View Full Version : How To Keep a ShopBot SQUARE

03-26-2001, 05:49 PM
The purpose for this new conversation is to get feedback from other botters on "squareness issues". What prompted me to start this topic is the fact that we just can not seem to be able to keep our SB square ie: y axis square to x axis. We have loosened the bolts on the carriage did the 3-4-5 deal, checked for loose pinion gears (not), I do recall that one fellow mentioned that he squared his SB by bumping it up against some bolts which he placed somewhere on his X axis rails. Is this something that should have to be done? Come-on how hokey. Ted at SB if you read this I hope that you will respond as well, because I had such high hopes for the use of my SB and those hopes are waning with each NEW problem that occurs. I mean that this tool should be able to be kept in squareness with out having to constantly check it or bump it against squaring bolts. Are you guys welding your carriages once they are square, if so do they stay square? Are you certain...I'm truly trying to get to the bottom of this situation. Weather I am able to get my SB square and keep it that way or not this has been my first foray into CNC routing and I am so glad to be here. But dog-gone-it these things aught to stay square and if they can't, then all the SB's capabilities are for naught.
Come on you guys, yea you with the fancy smancy auto leveling table and you sir with the laser leveler and you fine gentlemen from all over, I pose these questions to all of you, can the SB's really be squared by you and can they stay that way? Who among you I ask, claim to be able to square their SB into submission?

03-26-2001, 07:28 PM
Sorry Roney,I'm the guy who was "talking" about
the laser leveler, I don't have your problem,
my table I made is all welded construction .250
wall steel. about a week ago I ran a pencil
around to check level and square it was within
a thou. "so sorry"

03-26-2001, 07:37 PM

My ShopBot is an early model utility strut tool that's been upgraded to R&P. I cut scarf joints for boat planking with my ShopBot, so I'm pretty fussy about it staying square. I spent a lot of time when I set it up getting the rails straight and everything square, then tightened all the bolts as tight as I could get them. I used to check for square every morning when I started my ShopBot, but now I just do it occasionally and it's always where it should be.

After seeing the PRT tools it's hard to imagine that you couldn't get the bolts tight enough to keep it square under normal conditions, so it sounds like it's somehow being driven out of square by the motors. Does it only get out of square when you're cutting, or does it do it when it's just moving back and forth?


03-26-2001, 09:04 PM

I had some "squareness" issues a while back. Once it was the fact that the bolts were not quite tight enough, the other time occurred when I racked the machine by mis-typing a command and missing a decimal point. Since I've torqued everything down tight and added limit switches to ensure that "I" don't hurt the machine, everything's been square.


03-26-2001, 10:21 PM
I had the same problem untill I adjusted the (bumpstops, limiters, whatever they are called on the bottom of the Y-axis). The main problem was actually driving my machine beyond the table limits and racking it out of square because the torque of the motors was enough to twist the connection at the Y-axis carriage bolts loose even though they were tightened pretty good. Since adjusting the (bumpstops) to square on both ends the problem has not reaccured. Anyway, It might not be the right solution, but it worked for me. I suppose if I exchanged the 305 oz steppers for a few 100 oz models this problem wouldn't have accured..N-O-W-A-Y!! By the way: watch out if you do decide to weld the carriage, metal tends to warp towards the weld side.

03-27-2001, 08:45 PM
In your response to my question I sensed that I may have offended you and possibly. In no way did I intend to, in fact I'm extremely impressed by what you, Mr Seim and others are able to do with your SB's. However, we feel that we have properly assembled our SB and overly if not properly torqued our bolts...Dan, my problem is not with our table frame but with our moving carriage...
To the other responses thanks and guys keep 'em coming...BTW if any of you guys find yourselves near Gettysburg, PA contact me, if by then I'm still not square maybe you can square me up (in exchange I'll show you what I can of The Underground Pentagon and Camp David...plus we have an popular and awesome hang-gliding site as well...

03-27-2001, 09:39 PM

What are the symptoms when it gets out of square; does it go out of square as soon as it starts to move, or does it start out OK and then gradually get out of square? Does your tool move smoothly, or does it seem like it's missing steps? What kind of computer are you using?

When it gets out of square, is it always out of square in the same direction? In other words, if you square your tool up and then tell it to move some distance, does one end of the carriage always move the correct distance and the other end always move the wrong distance? After it's been moving for a while, are both x-motors about the same temperature?


03-27-2001, 10:46 PM

No you did not offend me, But I have been thinking about your problem, I had a driver
transitor burn up "my own fault" the bot still
ran but the cuts wouldn't match up, I think
thats what Bill up above is getting at too.
Maybe one motor is not working properly and acting as a brake,driver board wiring, connections,if you sqaure it up does it go out
immediatly or does it take awhile,

Gerald D
03-28-2001, 01:14 AM
Roney, maybe you were thinking of me when you said: "I do recall that one fellow mentioned that he squared his SB by bumping it up against some bolts which he placed somewhere on his X axis rails". These are the photos of the arrangement: (the mods are in red oxide primer color)




You can see a short length of angle iron welded to each of the two y-carriages and four adjustable stoppers (long grub-screw and lock nut)welded to each point of the x-rails.

Regarding the locking of the y-gantry to prevent "racking" - we have got small tack welds of the rails onto the carriages, but this is not meant to stop the racking. The welds are mainly to keep an exact parallel distance between the 2 rails. This is absolutely critical for the v-rollers sitting on top of the "sharpened" rails.

While we were fiddling with the squareness setup, we found that the rails did not keep their distance apart, causing the y-carriage rollers to start lifting off the rails. This lift was never repeatable and so the squareness was a secondary issue until we could be sure that the v-rollers were properly seated at all times. My guess is that a lot of the squareness issues are really caused by rails that out of parallel.

May I add that we now only occasionally bump against the stops on purpose. If we have reason to doubt the squareness (after doing something silly) then it is a very quick and reassuring check to go and find the stops. Also, if we hit the stops by mistake, then we know that we have not stressed the squareness of the gantry, and we can immediately carry on working with the confidence that nothing was damaged.


03-28-2001, 11:04 AM
Gerald, I really like the way you've mounted the X rails.

Roney, is it possible that a gear might slip by a tooth or two under heavy load? I had that happen last night, so will be setting the springs a little tighter.

Gerald D
03-28-2001, 12:05 PM
Sheldon, thanks for the comment, but also be careful to copy this method! No, no patents or anything like that . . . . . .

You see that there are short bits of channel welded upside-down onto the main girders, but there is also a lot of shimming between the channel and the rail (see first photo). The big problem is to get a channel welded so that its top surface is level - in some cases this must also be compensated with different shims just on one channel. If one channel is slightly "twisted" then it will pull an "s"-shape into the x-rail when the 1/2" bolt is tightened. Conversely, if the supplied rail is slightly distorted, then one could use careful shimming to pull it straight.

David Fisk
02-02-2002, 02:19 AM
Arrrrgh my square is not square nor my shopbot. I squared the square and still can't keep the shopbot square. I used the cutting techique for squaring. Seems to me that it would be possible for some sort of squaring routine like the old calibration routine for the cable drive. Is this possible? You could cut a square and measure across the diagonals, input the the information and then if possible the shopbot software would do a vertual squaring.
Ok maybe just a low tech step by step cutting/squaring technique that doesn't require a square.

02-02-2002, 08:52 AM

With only a little knowledge of geometry one can use most machines to square themselves. Well, you can at least get them to show how much error is in the machine.

One procedure would be to cut parallel and perpendicular lines. Flipping the board and lining up the cuts will show double the error in the machine.

I am also sure a good math-whiz could write a program to 'correct' the alignment. It would have to rewrite the .SBP files and correct the numbers there - or maybe I should say miss-correct the numbers as the file would not be useable except on a machine with the exact "UN-square-ness" as the program was written for.

I used a compass to construct square lines across my "X" rails. I then marked these spots with masking tape and made sure my ShopBot was CUTTING square. I permanently scribed this line in my "X" rails. I also have marks that line up when the 'X' value is at Zero. This way a simple glance can tell me the tool needs some attention.

Ron Brown - wdyasq@yahoo.com (mailto:wdyasq@yahoo.com)

If Stupidity got us into this mess,
then why can't it get us out? - Will Rogers

02-02-2002, 01:57 PM
If someone could please verify my math.
My geometry stinks, but I do know "a" squared plus "b" squared equals "c" squared. "a" being the x axis "b" being the y and "c" the diagonal in the triangle. So if I use the retangle command CR 26.833,24,O Then then measure the diagonal I should get (if things are square) 36.0001 inches. If it is too long I need to adjust the zero side of my x carriage toward the right. If it is too short I need to go to the left. This is about the limit of my math. Anyone know how much I should move the rails. I mean is there a formula I could use to calculate how much I need to move them?

02-02-2002, 02:31 PM

Try using your CAD program to check your geometry.

Regarding altering part files to accomodate an out of square machine. I think that would be a nightmare to keep track of.

Would it not be possible to perform a square cut, measure the difference in diagonal and the extrapolate that to how much each X motor is out of sync with the other, then send a corrective move command to only one motor while the other remains stationary (possibly un-hook one motor and clamp it's carriage in position so it doesn't move). Once it's moved into square then tighten up the bolts and set your hard stops.

I've never tried this. If it works, let me know, maybe I will

02-02-2002, 02:35 PM
What about this. I take my newly acquired "c" value plug it back into the formula using the y number "b" fixed at 24 then calculate what the new x value "a" is subtract the ideal of x value of 26.833 from the new calculated x "a" and come up with a number that the x axis must be racked(Left or right). multiplied by 3 since the Y axis is 6 feet long not 2.

Seems like this could be the start of a .sbp file for squareness correction unless I've really messed up someware.

What if you clamped the far x side unplug its stepper motor and then move zero side via the shopbot software. That is after the bolts are loosened.

Ron how did you use a compass?

02-02-2002, 02:57 PM
To check my machine I draw a Rectangle in a cad program and let the machine cut it, them I check from corner to corner to see if it is exact.

If it is not then I just edit my dxf file to match the the offset, then just save that file as a master template.

I then check it to a framing square to see if it is true.

This saves me the time to stop what I am doing to re-square my table.

02-02-2002, 03:26 PM
Ron V.
How do you edit your dxf file to match the offset? Do you do this with TurboCad? I don't know how to pull something out of square with TurboCad.

02-02-2002, 04:51 PM
Ok I figured out how to skew a dxf file in corel draw, but it requires angles to make the change (can't just move the x or y axis). Does anyone know the formula that would give the angle between square and not square? I'm trying to do this without requiring physical tools so there is less chance of operator error.

02-02-2002, 06:05 PM
This program is as intuitive as trying to draw with your elbow.

02-02-2002, 07:40 PM
David here is a movie off my server

Ron V

02-02-2002, 08:35 PM
Thanks Ron, The node edit dosn't work for selecting multiple items and you can't do much more then a rectangle. What I was after was a way to modify a complex drawing with multiple irregular shaped polygons. I think I'll go back to the fixing the table idea. Corel draw can do it, but its difficult to figure out.

02-02-2002, 09:51 PM

I simply make an arc with a stick that has two points - I use ice-picks. I then take the points where the arcs cross a line- say the gear-rack and swing two larger arcs. Where the larger arcs intersect is perpendicular to the center point.

If my description is rough, look up 'line perpendicular to a point' in a basic geometry text

Ron Brown - wdyasq@yahoo.com (mailto:wdyasq@yahoo.com)

If Stupidity got us into this mess,
then why can't it get us out? - Will Rogers

02-02-2002, 10:55 PM
You can use the blend funtion of corel to create your mutiple items... create one extreme then duplicate and invert to create the other, then blend them at 2 too ? in between...

02-03-2002, 12:09 AM
Corel Draw is incredible it can do so much I find myself overwelmed and I often click things on I never intended and I can't do something simple. Corel draw is to a drawing program what a computer is to a typewriter. Much more difficult to use but when you can wow! Experimenting continues . . . I obviously like the idea of vertual squaring.

02-03-2002, 02:01 AM
. . . . and I believe that virtual squaring is completely over the top! We should be doing everything possible to construct a rigid, accurate, square machine and not rely on software to take care of "sloppiness".

Higher up in this thread, there are some photos of what we did nearly a year ago to maintain squareness. In our early days, we sometimes did silly things that knocked our gantry out of square - sudden massive cuts, hit clamps or screws, forget to turn the router on, answered the telephone without stopping the SB, etc., etc., Today, about a year later, we seldom check squareness, and are always pleasantly surprised when we creep up to our end-stops to see that both sides will always hit simultaneously.

02-03-2002, 02:14 AM
In David's program above, he says to clamp the gantry down onto the x-rail to lock it. The rollers and rails could be damaged by this clamp force - that is why we pull the gantry against our adjustable end-stops, without overloading rollers or rails.

We know that the screw pitch of our endstops is 1.75mm per revolution. If we start the motors against the stops and then cut a panel which is (say) 1mm out of square over 1 metre width, we know that this is about 1.5mm over the width seperating the x-rails - so we turn the end-stop about one full turn. That exercise was done a year ago and didn't have to be repeated again. It is a good stable mechanical reference.

02-03-2002, 12:46 PM
Hi David,

With some of the ShopBot control boxes (I'm not sure which ones... maybe all?), you have to make sure you shut down the ShopBot software and the control box before you unplug the motors or plug them back in. It can fry the driver chips if they are plugged in or unplugged while they're powered up.


02-03-2002, 09:22 PM
Yep I agree Gerald, best to do things right the first time, but it helps to have a trick or 2 too use if you have too.

All David had to do is draw his file and select all and change the degree to match the error he was geting on his table.

Ron V

02-03-2002, 09:44 PM
VectorCam has a function called shear that will allow you to keep sides parallel, but place a controlled angle between the X and Y directions.

Draw a rectangle, Change-Shear. It works on straight lines, but arcs can be interpolated to also correctly (temporarily) compensate for slight out of square conditions.

02-03-2002, 11:01 PM
Gerald, your stops are a great idea I didn't realize you used them to actually square your table I thought that they were used, after you got the table square, to verify the tables squareness. I'm planing on doing something simular. I agree your method would be a much more precise way of squaring the rack. I also agree that we need to have the shopbot as solid as possible and not rely on the software. I was very frustrated with my inablity to use a framing square and thus tried to come up with a way to do without. Maybe clamping the clamps to the rail (blocking movement) so the wheels don't undergo the stress would help. I guess this method wouldn't be so good for the new machines with the v shape on the top of the rail. My bot still has the old utility strut with the edge on the corner not straight up. If I could get my shopbot perfectly square which you have acomplished with yours, then I would forget about crazy things like vertual squaring. Thanks for the insite into your squaring technique.

Bill, I get carried away sometimes. I was a technician in the Navy and we did this sort of thing all the time. Its sort of second nature for me to unplug things with the power to them. Its really a risky thing to do though and if you misalign the pins "poof" there goes your driver card. Now I've blown up my driver card by pushing on my carriage "like you shouldn't", but I havn't done it plugging/unplugging my steppers. I guess I havn't worried about it since the replacement chips are fairly inexpensive if you know how to solder. Maybe I could re-write it so the unpluging is done without the power on. Then again maybe using Geralds squaring technique would be best.

Ron, If I could figure out that angle I could use corel draw. It shouldn't be so difficult. I guess I could just get the angle from my cad program like Sheldon suggested.

Fred if I can ever afford vector I will probably get it, unless the up and comming TurboCam is easier/cheaper etc.

02-04-2002, 06:05 AM
David, another ex-Navy bloke here! Sounds like you did weapons alignment between radar and guns. (where the software guys must correct for a flexible ship).

The MO command cuts the power to the motors until the next Move command is given. After this command, I move the carriages by hand, very slowly (and only 1 or 2 millimeters) and at my own risk - have gotten away with it so far!

02-04-2002, 11:15 AM
Ha Ha! No I was a CTM and never stationed on a ship actually. Spent lots of time in remote areas like Adak Alaska, Diego Garcia etc. I'll try add that MO Command to my program.

02-04-2002, 03:00 PM
My updated safer version of the Squaring routine SquareX.sbp is available if you email me. I'll try and get it on a ftp sight sometime but until then if you are interested email me. The censor(s) have decided that freedom of speech dosn't include shopbot parts files on this forum. Mine has been deleted. Another blow to our freedoms all because of fear
we all loose when this happens

Will Sellers
03-16-2002, 04:39 PM
I've read through the postings and I'm getting better at getting into square, but maintaining square is not happening. The latest incident happened on a y axis cut, no x movement. It lost square first cut. Granted there was an uneven load on the bit (making an end cut in 1x24 ash. I made a 3/8" deep cut with the edge of the bit flush with the end). It pushed the y axis out of whack. Y axis move speed was .25 and 1/4" down spiral end mill. I now believe I need to build a new y carriage out of some other materials. Are there any redesigns that work better. I have the PR machine.

03-16-2002, 06:01 PM

Do you have a large enough wrench?

I have a small Uni-Strut frame ShopBot. It is very seldom I can knock it out of square unless I fail to adequately secure a part I am cutting and have a large bit or I just plain mess-up.

I did consider getting some Allen head 1/2" bolts of the right length where I could more easily tighten them. I finally found a wrench that would fit inside the Uni-Strut and allow me to properly tighten the bolts on my "X" & "Y" carriages.

I would take a careful look at the fasteners (bolts) and think about replacing them and properly tightening them before I built a "New X" or "New Y" when it seems few others have a problem.

Ron Brown - wdyasq@yahoo.com (mailto:wdyasq@yahoo.com)

If Stupidity got us into this mess,
then why can't it get us out? - Will Rogers

03-20-2002, 02:07 PM
I just noticed this topic. I also had inconsistant results when I first recieved the machine until I accurately set manual stops at the zero end of both x axis. Remember that anytime you motors are turned off just a slight bump will easily move the square off an 1/8" or more.
After booting up my shopbot each morning I unplug the x motors & pull the gantry to the calabrated stops at the end of the x axis. Then pull them tight against the stops one by one & plug them in. Once connected it would take 60 to 80 lbs lateral force to knock them out of allignment. This is usually OK for the day unless a bit grabs an aluminum part at the end of a cut, jams and breaks the bit. This seldom happens but if there is a question I just start over manually squaring the y axis to the stops at the end of the x rails & then since I have proximity switchs I run the automatic zero x & y routine to set my x & y zero. This all takes about a minute and I am up & running.
In the beginning I resented having to do this "manual" routine but now it is just habit and I feel confident that my machine will cut accurately every day.
To calibrate the stops, I would by trial & error drill 4 holes approx 47" appart with the shopbot in a square configuration in the coverboard and move the stops slightly to adjust and re drill the square pattern. When the diagonal mesurement is equal, the setting is then square.
From then on it is just a matter to pull the carrage up the the stops & replug the x axis motors before starting.

03-20-2002, 02:21 PM
Larry, there is no need to unplug the motors - just type MO and they switch off electrically.

03-20-2002, 06:26 PM
Gerald Thanks for the tip. Just goes to show that I understand how a plug works better than the shopbot software

03-21-2002, 09:53 AM
I have found the wondering whether the shopbot is square or not can drive me crazy. I put self adhesive tape measures on the rails and mounted magnifying lenses from a mitersaw stop to read the measure accuratly. Any time you don't know, you can stop the machine and read the measure on each side. It will allow you to confirm movements and I have found it very helpful.

12-04-2002, 05:29 PM
Hi Folks. Got another puzzle. It goes like this. I just (re)squared my SB. With the x axis motors hanging, I roll the gantry the full length of my PRT96 without issue and it is squared up. All rollers are fully seated on the rails. Then I engage the motors, use the K command to manual drive the gantry back and forth in the x axis direction. Now the back left roller starts to travel up ontop of the rail at around +30.00".
It consistantly rides up at this point and will come back down at a position of <30.00". Only the one roller is acting like this. The back right roller remains engaged the whole time. Anyone have any thoughts on this. Of course, any cuts I make beyond the +30.00" has a change in the Z depth.

12-05-2002, 12:51 AM
Perry, what happens if you let the front x-motor hang while you drive the gantry up and down? (Wondering if the back motor has enough "power" to fully drive that end. . . . . ) And vice versa . .

When the roller lifts up, is it still in the center of the rail? ie. Does is hang in the air, or is there still contact with one flank of the V-roller? (Wondering if your x-rails are really parallel . . . . but everything is fine with hanging motors????)

12-05-2002, 06:06 AM
Gerald, when the roller in question is acting up, one flank of the v-roller starts to run ontop of the rail. This has happened enough for the rail to become scored at this section. Discussion with a few friends of mine at my other job suggest that this roller in question may infact be loose and the spring tension of the motor support is causing it to climb up on the rail at the same point (maybe a nick in the rail is initiating this?) I am going to verify this after I get home.
Any other thoughts?

12-05-2002, 07:33 AM
In addition to what your friend suggested, my thinking is that your x-rails cannot be equidistant and parallel all along the length of the table. Could it be that the mounting of your x-rails to the table is too flexible (loose maybe?), so that the motor spring tension causes your x-rails to twist inwards/outwards? (The rails are very flexible, particularly for twist, and your table must keep them in place)

Suggest that you accurately measure the distance between the rails, with a guage rod/stick, and see if they are changing position as you apply the motor spring pressure.

12-05-2002, 06:19 PM
Thanks Gerald. Sometimes you are to intimate with a problem to see clearly. Prior to moving my PRT96, I drilled and dowelled all pieces with the thought that the assembly would be a matter of just redowelling to set it back up exact. Well somewhere in transportation (I am assuming) one of the rails go hit/moved on the 8" C channel it is mounted to. I never bothered to check. (Duh!!!!). The long and short of it is, Gerald, you were right on the money. I went back to the manuals, loosened all the rail bolts and re-aligned everything. Just tried it out and it now seems fine. All rollers remain in contact 100% of the time.
Once again reinforcing my logic in doing business with this company rather than some other.
Thanks again.

11-05-2003, 03:28 PM
I am assembling my PRT at the moment. The beautiful blue beast! I'm working to get it square and all, like has been discussed in this posting. What I would like to get from you guys is your procedures to square the machine up after assembly. My X carriage got knock a little out of square and, I've got the Y on top of it. Can I just lossen the bolts and bring it back into square? Or did knocking the X out of square knock my Y out as well. I'm just looking for helpful hints and trick to bring it back to square. I'm enjoying putting it together. To bad that the weather outside isn't helping more with heat, it was 30 degrees last night and I still don't have heat in the garage. But hey, my comuputer processer stays cool.

11-05-2003, 11:02 PM
Squaring the X car can be a real bugger. What I like to do is take a T-square and mark a line in pencil in the Y direction. Then take a small pointed V-bit or ground to a point RotoZip bit, chuck it in the router and center it on that mark. I jog from the edge of my spoilboard in the Y direction and see how far the point of the bit deviates from center of the pencil mark.

Once you see which side of the mark needs to be favored, clamp the X car on one side of the machine. Do this on BOTH sides of the stepper motor on the ground rail to keep it from moving +X or -X. Then UNPLUG that X motor. Make sure that you clamp it tight and that you clamp the rail on both sides of the motor so that it doesn't move. Loosten the bolts up on the X car. Then tell the bot to move in the X direction whichever way it needs to in order to square it up. It may be something like MX .125 or whatever distance...Jog the Y car down the pencil mark and see how it looks down the entire line. If good, then tighten up the bolts and you are set...If not, keep tweaking.

That will get you pretty close. After that, buy a cheap sheet of Luan plywood for $10 and make a big 90° triangle and measure it for squareness and 3,4,5 proportions. (3X in the Y, 4X in the X, and 5X in the XY Hypotenuese ~ IE - Make a move 30 in the Y, 40 in the X and the XY measurement should be 50)

Hope that helps,

11-06-2003, 06:47 AM

The magic smoke may leave the board.


11-06-2003, 07:40 AM
If you are going to clamp a side of the machine, make sure that the V-roller doesn't carry the clamp force. You may press a dent into your rail, or worse.

11-06-2003, 07:47 AM

I have seen that magic smoke leave motors before. And thay never seem to work after it's gone. But how do thay get it in there to start with ?

11-06-2003, 07:56 AM
I forgot to add that part...Bill Young already advised this up the thread a bit...

Yes, don't put the clamp on the actual X car. Just clamp the rail. Unless you are King Kong, I doubt that you are going to dent the rail with a clamp.


11-06-2003, 08:58 AM
Look at my photo in my profile.