Misumi RSRR25 and RSRR25... inexpensive linear guides

[dirktheeng]

This is a continuation of "Linear Bearing Idea". I've been trying to post some stuff on the buildlog, but can't post images. I think it is becuase Bart own's that buildlog and it won't let me post images and such... only write text. I don't have that trouble when I'm blogging here. I think it has something to do specifically with buildlogs. I don't know.

Anyhow, Here's the images of the fairly low cost misumi rails and car. I can get the guides set up for about $58 for 2 guides and cars. This is about half the cost of v groove bearings and v groove rails. The cost for the v-groove bearings alone is about $40 to $60 depending on where and how many you get at once. Granted this is not going to be as precise as a v-groove solution, but man, how much precision do we need? That cheep chinese laser system that FSE uses has square tubing and plastic bushings for bearings... that tells me that the largest advancements to be made are not in hardware, but control software at this point... so get the cheapest rails that give reasonable accuracy. Even still, I think this will give about 0.001" of accuracy for short runs and be extremely easy to construct.

Misumi_480 Rail.pdf

(135.27 KiB) Downloaded 1967 times

Misumi_480 Rail 2.pdf

(76.22 KiB) Downloaded 1639 times

Misumi_480 Rail 3.pdf

(49.39 KiB) Downloaded 1585 times

[bdring]

That is about what I thought they were. Why don't you buy some and give us a report?

[lasersafe1]

I generally agree with all of this "build it yourself" train of thought, but the rails are the one place where we shouldn't skimp out to save a couple of bucks.... That is, unless there really is a good cheap solution. Yes, the Chinese rails look skimpy, but when they roll they are smooth and solid. There was zero flexure in my first M40. I have a buddy who is getting ready to build a fairly large system and I am advising him to buy the best rails he can afford. That is heart of the machine.

[dirktheeng]

I don't want to sacrifice accuracy unecessarily , but we have to try and understand what level of accuracy is needed and not pay for more. In general the cost of attaining a tolerance is proportional to the inverse exponential of the tolerance measurement.

We need to come up with a set of tolerances for rails that is considered acceptable. After that, how do we come up with an objective way of measuring different ideas or products to see if they can attain that tolerance?

Can we come up with a test or two that we can measure Bart's current rails with and then do the same things with other rails? I can buy the rails and measure them, but without having something to compare to, it won't mean a lot.

I want to work with you to figure out how we can come up with a machine that is as good as Bart's design, but perhaps less expensive. In my opinion, the biggest area to cut some costs is in the rails. The V-groove rails and bearins are quite expensive and accounts for about 1/3+ of the cost and is about the only place we have to make some gains because all the other parts have a very fixed cost and have no room for design improvement in fabrication.

The question is, what is good enough and what can we find that is good enough and not pay a cent more for it.

[bdring]

All the recent traffic from Hackaday, Make, Fabbaloo and other blogs has generated a few donations. Thank you to everyone who donated. I try to plow that money back into the site, ideas and experiments.

I would be willing to to pay for one chunk of this material and a car. I need to buy some other material too, so I will have it all shipped to me, I will check it out and can send it to dirktheeng after that. If you tell me the exact length you need I will use that length. I want to order ASAP, so if you want it free, let me know soon.

[dirktheeng]

Ok... that's great!

lets get an RSRR25-480 and an RSRS25 and see how we do. I am going to call Misumi today and ask about a few things... let you know what I find out. They also have other rail/linear guide profiles that I found by mistake. I want to ask about these things too.

[dirktheeng]

Hey,

I talked with Misumi... They weren't able to tell me a whole lot about this product. That doesn't instill a lot of confidence. For these kind of designs, you want to be able to pre-load the wheel asembly in the track... that's how they (along with PBC linear) deal with a C-track arrangement. They could not verify one way or another if this could be done. We could get the product and it could really suck, but it could be ok too.

[bdring]

It might just pre-load by it's default dimensions. It is force fit in and stores the tension in the bracket and studs, like a ball bearing drawer slide.

The slide looks really small at about 20mm across the wheels and the wheels are very close together. The resulting deflection affects and forces due to load are directly proportional to the distance between wheel. That is how blacktoe can get away with v rail and MDF. They are spaced several inches apart.

On my design, the wider the carriage, the stiffer it is. Twist force has to deflect a wheel more to get the same affect when the wheels are wider apart. It is a trade off you need to find a balance for.

Well, my deal for the slide ends tonight. I need to order my parts. With that said, don't order unless you have good confidence this is a likely solution, I don't want to waste money.

[dirktheeng]

I just looked again and called the engineers again. It looks like the clearance between the rail and wheel on the RSR25 is 0.1mm. They are not preloaded. 0.1mm is about 0.004", not so good when you want to go to 1000dpi. However, they did direct me to the RSR40, which has a clearance of zero mm. The engineer said that these are preloaded. In addition, there is a 40mm spacing between the rails... more ability to resist torque.

I also found that they sell a kit with rail and slider together. The price for the RSR25-480 is $38, the price for the RSR40-500 is $47.10. With first order discount this comes to $26.60 and $32.97.

Lets also consider the 2 different applications for the rails. The Y-axis is very different from the X-axis. 2 railse are used ont he Y-axis. The Y axis is rigidly locked in the direction of travel by the belts and cannot torque in the plane of the gantry because it is bolted to the gantry. Therefore, it can only roll about the axis of the gantry. Therefore, we only need to be concerned about controlling that roll. Since 2 slides are used, even if they aren't pre-loaded as in themselves, we can apply a preload by applying a small load in the direction of the gantry. In that case, we would want the side of the car with 2 wheels on the outside. And what we really care about is the roll with the common axis of the gantry. If they are preloaded by appying the small "pinching" force to the rails, the car wheels will not allow that motion. Basically, this would be akin to removing the inside wheels from your rails and applying a load to spread the outside rails. This would be necessary use the RSR25 series since there is clearance between the rail and the wheels.

This would not be necessary for the RSR40 series as the clearance is zero. With jurry rigging, the RSR25 could work for the Y axis, but definately not the X axis. It would seem that the RSR40 would be a better solution all around and could be confidently used for either.

The RSR40 could yeild a good Yaxis solution for about $66 total with the discount from misumi. I would suggest ordering an RSR40-500 after talking with Misumi. This is an all metal solution with apparently a zero load slop of zero. Givent That it is all steel (rails and cars are chrom plated, wheels are plain steel), the stiffness underl load is bound to be high. If you can spare the money, I think this would be a better thing to try first.

[bdring]

Since these are single row ball bearings you might consider the bearings to run in the vertical orientation....just a thought.
.

I probably don't get the discount since this far from my first order, but I will try. I can get one axis for you. I should probably get the lower quality one. Give me the exact part number you need. Make sure you can actually use that number in a "shopping cart". I have had trouble with some of the numbers.