Saturn 2 4x4 - Nema 23

[PLJack]

Jack,

Excellent! Did you use regular shim stock (like in the photo below), washers, a combination of the two, or something else? I'm thinking ahead for when my Saturn 2 arrives.



Just curious, but did you drop the rack & pinion drive and move the gantry by hand to see how it felt? There's nothing like touch to disclose small irregularities. It sure looked and sounded good, though.

If your off a few thousands in gantry perpendicularity, you can correct it when you tram front to back. Not sure what you using to tram, but I can highly recommend this:



https://www.amazon.com/Edge-Technolo...dial+indicator

I've tried various methods, and found the Pro Tram to be the fastest and easiest.
In my order of setup, it's time to surface!

Gary

Hi Gary.

I used only the shim stock you list. Cut a bunch of U shaped shims to slide around the bolt.

Definitely need to drop the pinon to shim the gantry. With all the bolts loose I moved it around by hand. Super smooth as you you think. Then after each shim or bolt turn I moved it again by hand.
If you are tweaking the bearings you can feel and hear it pretty easily. I ended up rolling back a lot of times when I felt the bearings starting to resist.
Oddly enough the tension bolts also have this effect. I dropped the pinions twice to move the gantry by hand convince myself that was happening.
Not a big deal but I was not expecting that. Makes sense really.

Also had to do a redo because I was not paying attention to the squareness on the Y rail.

I might as well write it down while its fresh, In the end I did...
Square the X rail off of the bearing shoe.
Verify other side.
Tighten a back and front screw that have the closest gap to the shoe. Square the X rails off of the Y rail.
Shim as needed to tilt the gantry up or down.
Verify other side.
Push and pull the gantry down the Y rails by hand after every change to verify there is no binding.

That was about it. Takes a while but not that hard if you do it one step at a time.
Took me a couple of hours to figure that out. Only took about thirty minutes once I had a plan.

If you don't remove the gantry it should come pretty close from the factory.
My issues here are self inflicted. I had no choice but to remove the gantry.

Jack.

[PLJack]

First square cut. 80mm.
Diag should be 113.137.



Not perfect but I think I can try some facing. Seemed to run fine. I might experiment with motor tuning and speed first. I did try moving the gantry at 6000mm/pm but the right pinion disagreed.
The faster I can go the quicker the facing will go.


Jack.

[difalkner]

Looking better. Now try a full size square, 48x48, and measure the diagonals. You don't actually have to cut into your spoilboard the entire distance. You can do 4 right angles, one at each corner, and measure the diagonals off of those. Or you can use a V-bit and just put a dot at the 4 corners.

David

[GME]

Another option is to use a 1/4" bit and cut one hole in each corner. Then, insert a 1/4" metal or wood dowel in each hole and measure the diagonals. Metal dowels are better. Since you're interested in the relative difference between the two distances, you can use something like like shown in the photos to snug up to the inside of the dowels along the shorter distance, then move the gauge to the longer distance and use feeler gauges to measure the difference in the two. Easier on the eyes than a tape measure or steel rule, and more precise. There are links below the pictures.



Veritas® Bar Gauge - Lee Valley Tools





Veritas® Bar Gauge Heads - Lee Valley Tools



(Woodpeckers OneTime Tool - No longer available).


Gary

[difalkner]

I agree with Gary in that inserting dowels would be far easier for measurements purposes. I bought some 1/4" x 1" ground steel dowel pins for that purpose but decided to just engrave the spoilboard.

But I was also interested in the actual measured distance of the diagonals. When I first cut a large rectangle and measured the diagonals I could have moved the left side of the gantry back or the right side up. Which way, I pondered... Well, I went back to Fusion 360 to see what the actual measurement should be and it turned out that measuring from the right lower corner to the back left corner measured exactly what Fusion 360 said it should be (X+Y- to X-Y+). So that told me to move the right side of the gantry.

So yes, getting the diagonals the same is important. But equally important, if not more so, is making sure your machine is cutting the exact size and dimensions it should be.

David

[PLJack]

Thanks guys. More good advice. I like the four peg idea. I think I will do that next.

I just did an interesting test I thought I should share having to do with motor tuning.
Last night I heard the right pinion complain for a sec. Like a skipping sound. I checked the squareness of the X rails in regards to the Y shoe and sure enough it was way off.
So I dropped the pinion and re-squared by eye. I've been suspecting that my acceleration was too high for these pinions.

I've been running at a velocity of 4000 and acceleration of 1000 because if I went any faster the pinions would have an issue with it.
After some testing I settled at a velocity of 6000 and acceleration of 700. Jogging back and forth there were no issues.
So I ran one of my peg trimming jobs. Air job. No spindle. The pinions now moving faster did not sound right so I kept running the job adjusting the tension down.
Eventually I just dropped the pinions and adjusted both to 1 1/2 tuns from seating. Ran the job again and the pinions sounded great.
Before I had to make those really tight. At least 2 1/2 to 3 turns.

Anyway, I figured that although they sounded good I probably just introduced a lot of backlash. So with Y set to zero I put a dial indicator on it and ran the job again.
The job goes the full length of Y back and forth for about 3 minutes. When it finished I ran Y back to zero and was stunned to find the indicator back at zero.
I'm going to have to do some more testing and tweaking of the pinions but that was encouraging.

The takeaway is that if you really need to reef on those tension bolts your issue might be somewhere else.

Edit: Just ran the job again without making any changes. Came in at 3/4 thou. Definitely on the right track.

Thanks
Jack.

[GME]

Thanks guys. More good advice. I like the four peg idea. I think I will do that next.

I just did an interesting test I thought I should share having to do with motor tuning.
Last night I heard the right pinion complain for a sec. Like a skipping sound. I checked the squareness of the X rails in regards to the Y shoe and sure enough it was way off.
So I dropped the pinion and re-squared by eye. I've been suspecting that my acceleration was too high for these pinions.

I've been running at a velocity of 4000 and acceleration of 1000 because if I went any faster the pinions would have an issue with it.
After some testing I settled at a velocity of 6000 and acceleration of 700. Jogging back and forth there were no issues.
So I ran one of my peg trimming jobs. Air job. No spindle. The pinions now moving faster did not sound right so I kept running the job adjusting the tension down.
Eventually I just dropped the pinions and adjusted both to 1 1/2 tuns from seating. Ran the job again and the pinions sounded great.
Before I had to make those really tight. At least 2 1/2 to 3 turns.

Anyway, I figured that although they sounded good I probably just introduced a lot of backlash. So with Y set to zero I put a dial indicator on it and ran the job again.
The job goes the full length of Y back and forth for about 3 minutes. When it finished I ran Y back to zero and was stunned to find the indicator back at zero.
I'm going to have to do some more testing and tweaking of the pinions but that was encouraging.

The takeaway is that if you really need to reef on those tension bolts your issue might be somewhere else.

Edit: Just ran the job again without making any changes. Came in at 3/4 thou. Definitely on the right track.

Thanks
Jack.

Jack,

Sounds like you're making good strides at getting it dialed in.

It sounds like the gantry being out of square was causing some binding issues under power, and that was causing your problems with the pinions. Makes sense. However, I wonder why you aren't able to run a whole lot faster than 6000mm. I've been running at velocity=600 ipm, or 15,240mm, for a couple of years with no issues. I have NEMA 34s. It's true that I'm presently running a kit machine, but the mounts on my machine are like yours, although without the upgrades, and mine has a ratio of 3.2:1 versus your 3:1. I doubt any of that would make any difference.

Nate advertises rapids at 1200+ ipm for NEMA 34s and 750+ ipm (19,040mm) for NEMA 24s for the Saturn 2.. I've never tried to run at 1200 ipm, or even 750 ipm, for that matter. 600ipm is plenty fast enough for me. Going twice as fast would be scary fast, IMHO. I should add that I'm running the steppers Nate sells. I liked the idea of lower inductance motors and have been very happy with them.

I wish I had some sage advice, but I'm stumped. Maybe going back to a tighter spring tension?

Gary

[wmgeorge]

I run mine with NEMA 34s at 500 IPM and acceleration is not set that high either I think around 60 or so but I would need to look to make sure.

[GME]

I run mine with NEMA 34s at 500 IPM and acceleration is not set that high either I think around 60 or so but I would need to look to make sure.

Good point. I checked my setup. My acceleration is set at 50. I'll bet that's the problem. Maybe it be worth a try for Jack to lower the acceleration and increase the velocity?

Gary

[PLJack]

OK guys, you got me curious and forced me to fire up the machine and take a video.
(spoiler, apparently I can)



Looks she does do 12000 at least. With a higher acceleration that would have totally jumped a gear. Say you are coming along on a horizontal line on a square and then start heading down vertically. That is where it was failing. I could reproduce sometimes with the arrow keys but in a run it would go badly.
The setting of 50 is way to slow. Once I letup the arrow key it keeps going for a couple of seconds because of the slow deceleration.

But with the new acceleration setting and the nice and , not so tight, tension it seems much happier.
I ran an air job of the facing with the 6000,700 settings and it went pretty well.

I don't know why this is the case for me. The numbers I shared in the previous post were just some numbers I hit upon.
Thanks to your suggestions it looks like I have some room to play with here.

Of course there is no load on the Y axis yet. When the bit hits wood I may need to go another 1/2 turn or so.
On a side note I noticed in the previous pictures that my tension bolts are different than yours. For starters its a hex head, not a hex drive.
They may be shorter as well. All I know is that up to now that spring had pretty much all the air twisted out of it. With the 1 1/2 turn I have now I can see the spring move under heavy torque situations. Pretty sure that is it's job.

Edit, also some testing earlier today showed that I could introduce about 7 thou difference on Y by moving the gantry back and fourth about ten times and returning to zero.
Assuming that is backlash.

Thanks.
Jack

[difalkner]

Just for comparison I'm running 600 ipm and 50 on acceleration with 637 oz. in. NEMA 34 steppers.

David

[DDgitfiddle]

Good point. I checked my setup. My acceleration is set at 50. I'll bet that's the problem. Maybe it be worth a try for Jack to lower the acceleration and increase the velocity?

Gary

For the tension on the pinions, acceleration is definitely more of and issue than the top speed. The tension keeps the pinions from separating from the rack. The more torque you apply, the more the pinion tries to lift off. The more acceleration you ask for, the more torque you need.

I am running a max speed of 400IPM with acceleration at 100 IPS^2, so I would think that you should be able to dial up the acceleration and speed more with your motors.

Is it possible that the noise you are hearing is related to the motors, not the drive mechanism? Could you be losing a step for some reason (electrical noise, bad connection, insufficient step rate from the controller, etc.)?

-Robert

[DDgitfiddle]

With the 1 1/2 turn I have now I can see the spring move under heavy torque situations. Pretty sure that is it's job.

Edit, also some testing earlier today showed that I could introduce about 7 thou difference on Y by moving the gantry back and fourth about ten times and returning to zero.
Assuming that is backlash.

Thanks.
Jack

That sounds too me like the springs are still too loose. You shouldn't see them compress under heavy load. That is a sign that the pinions are pulling away from the rack. It may not lead to backlash when you come to a stop, but it will make things alot less accurate as the system is moving (squares not square, circles egg-shaped).

-Robert

[GME]

Looks she does do 12000 at least. With a higher acceleration that would have totally jumped a gear. Say you are coming along on a horizontal line on a square and then start heading down vertically. That is where it was failing. I could reproduce sometimes with the arrow keys but in a run it would go badly.
The setting of 50 is way to slow. Once I letup the arrow key it keeps going for a couple of seconds because of the slow deceleration.

Could it be that your very heavy gantry is overpowering your NEMA 23s? I'm guessing that [let's call it coasting] isn't a problem with your X axis, since the stepper isn't pushing much weight. If it isn't a problem with X, it lends some weight to my supposition. In any event, this is the first time I've run across a situation where the gantry was coasting after the run command stopped. I'm getting over my head on the topic of acceleration/deceleration, so I will eagerly await others' explanations of what's going on, and what to do. Just glad I haven't had to deal with it.

Are you out there Gerry?

Gary

[wmgeorge]

There should be no coasting. When steppers are not running they should be locked and holding. I think you have issues either with your power supply or stepper driver board.

BTW there is no way my machine is going to run at 1200 IPM and I am not sure where or how those numbers came from.

[PLJack]

Hey guys. I need to head out but real quick, there is no coasting. The steppers were under control of the gantry the whole time. It just that with such a slow ramp up (acceleration) to speed it also has a slow ramp down (deceleration). I know that if I manually entered some gcode it would slowly ramp up and ramp down and hit that number just fine.
I've seen that behavior on my small desktop machine as well. If you change your 50 to 5 I would think your would see the same thing.
Maybe it's a Mach3 thing? I'll list some numbers on that screen later today.

Thanks.

[ger21]

The setting of 50 is way to slow.

One of you is using inches/sec/sec, and the other is mm/sec/sec.
25x difference in acceleration.

[wmgeorge]

Hey guys. I need to head out but real quick, there is no coasting. The steppers were under control of the gantry the whole time. It just that with such a slow ramp up (acceleration) to speed it also has a slow ramp down (deceleration). I know that if I manually entered some gcode it would slowly ramp up and ramp down and hit that number just fine.
I've seen that behavior on my small desktop machine as well. If you change your 50 to 5 I would think your would see the same thing.
Maybe it's a Mach3 thing? I'll list some numbers on that screen later today.

Thanks.

Just wondering why your using mm's instead of Inches. I use mm on both my lasers.... as thats the way they came and I do not have an issue. But inches makes it a lot easier for me to do things for customers.

[PLJack]

Just wondering why your using mm's instead of Inches. I use mm on both my lasers.... as thats the way they came and I do not have an issue. But inches makes it a lot easier for me to do things for customers.

That is a good question. The answer is that I CAD most of what I mill and I find that designing in mm to be a lot faster and the math is less complicated.
Although I do have to convert back and forth every now and again.

Jack.

[GME]

One of you is using inches/sec/sec, and the other is mm/sec/sec.
25x difference in acceleration.

Gerry, thank you for helping out. Gerry nailed the problem, as usual. My apologies for not making all of the right conversions. In previous posts, I was looking at the acceleration numbers and treating them as though they should be the same whether working in inches or millimeters. As Gerry tactfully pointed out, my assumption was completely incorrect.

I used an acceleration calculator to compare the effects of inches vs millimeters. An acceleration rate of 50 in inches (my setting for NEMA 34s) equates to 1270 in millimeters. Jack is using 1000, which seems reasonable to me for NEMA 23s. I don't think this is a Mach3 thing. It's just a matter of recognizing the need to make the conversion and using the right formulas.

Speaking of Mach3's quirkiness, FWIW, I converted from Mach3 to Mach4, when Mach3 stopped opening up under Windows 10. I found the patch to make Mach3 to work, used it for a few days, but decided to migrate to Mach4 anyway. Having used both, I like Mach4 way better. I also found that the Mach4 setup fairly easy and mostly straightforward. I was able take most all of the numbers I needed from Mach3 settings, and although some of the settings were done differently, I did not find the conversion particularly confusing. I am using a PMDX-126 (+ PMDX-107 spindle control) with SmoothStepper ESS. The SmoothStepper setup instructions were excellent. That's good, since most settings are made in the SmoothStepper config. I'm less enamored with the Mach4 setup docs. IMHO, they are just okay at best.

Gary