toward the other?side.? That keeps the inner moving race free to move on the bearing balls.? We only want the bearing balls affecting the tracking and PE.? 2. - why can't you just push the far block in by hand, like the Losmandy video shows, to keep the worm from lateral motion? (Derek has it right... The problem is inevitably thermal effects that the Belleville spring addresses). In the original drive design, the far end worm block is on a stud that mounts into an oversized hole.? In that case, you can squeeze the far worm block toward the opposing one.? You must simultaneously keep the two worm blocks lined up, so the bearings are on axis with the worm axis, and simultaneously push the far block toward the ring gear, to keep the worm engaged.? (David Partridge did that with a Belleville washer, and ABEC-5 bearings,? and he got nice low PE results.) In the case of the second drive design, the OPW version, the worm blocks are kept in line automatically by the OPW metal channel.? But... the holes mounting the worm blocks are pre-drilled, so you can't squeeze the worm blocks toward each other.? If the worm happened to be slightly long, or the holes in just the right place, the worm might be automatically held firmly. If the worm length slightly short, etc. the worm can shift.? It's left to chance.? The OPW itself then gets bolted to the mounting plate, and there the enlarged far mounting holes allows the whole OPW to rotate in toward the ring gear. Likewise, on the newest spring loaded system, it has an OPW design too.? There the worm blocks mounting is pre-drilled, so you can't really compress the worm bearings.? True the spring pushes the worm inward, but if the worm can shift along it's axis, you get hysteresis in the axis from the ring gear rotation.?? Bottom line (my view of the various systems): best to follow the OVision worm system, and employ a Belleville spring washer to always put a force (the R4 size puts about 7 lbs) on to the outside of the far worm bearing, pushing it into the worm and the "near" worm bearing (it is "near" the gearbox). The thing that is hard to imagine is the effect of time and temperature variations on mechanical systems.? The subject is known as "reliability."? The brass, aluminum, and steel parts all expand and contract at different rates and therefore the gaps between them are temperature dependent.? That's why you need a spring, to ensure that under all thermal conditions, the parts are in contact and there are no gaps.?? Hope these notes make sense! Very best, Michael PostScript: I was working on the question about how much distances amount to 1 arcsec.? My calculation for a G11 is: Ring gear diameter D = 5.625 inch Ring gear circumference ~ ¦Ð ¡ã D ~ 17.67 inch. (there are 360 teeth to ring gear, but that has no effect...but note that a GM8 has 180 teeth so will have twice the rotation effect) 1 degree of rotation = 17.67 inch/360 deg = 0.05 inch circumference movement per arcsec.? There are 3600 arc secs per degree. So 1 arcsec circumference movement = 0.05 / 3600 = 1.39 E-5 inches. So a gap of 13.9 micro-inches would give a 1 arcsec hysteresis. The worm rotates 1 cycle in ~ 8 minutes = 240 sec.? There are ~ 24 ¡ã 3600 = 86400 secs per day (slightly less for sidereal) so there are 86400/240 = 360 worm rotations per day, or 1 worm rotation per arc degree.?? There should be 0.05 = 1.27 mm thread pitch to the worm.? (I'm not home to see if that is correct). So 1 arcsec = 1.39E-5 inches (13.9 micro inches) = 0.05 inch ¡ã Rotation angle fraction.? So the worm must rotate 2.78E-4 ¡ã 360 deg = 0.1 degree of the worm rotation to give 1 arcsec. The motor is geared down by 25:1 gearbox so the motor must rotate 2.5 degrees for 1 arcsec on the ring gear.? There is an optical encoder in the motor to get that precision.?? The small angles of the worm rotation require there be essentially no gaps in the mechanical parts.? That's why earlier recommendations were to hang a weight on the RA axis, or unbalance it, etc.? But those techniques also only load one if the worm bearings, and leave the other bearing unloaded axially. And the bearings don't work as designed unless they have an axial load to keep the balls in their intended races.? Best, Michael On Sat, Aug 17, 2019, 8:59 PM Anthony Q <anthony@...> wrote:
|
Michael Herman