Encoder modification

[dale]

I cannot affort either chain or R&p drive and I was having a lot of trouble with accuracy. I decided to try an idea posted on the forum some time ago.
I initially purchased a Shopbot because it was a "closed loop" machine. My problems are with the bathtub chain for driving the encoder. I turned new wheels that have a groove with a radius for seating a length of cable. The wheels are 1.5" diameter and I still use the same idler pulleys for the chain. This gives me very little clearance between the 3 and I get almost 180 degrees of engagement.
A turnbuckle on each axis will enable me to tension the cable.
I am hoping that this will solve my problems with the chain stretching differently at periodic intervals.
I will mount the new wheels tomorrow and I hope my accuracy problem is fixed until I can get a R&P.
Dale

[robin]

I tried this modification (I think it was suggested by Brian in Australia) and it solved my accuracy/repeatability problems.

[seaside@esva.net]

Didn't the guy that came up with this idea originally (Brian?) wrap the cable all the way around the new encoder pulley instead of just running the cable on the under side of the pulley, so he was getting 360 degrees of engagement?

It sounds from your posting that you are going to run the cable on the underside of your new pulley the same way as the old chain was run; I would worry about side load on the encoder shaft when you tighten the turnbuckles.

How did you run your new encoder cables Robin?

Bill

[robin]

As you say, wrapped all the way around. Thanks for the clarification.

[dale]

There is very little tension on the cable. Just enough to turn the pulley. I didn't want to wrap the cable all the way around for fear of the cable rubbing against itself.
I ran the cable over the top of the pulley. I thought I did something wrong, but, my manual shows the chain running this way.
I have been testing the accuracy all morning and have found it to be very consistent.
Now maybe I can get some work done
Dale

[rgbrown@itexas.net]

My SB was not registering the same on different areas of the table; the calibration would not stay consistent. All things being frustrating, I decided to use try the modification Robin and Dale mentioned above.

I removed the sprocket encoder wheels. If the assembler at ShopBot was not trying to win an award from Loctite, it would have been a thirty minute job. Remove a wheel bar from each axis; raise carriage; loosen set screws; pull encoder wheel. On mine, I had to heat the allen set screws to help kill the Loctite; build puller; build a heavy duty puller; heat encoder wheel; pull encoder wheel; clean the Loctite from all of above.

I chucked the encoder wheel in a Jacobs Chuck in a wood lathe and turned off the bead sockets. I reinstalled the wheels and decided to change the carriage rollers as I had a spare set. I noticed on removal a couple of them were shedding metal from the races.

During reassembly, I looped the stranded cable around the 'new' encoder wheels, built some brackets to hold the ends of the cable and ran UC. I don't see how anything can get out of adjustment on this system. The only thing I need to change is change from stranded cable to a 'music wire', where the cable will not vibrate when they rub together.

Ron Brown - rgbrown@itexas.net (mailto:rgbrown@itexas.net)

[jimt]

Just curoius - How are you guys doing with the encoder modifications? It seems like there should be some creeping of the registration as there is no locking positive drive (not that the beaded chain is great - but the idea is correct). It's been a month since the first post. How are things going?

[phil]

With the beta version 2b8 software I have turned the encoders off and don't even use them. I've also disconnected the cables to the encoders. I must say that I have never had such an accurate machine before. It seems that the encoders are more headache than benefit. That is the status of my encoder modification. So far I don't see any reason to use them. I've also modifeid my x drive system to use only one motor (larger one). I've mounted it at the back where the wires used to cross. I also built a 2:1 reduction for the x drive which gives me superior resolution. So far I am very happy with the performance of my machine.

[rgbrown@itexas.net]

Still happy with my arrangement. I haven't done anything to really 'test' the device. I will be cutting 'multiples' all over the table next week and will be able to tell you.

Phil,
Are you 'gear driven'? And how do you power both sides with one motor?

I am planing a modification using 'Gates' belting and geared boxes with a cross-shaft. When I get all the bugs worked out, I'll tell all.

Ron

[phil]

I use a cable drive system and the electronics and software from shopbot; but I use a linear bearing set-up for my x and y axis motion and a screw driven linear motion Z-axis. My frame for the machine is made from 2 x 4 inch aluminum tubes welded into a rectangular frame.

How I power the x axis with one motor:
I have three pulleys on each X-axis side (the near side and the far side) mounted on each X-axis strut. The pulleys are all in the same plane just above the x axis struts. Lets call a point on the top part of the x axis strut centred on the top 1 and 5/8 surface as zero. The first pulley is located at location X0,Y0 mounted on a clevis to allow tensioning, (just a u shaped bracket with a bolt through the ends of the legs of the "U" holding the pulley.) A captive nut at the top part of the U-shaped clevis accepts a screw through a bracket that is mounted to the end of the strut. Tightening the bolt tightens the cable.
The second pulley is at X100, Y1 while the third is at X102, Y-1. I repeat the same three pulley arrangement on the far X-axis strut.
I mounted the X-axis motor on a bracket off the the traverse strut that bridges the two X-axis struts. The motor shaft points up vertically. The cable arrangement is very simple: attach the cable to a point on the X-axis carriage , run the cable towards the second pulley, run the cable conter clockwise starting at six oclock and exiting at three oclock, go around the bottom notch of the motor spool clockwise from 9 oclock and exit a 3 oclock, go around third pulley clockwise from 3 oclock and exit at 6 oclock go around first pulley clockwise from 6 oclock and exit at 3 oclock and attach free end of cable to the same point that the start end of cable was attached. Repeat procedure for the far axis strut, except use top notch on motor spindle. I actually have a belt drive on my motor. I go from an 11 tooth sprocket to a 33 tooth sprocket and connected with a 3/8 wide timing belt. The 33 tooth gear drives the notched spindle from the same shaft. (I have a common shaft for the 33 tooth sprocket and the drive spindle ).
I have two marks on my X-axis struts that I use to ensure a perfect perpendicular line between the struts. I then sequentially tension each X-axis cable and using the marks to keep everything squared up. I keep tensioning the cables until I just barely get slippage when i move the x-axis to it's extreme positions and it bottoms out on the stop blocks that I have installed at each end of the X-axis travel. I also run the X-axis up and down several times and keep adjusting the tension until I have a good perpendicular setup and no backlash on forward and reverse movements.
I am still using the shopbot power supply but plan to install a 30-36 volt power supply as I have very little friction in my system and could get quite good travel speeds but my motors cut out at about 3.5" with a 386 computer. I'm sure the higher voltage supply would help.

I also use a cable to drive the y axis from a motor that is also mounted at the back near the X-axis motor. The cable arrangement for that is kind of tricky and I actually got the idea from an old photo-copier that I took apart. It uses a total of fifteen pulleys (six per side plus three at the motor drive); but it works very well and is extremely easy to set up despite how it sounds complicated.

I have run programs that exceed 20000 lines and the machine always comes back to home origin perfectly.

[jimt]

Phil,

Thanks for the info. I have been unable to use ver2.8 as it was very erratic with accuracy. Could you post a drawing of you cable arrangement somewhere or e-mail me a .dxf file? I've read it several times and it's still a puzzle. also, how big is the single motor you use for the x-axis?

Thanks,

Jim

[jimt]

I tried the cable-encoder drive with no success at all. After 10 air cut rectangles, I was off the zero point by 1/8". I was using the system with the cables wrapped 360 deg. around the modified (actually custom made) encoder pulleys. Anybody got any suggestions? I tried tight cables and looser cables and nothing worked. HELP!!