[JohnZ]

Almost 2 months ago I mentioned that I had basic motion but my limit switches were not attached so everything moved a touch and then stopped when the PC realized there were none. Jeremy from Probotix came over a few weeks ago and actually wired the switches for me and made some software changes.

The one hangup is on my Z axis. The moving portion of the Z axis weighs about 100 lbs, more than a few of the machines out there. I thought the 1150 oz-in motors I am using to drive everything would have plenty of power and my Acme screw on Z has only 4mm lead so i thought I'd speed up the screw. I put a 2" pulley on the motor and a 1" pulley on the screw. In the end, the motor stalls out too easily so I need to change the ratio back to 1:1. To make this work with timing pulleys as I use now, I would need to buy 2 completely new pulleys at about $80 total and I want to verify this is workable before spending that much, so I am going to try 1/4" chain and sprockets. This can be done for about $20 but I have been waiting on cut to length chain for 4 weeks so my Z has been disconnected all that time. A friend is going to give me a length of chain he has sometime soon and I should be back in business.

Andy

Andy - I am also lifting well over 100 lbs on my Z axis, but only using 960 oz in nema 34 steppers. I found that I could lift over 200 lbs on my Z axis (I physically stood on the router mount for 2x spindle weight for when I put the cast iron spindle in place). To do this I had to reduce the acceleration and top speed. But once tuned in, no problems. Note, this was both before and after a ballscrew upgrade. Before acme screw, 10 tpi, after ball screw with 5mm pitch.

You should be able to lift just about any weight, but how fast may be a different matter.

Thanks - John Z

[ameinert]

Andy - I am also lifting well over 100 lbs on my Z axis, but only using 960 oz in nema 34 steppers. I found that I could lift over 200 lbs on my Z axis (I physically stood on the router mount for 2x spindle weight for when I put the cast iron spindle in place). To do this I had to reduce the acceleration and top speed. But once tuned in, no problems. Note, this was both before and after a ballscrew upgrade. Before acme screw, 10 tpi, after ball screw with 5mm pitch.

You should be able to lift just about any weight, but how fast may be a different matter.

Thanks - John Z

Thanks for the reply, John Z.

It is able to lift, but as you say, speed and acceleration are another thing. I think I had it set at 9 IPM and 2 IPS^2 with the 2:1 pulleys and it would stall. That's pretty slow - hard to do much work when it takes a couple seconds for every quick Z retract. What do you have yours set at? Is it acceptable to you?

Andy

[ger21]

The one hangup is on my Z axis. The moving portion of the Z axis weighs about 100 lbs, more than a few of the machines out there. I thought the 1150 oz-in motors I am using to drive everything would have plenty of power and my Acme screw on Z has only 4mm lead

There are numerous other cases of this issue posted here.

A 400oz 6amp Nema 34 stepper and a gecko drive would probably be 5x faster than your big motor.

Those big steppers are slow. To get any speed out of them, you'd usually need to supply them with well over 100 volts.
Here's an example:
http://www.cnczone.com/forums/1170266-post313.html

[JohnZ]

I did a lot of testing and have performance to report using four different drivers and two different drive screws with two of the drivers.

Gecko G213V, 78V, 1/2-10 ACME Screw - 120 IPM max, ran at 90
5mm Ball Screw 190 IPM possible, 120 IPM lifting over 220 lbs

Older Kelling 8060 driver, 78V, 1/2-10 ACME - 90 IPM max, ran at 45
5mm Ball Screw 150 IPM possible, 90 IPM lifting over 220 lbs

Generic Chinese 50V 6A driver, 48V, 1/2-10 ACME - 40 IPM max, ran at 22
5mm Ball Screw 80 IPM possible, 40 IPM lifting about 60 lbs

Old "Dragon" driver, 38V, 1/2-10 ACME - 60 IPM max, ran at 40
Died before I upgraded to ballscrew

Note, the Max speeds lifted were the fasted speed I could get everything to move without stalling with just my porter cable router and mount, no additional loading. The speeds I ran the system at were where I could not stop the motion with my hands pushing as hard as I could. The lifting 220 lbs was by standing on top of the router mount to determine at what point it would stall out lifting my weight plus the weight of the rest of the moving assembly. My new spindle weighs about 120 lbs so I figured 220 lbs would be provide a good safety margin. I finally settled on the Kelling driver running at 90 IPM, and acceleration at 20 IPS - So I could use the faster driver on the X axis.

I agree that your Z-axis is pretty slow and would drive me nuts. What driver and voltage are you using? I suspect a generic Chinese driver, I was extremely disappointed in the one I got to replace the Dragon Driver when it died. Might I suggest using a Gecko driver as close to 80V as possible rather than swapping the pulleys, and be certain the screw / nut are not binding along with the motion of the Z axis along its entire travel.

Thanks - John Z

Sent from my DROID4 using Tapatalk 2

[ger21]

A smaller motor with lower inductance can give you more speed, and actually have more power at the higher speeds your trying to spin.
Also, 1/2-10 acme is only about 30-35% efficient, so your losing 65% of your power right off the top.
Switching to 1/2-8 2 start acme would double or possibly triple the speeds you saw with 1/2-10 acme.
A 10mm pitch ballscrew would probably give you the best performance.

[ameinert]

I did a lot of testing and have performance to report using four different drivers and two different drive screws with two of the drivers.

Gecko G213V, 78V, 1/2-10 ACME Screw - 120 IPM max, ran at 90
5mm Ball Screw 190 IPM possible, 120 IPM lifting over 220 lbs

Older Kelling 8060 driver, 78V, 1/2-10 ACME - 90 IPM max, ran at 45
5mm Ball Screw 150 IPM possible, 90 IPM lifting over 220 lbs

Generic Chinese 50V 6A driver, 48V, 1/2-10 ACME - 40 IPM max, ran at 22
5mm Ball Screw 80 IPM possible, 40 IPM lifting about 60 lbs

Old "Dragon" driver, 38V, 1/2-10 ACME - 60 IPM max, ran at 40
Died before I upgraded to ballscrew

Note, the Max speeds lifted were the fasted speed I could get everything to move without stalling with just my porter cable router and mount, no additional loading. The speeds I ran the system at were where I could not stop the motion with my hands pushing as hard as I could. The lifting 220 lbs was by standing on top of the router mount to determine at what point it would stall out lifting my weight plus the weight of the rest of the moving assembly. My new spindle weighs about 120 lbs so I figured 220 lbs would be provide a good safety margin. I finally settled on the Kelling driver running at 90 IPM, and acceleration at 20 IPS - So I could use the faster driver on the X axis.

I agree that your Z-axis is pretty slow and would drive me nuts. What driver and voltage are you using? I suspect a generic Chinese driver, I was extremely disappointed in the one I got to replace the Dragon Driver when it died. Might I suggest using a Gecko driver as close to 80V as possible rather than swapping the pulleys, and be certain the screw / nut are not binding along with the motion of the Z axis along its entire travel.

Thanks - John Z

Sent from my DROID4 using Tapatalk 2

Thanks a lot for the information, John. That's a huge amount of spindle weight.

I considered a ball screw for this machine but was concerned about the wood dust gumming it all up. Since I have been directed by the wife to ease up the spending for a while, I will play with the hardware I have and work on what I can change with dip switches and EMC2 settings.

I am running re-branded Chinese drivers so I wonder what affect that is having. some of the stuff coming out of there seems pretty good and other stuff, not so much.

Andy

[ameinert]

A smaller motor with lower inductance can give you more speed, and actually have more power at the higher speeds your trying to spin.
Also, 1/2-10 acme is only about 30-35% efficient, so your losing 65% of your power right off the top.
Switching to 1/2-8 2 start acme would double or possibly triple the speeds you saw with 1/2-10 acme.
A 10mm pitch ballscrew would probably give you the best performance.

Thanks for the reply, Gerry. I am now starting to see the value in posting here.

Now I am starting to wish I had paid more attention to the motor and driver details when I set this thing up. I confess that I focused on the mechanical aspects of the machine (I'm a mechanical engineer and have never really worked work with steppers and stuff before) and leaned on someone else to spec out and build the electronics and motors and so now I am weak on understanding the eletrical issues and solutions here. The post you linked to is very interesting.

I am using Probotix Mondostep 7.8 drivers, which I believe are re branded Chinese drivers. My motors are their 1150 oz-in units, bipolar, 8 wire, half stepping. As a result of your post, I looked up the manual on this drver on the probotix site and it describes how you can set the motor up in series or in parallel and it mentions half coil configurations or "half-chopping." Since I am not in a psoition to go buy new motors and drivers right now, is changing something with these motors and drivers to reduce inductance workable? It looks like this implies re-wiring the motor and I would probably get the Probotix guys to do that for me if it makes sense.

I also wondered if swicthing back to full stepping would be helpful?

Andy

[ger21]

Your motors appear to be 8 wire motors. They should be wired bipolar parallel for best performance. If they are wired bipolar series, it shouldn't be difficult to change them to parallel.

What is your power supply voltage? For best performance you'd want to use about 72V with bipolar parallel. Wired bipolar series, you'd need to use about 150V to get the most out of those motors.
Be aware that bipolar parallel will double the current requirements. For 3 motors, I'd recommend a 12-15amp power supply.

I'd set the microstep setting to either 8 or 16. FUll step is not recommended, as it usually results in resonance and rough running.

Those drives appear to be the same basic chinese drives sold by a lot of vendors.
They are not as good as Gecko's or the digital chinese drives, but they are not bad.
The real issue is that your screws have too fine of a pitch, and the motor is just too big and slow.

[ameinert]

I really appreciate your help, Gerry.

I actually have 5 motors because the gantry is so long that I have one on each end and I will have a 4th axis when I get done, so I bought the whole thing at once. I have 2 40V 10A power supplies. So it sounds like voltage is probably an issue for me, as well.

I thought going with more steps than 2 would actually make this situation worse. Am I misunderstanding that?
Andy

[ger21]

Yes, microstepping allows the motors to run smoother, which will usually allow them to run faster without stalling. You don't really need to go higher than8, or possible 16.
One thing to be aware of, though, is that as microstepping goes up, you need to send a lot more steps to the drives for a given speed. You may get to a point where the PC becomes the bottleneck, and not the motors and drives.

If they're not already, I'd wire the motors bipolar parallel, and you should see an immediate improvement if they are wired bipolar series.

[ameinert]

So, I spent the last couple of weeks of (very limited) spare time looking at this problem with my Z axis being so slow. I verified that the motors are wired parallel, bi polar and now I actually understand what those both mean. The best I was getting out of my Z was 18 IPM and like 2 ips^2 acceleration. No amount of parameter adjustment did any good and changing from 2:1 speed increase to 1:1 did improve the travel speed from 9 to 18 IPM but then the improvement stopped. I tried a 1:2 speed reduction (more torque, less speed from same rpm) and it made thinsg worse, which makes sense when you look at stepper motor torque curves.

I don't have the torque curves for my specific motors but I looked around and found some for similar motors so at least I could get close to a reasonable understanding of my situation. I also found some good tools on the internet to help me understand what amount of torque should be needed to lift my Z. I put all this in a spreadsheet and did some wonderful graphing of my results and charted all the EMC2 parameters and results. It all showed that I should have plenty of power to do what I am trying to do with the parts I have. (I really like the internet.) In the speed range where I am running out of motor, the curves show I should be getting 800 oz-in or above and the screw curves indicate I should need less than 400 oz-in to get that speed. Obviously somethingw as wrong.

All along, I wanted to try swapping out the motor and/or the driver to see if one of those was possibly bad. That takes a fair amount of time, but I finally got a few hours today and made it happen.

Sure enough, when I moved the X axis motor over to the z axis and made the requisite software changes, everything was dandy. I got 60 IPM velocity and 30 IPS^2 acceleration, which was like night and day to me. I tried combinations of drivers to make sure, but the driver was not a variable. I went ahead and stuck the Z motor on the x axis and got the same result on that axis - barely moved before it stalled. So, I pulled the motor back off and unhooked it. I tried to manually turn the shaft and it would not budge, even with a big pair of pliers grabbing the flats. I probably have a bad bearing. I suspect it was going bad all along and just got worse along the way because it never gave me good performance and an identical motor is doing fine. I'll get a replacement from Probotix next week.

As I was working thru all these torque curves, one thing that kept being obvious is that my 40 V power source was definitely keeping me from getting the most out of the motors. A 72 V power source shifts those torque curves dramatically. So, before I knew that I had a bad motor (I really didn't think that would be the case.), I ordered a new 72V and 20 Amp power source from Keling. It should arrive on Tuesday. It looks like it is not necessary to fix the problem, but that much additional power should definitely make this big beast scoot along. I am almost scared to think of the amount of inertia I'll be able to put into this system. I am also suddenly wondering if the wires I have going to these motors is adequate. They are 22 GA shielded cable, I believe. I nede to get a reference to show me what 7.8 amps max and 80 V requires. Longest wires are 33 feet long.

Andy

[ameinert]

Actually my motor wires are 18 GA. Misspoke.

[ger21]

18 ga should be fine.

[ameinert]

Thanks, Gerry.
I looked all over the internet for tools to help me determine that and they are all specific to AC or to 12,24,48 VDC.

[OCNC]

Thanks, Gerry.
I looked all over the internet for tools to help me determine that and they are all specific to AC or to 12,24,48 VDC.

Wire Selection

[ameinert]

Over the last many weeks I have played with ideas around managing the many cables on the Beast, especially the Z axis. I spent a while looking at the "Cable Carriers" thread and saw some cool designs. For now, I just want something that will keep me from ripping cables apart (happened yesterday) so I can start running this thing and starting to work out the kinks.

I liked the ideas I saw of using PVC pipe and stuff from the home stores. However, it seemed like something rectangular made more sense than a round pipe and I did not want to do any machining other than sawing. (I don't have a mill or anything like that.) My first experiment was with PVC gutter material. I did not have high confidence in it but it was cheap and I thought it would be a good experiment. I did some experiments as shown in the first 4 images below. I bought the standard stuff and cut it at some angles and then tried hooking it together with Gorilla Tape, which is really nice strong stuff. I also put a piece of linen material inside to try to keep the wires from leaking out of the joints a little better. The linen was pop riveted in place to also provide some stiffness on the outside of the bend so it would self-support both ways. What I found was that it worked fine on a horizontal axis (at least sitting on my bench for a couple of days with no wires in it) but the tape started to loosen up and make the whole thing sag after it had been mounted vertically for about 4 hours. The 4th picture shows this. I thought maybe using aluminum sheet instead of Gorilla tape might help, so I undid the tape and replaced it with aluminum sheet riveted to the PVC. That did not work much better - the aluminum was too stiff in some ways and had the same sagging problem as the tape once upright. I concluded that the link material needed to have more bearing surface and bearing strength in order to support itself. The gutter material is just too thin.

I went to Menard's looking for some rectangular conduit material that some guy on the other thread had, but I could not find anything like that. However, I went into the garden section outside and found this 2x4 inch composite material that is used for fencing. I think you put it over a 2x4 to protect it from the elements and make it more attractive. It seems to be made from wood chips and plastic and it only cost $6 for a 6 foot length so I bought 3 pieces and took it home.

I decided I needed a pivot of some kind and decided on a 3/8" screw bolted all the way thru. I also decided that offsetting the links of the chain was a very easy way to make them connect and support each other. I spent a while on Pro/E building up different versions and experimented with different cut angles and chain lengths to see what would work best. I ended up with a 30 degree angle and 5 inch long links. The pivot bolts are brought in a bit form the end so the effective link length is less than 5. More like 4-1/4 or something. I made up some of these and found them to work pretty well so I set up some temporary fixturing on my saws and went to work. I have one chain made up of 12 links and 2 end pieces now on my X axis and it seems to be working pretty well. I have accidentally broken a couple of the links by bearing failure so I am making a new variation that is stronger by using a wider side tab and thus increasing the bearing strength of the bolt holes.

The 5th picture below shows the finished product on the X axis. The 6th picture shows one of the links sitting on the bench next to the raw material. Note that I decided to try this on X first and then possibly implement it on Z later but this material is really too much for the Z axis in terms of size - just don't need that much cross section. I have also designed some slots to put into these chain links to lighten them since the material is heavier than it is strong. I may actually make that the first machining I perform with this machine.

My last picture shows my temporary solution for the Z. I had some old Shop Vac hose in the shop and I thought It would be better than just leaving the wires bare. My conclusion is that this material is quite adequate for this purpose as long as you leave the material intact. However, I had to cut a slit all down the length in order to get the wires in. My other option was to do lots of unwiring of things like motors, switches, and my Super-PID box and so on to be able to feed the wires thru. So, I slit the hose and it got much less stiff. I went ahead and used it anyway but I have had to use generous amounts of Gorilla Tape to make it hold together enough to work. This is not a good long term solution. I was able to pull apart the hose ends and reattach them to a shorter length of hose. I also made little wooden mounts to put them in to attach the whole affair. Interesting experiment but too cheesy for me to leave there forever.

So, I am a little embarrassed about the gutter experiment but I am sharing for the amusement and edification of others. The fence material thing seems to work well, though so I am starting to build some up for my long axis. I will probably just buy some commercial chain for the Z since it will set me back like $40 and I have already put about $2000 worth of my time into it.

Andy

[ameinert]

Thanks. I knew if I whined someone would come thru.

[PaulRowntree]

re : CableChain
What about just using a 3-4" wide piece of flat vinyl siding (or similar) and use wire ties to make loops to constrain the wiring to the siding? I think CarveOne tried and said it worked but the aesthetics were ... less satisfactory. It is hard to imagine an easier solution.
Cheers!

[CarveOne]

re : CableChain
What about just using a 3-4" wide piece of flat vinyl siding (or similar) and use wire ties to make loops to constrain the wiring to the siding? I think CarveOne tried and said it worked but the aesthetics were ... less satisfactory. It is hard to imagine an easier solution.
Cheers!

It won't cost much to implement it while you continue the search for a better looking solution. It works so well for my smaller machine that I quit looking and do more project making. Other than for looks, I have seen zero reasons to change it out.

[ameinert]

I have followed CarveOne's work a bit and like his solution - simple, effective, clean. Before I did any of the things I described here, I went to the home stores and looked for something like that, but could not find anything suitable. There was some really nice looking stuff but it was 1/4" thick or something like that and would not bend tight enough. Bend radius was around 6-8". I actually considered buying some and running it through my thickness planer but envisioned this enormous pile of plastic shavings sticking to everything due to static electricity and decided to skip it. I have a bunch of vinyl siding left over from my shop build but I have already had to use some for repairs so I don;t want to use that.

I thought about using aluminum sheet, which I had in my shop already. I cut some out and played with it. It seemed to meet the physical requirements but I was concerned about the electrical implications of running 120VAC and a bunch of low voltage DC cables along several feet of conductive material.

Andy