Saturn 2 4x4 w/ Nema34 closed-loop - In Canada

[wackocrash5150]

Dumb question but here goes, does it have the problem when using the keyboard? Just wondering if it is the hand held unit. I'm sure you already thought of that, but I figured I'd ask.

Dan

Hi Dan. That is NOT a dumb question at all. Sometimes the most obvious things point to a solution. However in this case it does it with more than just the keyboard or keypad. It's done it while running test code as well. It will start off fine and then part way through I will be looking at it thinking "hmmm .... something doesn't look right here..."

[wmgeorge]

FYI I have had Chinese made circuit boards have poor or cold solder joints and pins that have not even been soldered. I would pull the board off and carefully go over it with a good light and magnifying glass. My guess you will find one of the above. My circuit board was on a mini milling machine and it was a bugger to get out and back in.

[Ntl]

Hi Dan. That is NOT a dumb question at all. Sometimes the most obvious things point to a solution. However in this case it does it with more than just the keyboard or keypad. It's done it while running test code as well. It will start off fine and then part way through I will be looking at it thinking "hmmm .... something doesn't look right here..."

That sucks hope it didn't break anything when you were running a program.

[bhurts]

Could it be the cable causing the problem. If it happens sometimes and not others that could indicate a break or intermittent connection.

Ben

Sent from my E6810 using Tapatalk

[wackocrash5150]

That sucks hope it didn't break anything when you were running a program.

Nope. Was running an "air cut" with the code. Learned my lesson on that once or twice on my smaller CNC. Was lucky before but the Saturn 2 with Nema 34s on it I imagine would be much less forgiving.

Could it be the cable causing the problem. If it happens sometimes and not others that could indicate a break or intermittent connection.

Ben

Sent from my E6810 using Tapatalk

I thought of that, but that's where the encoders come into play. If the motor did anything that the driver didn't tell it to do, it would alarm and fault out. I'm using closed loop steppers and drivers.

[wackocrash5150]

Hey there everybody... so I've finally had a bit more time to dedicate to my machine the past couple days. After a whole bunch of troubleshooting, I did narrow down the previous issue to a faulty driver. I contacted the seller and after a couple weeks, I swapped out the faulty one with a replacement. No more issues on that front. Since then I've been able to cut all the holes in my wasteboard and spoilboard and it turned out pretty good. This brings new to a new issue that I just can't seem to wrap my head around. I know a possible cause but I really hope I'm wrong. After the baseboard and wasteboard was finished and fastened, I thought I'd move on to calibration. Steps per inch went fine on the x axis. I think I was at 2036 instead of the starting point of 2038; fine that way. Then the Y axis. One would assume that it would be 2036 as well but I thought I would check it just to be sure, but it was out. It seemed closer at 2038, which I thought was odd but entered the value. This got me curious. I move the X axis almost all the way to the right and measured the travel on the A (slave) axis. It was not the same. Very odd. Something I also noticed is that when down on the homing end (front) of the machine, when I home and then hit the E-stop to cut power to the drivers, it stays right where I left it. However if I home, travel about 40 or so inches up the Y axis and then hit the E-stop, the gantry seems to spring back about an 1/8th" or so, removing the tension on the R&P gear as I can see it pull back up and into the rack. I think that this is the same issue from back in post #5, however it just "APPEARED" to go away when I tightened the R&P spring. Seeing that the gantry travelled smooth without the R&P being engaged since day one, there was something so obvious that I never thought to check. I measured the distance between the Y and A rails at either end. They're out by just over a 1/16th of an inch. Entirely not what I was expecting given that this machine was advertised as within 0.1mm.

So my fellow CNC Zone members, what does this all mean? I'm a bit frustrated and therefore can't wrap my head around the geometry at this point to find a cause and then a solution. I should mention that I've also swapped the X and Y racks, thinking that it might be a difference in the manufacturing of the rack. Same result.

I haven't reached out to Nate at this point as I'm looking for ideas to troubleshoot before going down that road.

[Ntl]

It's not going to matter if one side is a inch longer or a 16th of a inch. The only thing that could cause a problem is if the gear rack is off and even then you just need to square up the gantry. Don't go off the hard stops since they aren't perfect. What I did was use my laser and lined it up with the linear rail on one side marked the machine on both sides, then went to the mark on the other side of the machine and verified that the marks were in the same place. Doing this gave me a real reference point to square the gantry off of you can't square it up off any table measurements since the steel fabrication that was done on these is anything but precision unfortunately. Granted I have a high end very accurate PLS laser it has a accuracy of +-0.004 per foot. Get a good 4 or 5 foot t square if you don't have a good laser since it will be even more accurate and mark the rails then see if it is square. You may need to loosen up the gantry if the teeth aren't meshing properly. Let us know what happens.

Dan

[wackocrash5150]

It's not going to matter if one side is a inch longer or a 16th of a inch.

Dan

Hey Dan. Thanks for the reply. I think I might've not explained the 1/16th of an inch properly. It's not in the length of each the Y and A (slave) rail, but rather the distance between the two rails from one end to the other. They are not perfectly parallel.

Leon

[GME]

I measured the distance between the Y and A rails at either end. They're out by just over a 1/16th of an inch. Entirely not what I was expecting given that this machine was advertised as within 0.1mm.

So my fellow CNC Zone members, what does this all mean? I'm a bit frustrated and therefore can't wrap my head around the geometry at this point to find a cause and then a solution. I should mention that I've also swapped the X and Y racks, thinking that it might be a difference in the manufacturing of the rack. Same result.

I haven't reached out to Nate at this point as I'm looking for ideas to troubleshoot before going down that road.

It appears your issue is the same as one I had on the machine I sent back. Yours just isn't as bad as mine was. When I measured the difference front as compared to the back of the right and left linear rails, it was .086" out. When I measured the frame tubes, the difference was .026" - significantly less. Here is a link to my build thread: https://www.cnczone.com/forums/finel...c-posts-3.html

To put just over 1/16" (.0625+) into perspective, the maximum allowable deviation from parallel for linear rails is somewhere between .001" and .0017", depending upon the source. The max for Hiwin is .001. So, you are way outside the range. Your discrepancy is going to throw pretty much everything off. As ger21 put in my thread, "If the machine can't flex enough to accommodate the misalignment, the bearings will wear faster, or be damaged. Or they'll just bind and not slide at all." I'm betting that the "spring back" you're experiencing is related to the extreme lack of parallelism. Here's a link to the part of my thread where the subject was discussed: https://www.cnczone.com/forums/finel...c-posts-5.html.

My Saturn 2 had so many problems, I finally stopped measuring and Nate took it back. However, if you have a long precision straight edge, you might want to run it along the bases of the rails and look for gaps. It could be that the holes for the linear rails weren't drilled in a perfectly straight line. You already know that the rails are not parallel, but for diagnostic purposes you might want to see of they are bolted down in a straight line. Maybe yes, maybe no. If they aren't, you have yet another layer of of problems. You may also find that the distance between the tubes is different (out of spec). It could matter, because the tubes are dished in the center (like most steel tubing). If the rails were mounted parallel on out-of-parallel tubes, there well still be a disrelation - one rail could be flat in the center of the tube on one side, while the other rail is angled slightly on the other side (it ride on the side of the dish shape). Think about checking the steel frame tubes to see if they were welded square, plum. There is a reasonable probability they weren't.

Keep in mind that one problem often leads to one or more others, and errors can become cumulative (depending upon the nature and cause).

I suggest you may want to read more of my thread that just the links I posted. Maybe it will give you some ideas.

As the end of the day, I concluded that trying to fix my machine wasn't for me. The problems were more than what is contemplated in satisfaction guarantee. What I was sent was defective when built and far from the tolerances promised in the advertising. To Nate's credit, he gave me no hassles and did the right thing. I'm sure that my taking the time to photo document the issues made it easier for both of us.

Gary

[Ntl]

Oh so you mean the span over x axis, I thought you were talking about the y axis. Dumb question but you're measuring in between the rails that the bearings ride on and not the steel tube that they bolt to right?

If so I would pop out the plastic bolt covers the easiest way is to use a screw and a drill and just screw it in the middle of the plastic plug and they pop right out and then have Nate send you new ones. Then drop the motors and loosen all the bolts move the gantry to one side and tighten the bolts as you move the gantry to the other side the gantry will line up the rails if there's enough clearance in the holes of the linear rails with the bolts. If there's not enough clearance you could always either use helicoils and smaller bolts in the area that is off or you could bore out the holes. I'm not sure if any of this makes any sense? If not let me know and I can try to explain it better.

Dan

[GME]

If so I would pop out the plastic bolt covers the easiest way is to use a screw and a drill and just screw it in the middle of the plastic plug and they pop right out and then have Nate send you new ones. Then drop the motors and loosen all the bolts move the gantry to one side and tighten the bolts as you move the gantry to the other side the gantry will line up the rails if there's enough clearance in the holes of the linear rails with the bolts. If there's not enough clearance you could always either use helicoils and smaller bolts in the area that is off or you could bore out the holes. I'm not sure if any of this makes any sense? If not let me know and I can try to explain it better.

Dan

Loosening both linear rails and trying to use the gantry to get them straight probably won't work. The technique assumes linear rails are perfectly straight. Often, the are not.. That's one reason why it's recommended that an edge be milled at the mounting site. The rail is then referenced off the edge and any irregularities can be worked out as you work down the length. Here's a link to the Hiwin PDF that covers all manner of info. Mounting starts on page 14. http://www.file-vault.us/pdfs/hiwinguideway.pdf. There are various options listed, but none of them involve loosening all of the fasteners and using the gantry to line up both sides simultaneously. What they all have in common is using precise techniques for alignment. For my newest machine, I will be using precision straight edges (.0002 in 12") for alignment. Not the preferred method, I don't have the option of milled edges.

I've never seen helicoils used where there would be so little thread engagement. The walls of the frame tubing aren't all that thick. I'll defer to you, if you've successfully used them in tubing of this sort. However, there is another issue worth considering. 30mm rails use an M8 cap screw. The head is typically about 13mm. The screw hole size is 9mm, so there is 2mm head contact area all around. The next size down is M6. The head on an M6 is 10mm, so the head contact area would only be .5mm all round (25%). For my taste, .5mm is way too little. Also, note that in terms of adjustability, there is only .5mm (.0197") of adjustment on a centered cap screw. Not nearly enough to correct a 1/16" (.0625) misalignment. Leon has more than 1/16", although how much more is unclear. Specs on rail parallelism allow for a deviation of .001 to .0017, depending upon the source. My Saturn was worse. I couldn't find a reasonable way to correct it.

I have no idea what Leon plans to do, but he didn't get what he bargained for. If he takes the time to take measurements of other areas, he might find other things out of spec. My frame was welded out of square/plumb. I cannot help but wonder whether something like that is contributing to his issues. Anyway, I've been there. I sent mine back.

Gary

[wmgeorge]

As I pointed out before the design of the Saturn 2 does not allow for the parallel re-alignment of the linear rails as its shipped. The CNCRP design appears to be fully adjustable. But if you had the linear rails of the S2 mounted on a 3/8 plate with slotted holes cut for parallel alignment and that rail plate could also be shimmed up and down if needed that I believe would work.

Gary says the linear rails are referenced off the frame rail, I do not think that's true. I think the frame rail is just a mounting surface.

My machine is fine but it all depends on how close the frame can be welded to true, and the placement of the rails when mounting.

[wackocrash5150]

So, just for the sake of trying to wrap my head around it, am I to understand that the rails being out of parallel from approximately 1/16th to 3/32nd of an inch over the length is indeed enough to cause the racking that I'm experiencing?

Humor me and let me speak hypothetical here on a solution. So the lack of any adjustability if it's not set up perfectly upon assembly would seem to be a design flaw. Here's the what-if:

ASSUMING that it was just and only a matter of the rails being out of parallel due to human error when drilling and tapping the holes for them, what if it were possible to make a hole in each end of each rail support tubing, allowing a bar to be inserted within the square tubing. This bar would be roughly the same length and width of the rail itself and would have threaded holes that precisely match the holes of the rail. The threaded holes currently in the square tubing could be then drilled a bit oversized to accommodate adjustability. The screws would go into the rail, through the tubing and thread into the inserted bar below, essentially sandwiching the square tubing Make any sense? Again, assuming that it was ONLY misalignment of the rails.

[ger21]

Yes, that would work.

It's not a design flaw, as there's no easy way to provide adjustability. It's a manufacturing defect.

[wmgeorge]

Sure that would work and not effect the height of the Z like my idea would. You could lift the gantry high enough to unmount the rails one side at a time or maybe just slide them out as is. Then drill and tap those holes needed in the new flat steel bar. Drill out and elongate (if needed) the old mounting holes. In fact you might only need to do one side. Nether one of my 3" square tubing mounting rails have caps anyway.

If you were 100% sure that in the future you will never move those linear rails, you and drill a hole and put ether a roll pin or tapered pin to hold any movement.

[wackocrash5150]

Nether one of my 3" square tubing mounting rails have caps anyway.

Where I would propose the hole would actually be the tube directly under the linear rail. The ends are what the hard stops screw to. Borrowing PLJack's image here:

https://www.cnczone.com/forums/attac...d=403344&stc=1

[GME]

So, just for the sake of trying to wrap my head around it, am I to understand that the rails being out of parallel from approximately 1/16th to 3/32nd of an inch over the length is indeed enough to cause the racking that I'm experiencing?

Misalignment can definitely cause binding, and if the misalignment is bad enough, limit your ability to move the gantry at various points, depending upon where the gantry was located (relative to the front and back) when it was tightened down. Whether it definitely the cause of your problem is an open question. The problem your having seems consistent with misalignment, but other things could also cause the problem, line the frame rails being welded out of plane. Diagnostics can be difficult in person; much harder remotely in a setting like the zone.

Humor me and let me speak hypothetical here on a solution. So the lack of any adjustability if it's not set up perfectly upon assembly would seem to be a design flaw. Here's the what-if:

ASSUMING that it was just and only a matter of the rails being out of parallel due to human error when drilling and tapping the holes for them, what if it were possible to make a hole in each end of each rail support tubing, allowing a bar to be inserted within the square tubing. This bar would be roughly the same length and width of the rail itself and would have threaded holes that precisely match the holes of the rail. The threaded holes currently in the square tubing could be then drilled a bit oversized to accommodate adjustability. The screws would go into the rail, through the tubing and thread into the inserted bar below, essentially sandwiching the square tubing Make any sense? Again, assuming that it was ONLY misalignment of the rails.

I don't see why your hypothetical wouldn't in theory In actual practice, I'm not so sure.

If you think about it, it is similar to how it works with 80/20 extrusion machines, only you are using a solid bar versus a t-nut in an extrusion slot. However. there are differences. With 80/20, you have a flatter surface to work from than you have with steel tubing. So, you have a better chance of establishing a good reference edge. Since the tubing on the Saturn is dished, getting a good reference edge (straight edge?) is more problematic. Getting to straightness may be a significant limiting factor. Also, there is the question of the surface against which your bar will rest. On the inside of the frame tube, the surface is convex. That means the bar will have a tendency to rock from side to side, due to smaller contact area than you would have on a flat surface. Hard to predict how that would work out. With Bill's suggestion (wmgeorge), at least you would be spanning the edges of a concave surface. It would probably sit pretty flat, but could distort with tightening. If you put a bar on top, you would probably want one thick enough that the chances were better that the concavity would straighten out, rather than the add on rail.

Because straightness of the rail mounting is also very important, you would need to ensure that the holes were drilled and tapped in a straight line - straighter than you can likely do by hand. That will take a mill. I believe you would probably need a setup with at least two matched vises, because most mills won't have enough Y axis travel to drill and tap the holes on one fixed pass.

I considered the same options. Not worth it for me. Your mileage may vary.

The discussion assume your premise, i.e., that the only problem is with the holes not being drilled and tapped straight and/or parallel, and there is nothing else contributing. I question whether that is the case.


Have you fully loosened all of the gantry mounting bolts and used feelers to ensure that the gantry sits perfectly flat on the interface plate? If it doesn't sit flat, tightening everything down will also stress the bearing blocks/linear rails can, by itself, cause problems, including some pretty nasty binding. The gantry may seem to move okay, but still be binding. IMHO, you need to check everything - gantry, frame rails plumb and square - everything - before even thinking about how you might fix the machine. Too much is interconnected. Admittedly, I had the tendency to think, "how to I fix this" or "how do I fix that" but finally came to the conclusion that I had to take a more holistic approach to understand what I was truly up against. For me, the best starting point is the frame. If the frame wasn't welded plumb and square, every fix will probably result is a significant compromise. Anyway, looking at segments in isolation will inevitably lead to the question, "but what about . . . ?"

Gary

[Ntl]

That would work great. Good idea for not over complicating it, you could even use flatbar if you wanted to.

Good idea

[he1957]

But why should you have to fix a machine that was designed and built so you could buy it and start using it?

Kit or no kit, a kit needs to be designed and made to allow for home-shop/user assembly and adjustment to within published/advertised design tolerances without horribly expensive or possibly impossible to build-it-yourself alignment jigs and fixtures and so forth. Let alone re-designing and modifying it.

Just saying...

Send it back.

A$0.02

HarryE.
===

[wackocrash5150]

But why should you have to fix a machine that was designed and built so you could buy it and start using it?

Kit or no kit, a kit needs to be designed and made to allow for home-shop/user assembly and adjustment to within published/advertised design tolerances without horribly expensive or possibly impossible to build-it-yourself alignment jigs and fixtures and so forth. Let alone re-designing and modifying it.

Just saying...

Send it back.

A$0.02

HarryE.
===

Unless it was an easy fix, I would normally agree with you Harry. However in my case, I did not have the machine shipped. I drove from about an hour north of Toronto Canada to Lebanon Pennsylvania to pick the machine up. I encountered a small hassle coming back across the border with it but nothing serious. It was about a 9 hour trip each way, but in the end I saved around $1000 by transporting over the border myself. I am sure that return & replacement is an option as that is what the site says and another has taken Nate up on this, however I view this as close to a last resort. Admittedly, I did not check the rail distance onsite. The gantry did roll smoothly and made an assumption that also meant parallel. In fact it still seems to roll smoothly when the rack is not engaged. Not quite as smooth as the X axis but it also has considerably more weight on it. I rolled with that assumption all through the stepper motor and controller issues that I've only recently resolved. It was only over the past few days while trying to calibrate that it became more apparent something was up.