Chopper,

You are correct and I phrased the faster comment incorrectly, the steps per inch will be relatively close. The motor turning faster as it does on a 7.2:1 minimizes the resonance that is created. Even when making minute movements the faster spin minimizes the stepping (bump, bump, bump) vibration that you have with 4:1. I have always read about it but until I actually heard the difference when I installed the 4:1, the audible sound difference was incredible. Right from Oriental Motors website "Stepping motors at low speed produce a relatively high amount of vibration" Logically it adds up. The step per inch for the 4:1 is around 1700 and the steps per inch for the 7.2:1 is around 3100. It is night and day on the ears and even the feeling of the vibration in the spoilboard. With all 7.2 there was no vibration. Now I would not trade the 4:1 on my Z for anything as my cut quality increased exponentially but the resonance I could definitely do without.

Is that a better explanation?

Well I figured since the belts would be here tomorrow, I would get everything mounted on the machine. All plates fit nicely however I forgot to account for the manual release of the pinion. Have to machine out one place on each bottom plate. Will try to accomplish that tomorrow so I can get them assembled and post pictures of the assembly and mounted units.

it sure is...I understand what you mean about the noise from the motors,
it hasnt been much of an issue yet, looking forward to seeing your drives in action.
//chopper

All drives changed over, steps per inch set, ready to run. I will try to make a movie for you to hear the difference between the 4:1 and 7:1. Just as quiet as before. Some pictures for your viewing pleasure.

http://farm9.staticflickr.com/8232/8...2b957e02a0.jpg
Before Assembly, on Flickr

http://farm9.staticflickr.com/8109/8...a2132631a3.jpg
Side View Assembled, on Flickr

http://farm9.staticflickr.com/8389/8...84870ef87f.jpg
Top View Assembled, on Flickr

http://farm9.staticflickr.com/8239/8...f7b9ffddb0.jpg
X axis on machine, on Flickr

http://farm9.staticflickr.com/8518/8...7a0ff8412d.jpg
Y axis on machine, on Flickr

http://farm9.staticflickr.com/8093/8...0590f470b0.jpg
B axis on machine, on Flickr

Spring project on the way. 4th Axis Start. Need more money!!!!! Don't we all. Still need tailstock. Most likely will use 80/20 1530 for the bed and 6824 linear bearings with brakes to hold the tailstock. Looking at a Sherline 3702 for the tailstock.

http://farm9.staticflickr.com/8103/8...d0c17296_z.jpg
4th Axis, on Flickr

Man, im loving your projet with 4th axis. I am at the very end of my assembly and cant wait to get my hands on this project also.
Keep us posted, and awsome work.
Marko

Marko, I always try to post what new things I am trying.

Here is the video with a program running so you can hear the differences of the different reductions. It is uploaded in 1080 if you so choose. Sorry youtube does not provide a 1080p link only 720p so you will have to change it there. It is 3 minute worth so you can distinguish the drives. I gave a nice mention about Mechmate at the end.

720p youtube link here

360p youtube link here

Pete - Thanks for that.

A great learning experience to see them in action, plus I just love things taken to the extreme :)


Regards
Ross

No problem

Still cuts. One thing I did notice. When changing directions in cuts before there was a lot of vibration you could feel in the machine base, now it is almost not even noticeable. I was not expecting this. Here is a picture of another horse sign.

http://farm9.staticflickr.com/8383/8...aa3dd6e4_c.jpg
First cut after upgrade, on Flickr

Just some information.

Start of design for modification needed to crossbearers for mounting 4th axis. Looking at a max size of 12" (304.8mm) round capacity. It will have to span the length of X and require relocating the braces. I would like the capability to cut an 8 ft (2438.4mm) piece if I needed. This would require the both the A axis and tailstock to be beyond the spoilboard for maximum cut capacity. Option 2 is to have both inside the spoilboard area and limit the cut capacity to 5 ft (1524mm). Will make all the recess mounts prior to and based off a jig that has drill bushings in it. This jig will also be used to drill the existing crossbearers. It will have a 9" (228.6mm) support on each side to bolt onto the existing crossbearers to keep the alignment. Welding would cause draw and shift the existing alignment of the crossbearers so I am electing to stay with my current bolt together design. The A axis is built using aluminum and tapered bearings. The 6" chuck when we faced and mounted the chuck was within 0.0005" run out. This is more than accurate enough for wood I believe. I know I will have to use the machine and indicator to make sure I have the bed lined up properly when installing it.

Right now I am working on some other signs and getting ready to start cutting the kitchen cabinets. That is the only reason I built this machine was for the cabinets. Funny how things change. :D

Let me just say it sucks when someone hacks your email account. Eight hours of changing and update emails and passwords on websites. Vectric was easy to work with on the problem.

Pete sorry to hear about the hacking of your email account :(
On another note it has been a joy to follow you on your build.
Very exciting and informative and Im sure we all appreciate it.

James,
Thanks for the concern and I do try to post my progress. Seems like most guys here do this for a living whereas I do this for a hobby. I like to post the great things I have done as well as the dumb things that outnumber them. Until someone shuts me down I will continue to display my progress.

One of the things I have learned is information is hard to get on most things. Here at mechmate.com this is not so prevalent.

A lot of people are really good at what they do but do not like to share while others do not have a clue and like to lead other in the wrong direction because they like to post a lot. Some newbies are very discouraged because of this. I have always said there is no right way but many ways to get to the end solution. A smart person would listen to all of them, debate them and form a solution based upon all the relevant information.

Update

Was cutting out a larger project and ran into the problem with the steppers all of a sudden losing massive steps again when in rapid to tool change (more than 3"). I have had this happen before and found the solution then was to output single toolpaths. Thought I would try to get to the bottom of this time. The only thing I found that cleared it was to move my velocity down to 400 and set acceleration to 40 on all axes. I have no idea why worked but it did. Have posted a mach's forum searching for answers. Had to modify the layout of the sign to accommodate. Thank God I was there when it messed up or I would be out $60 in wood.

Physics always has an answer

Sir Isaac Newton said,
Force=ma, m is mass & a is acceleration.
Since mass is constant, When you set your acceleration more than the motor can provide, you will loss steps.

Power =WT, W is angular speed, T is torque,
When you set your speed too high & your motor can only provide that much torque, hence you loose steps.

Ken,
So why when the velocity was lower and the acceleration was lower I still lost steps? The version of Mach I have did have a bug with when Z was set different than X and Y it would have problems. I need to upgrade to further test. Before when all axes where the same I did not notice this. This first occurred when I changed the Z drive over to belt drive and the velocity and acceleration was different than X and Y. Before the change, I was 500 and 50, after that I changed the motor tuning. Since it appeared to be X and Y I lowered the acceleration on X and Y to 25. It stopped for a while but when there was a lot of rapids I found this reappearing. I then set all axes to 400 on velocity and it still did it. So I set all axes acceleration to 40 and now it has stopped. So I lowered V and increased A.

If its still loosing step, it is either you didn't set your acceleration low enough or you didn't get enough torque to overcome the total summation of external forces which you need to overcome. OR, you have too much external forces which you don't want acting on your motor, they can be generated from anything. such as stiction, friction, overweight due to extra super feature which you add on the car... misalignment... etc etc.etc...
IMHO, if Mach has a bug, everyone would had caught it, it won't be exclusive to you...
Mind you, the belt reduction box do contribute to the total mass which your y motor has to overcome.
BTW, "If there is a lot of rapid,..." this is the sure way to see loose steps when the motor is just barely making enough torque/power to keep up with your machanism combo. SO, lower the rapid speed & acceleration further.
Another way to improve speed/torque is to raise your PSU voltage. but without the measured motor characteristic curve, its not always gonna work.
IMHO, if you must run "rapid" in lightning speed, go back to basics, strip off all unnecessary weight from the car & see what happens.

I read your posts on the Mach3 forum and the discussion there is far from conclusive.
If you suspect a bug then it would always be best to include a build number.

You have converted small steppers to do the job and I believe that you are now seeing limits to your design.
The issue is that as the stepper motor RPM increases the torque rapidly falls off.
The 7:1 belt reductions cause the motors to spin even faster for the rapids putting them into their lowest torque on the motors overall torque curve.
Given the substitution for smaller motors over the perferred ungeared 34HS9801's and the increase in the belt reduction ratio you have received a double hit.

The solution will be to slow the rapids down to a point where the motor RPM can cope with the weight it is asked to push around.
My opinion but of course others may have their own and hey I might even be wrong :).

Ross

Ross is absolutely right. & I overlooked the fact that Pete actually using Nema24 motors... that explains all his problems... expecting a stock scooter motor to power a NASCAR & win.... hmmm.....

I do like how this was happening before and now since I changed the XY config it is the belt drive. Someone must have missed this post. Now that is funny.

I did see that Mach likes the acceleration to be withing the 8 to 10 % range of the velocity or it might cause problems. This was on a systems builder website. Why is that? That is were I got the idea to change from 25 to 40 and it went away. It has run almost 12 hours straight know without problem. I would think the higher acceration would cause it to happen even more because it increases the Gs. I agree it the build design is limited but if I wanted more then why not servos. Back to my original question. Why would lowering the V and increasing the A work. Especially when the calculated Gs increased from .06 to .1 ? I do not know that is why I am asking.

Ken,
I thought NEMA was designated by flange size? This is a hard one to find information on. All the motors are 3.54 and one is 3.35. Would they not be a Nema 34 not a 24. Nema 23 if I am correct is about 2.4 inches round?

I did find this on Oriental's site and it took some digging. Good reference for anyone who needs it. The pdf link is at the second post.

I was assuming NEMA 34, it was the motor length and by inference the magnet pack size that I was referring to.

Pete it is possible to increase the G's as long as it remains within the available torque. At the start lots of torque is available so as long as demand is not exceeded all is ok. Getting to the travel speed slowly or quickly will make no difference if it is the lack of torque at the travel speed itself that is at the heart of the problem.

The 8 to 10% estimate only holds for high travel speeds. Mach is totally happy to accelerate at 50% of a slow velocity. So the rule is just a generalisation not a commandment.

I checked the torque curve and it rapidly disappears after 200 RPM.
Suggest thinking in terms of containing and calculating your RPM rather than acceleration and velocity. Note that the curve is for the motor with the gear box in place. Changing to the 7:1 increases the torque but pushes RPM along into the lower part of the curve so their are trade offs. The curve is also not linear so rough estimations may not prove accurate.

I love your work, your workmanship and the fact you have chosen to experiment beyond the standard configuration. You should be rewarded with beautiful resolution but the only down side I can see is the trade off for some speed. To me this is a small price.

I hope this discussion adds to the mix rather than detracts in any way.
My intention is simply to understand.

Regards
Ross

Pete,
are you loosing steps on both x and Y?
or just on one?
are you loosing steps on the slaved access and the X at the same time?
or just on one of them?
the reason I say this is I have had this happen before, after running my machine for a bit after it was built I had to go back and check all the connections for the motor wires, I had a few loosen up where they attach to the motor drives.
do you use quick connectors between the motors and the machine?
I had xl connectors on my machine and removed them, ( motors are wired directly to drives ) once I checked the wires and removed the connectors all was well and I have not had an issue since. things loosen up and move especially on a new machine, also check your computer to see that all programs except mach are removed from the computer, if something stats up in the background like anti virus you will have similar issues
I would think one of these would be a more likely issue since you were having the issue before you switched to your new design. also check you kernel speed maybe you need to raise it ( data not fast enough ?) also raise the debounce.
this could also be a bad proxi if you are running them.

//chopper

Ross, that makes a little more sense. Nice explanation.

Chopper,
The action would be better for me to video than try to explain. More times than not, it will start with the Y then immediately the X will stall. There is a very distinct low hum that happens then the Z will stall. That is with the one program. With it, the movements never continue and mach will not index to the next line of code.

On the other programs, it depends on which linear direction it is moving the farthest. If X is longer than it is X, if Y is longer then Y. With those programs it will make the same low hum then resume. The only good thing if there is one is that it will affect the Z height by losing the position as it is lowering. At least this way the machine is not breaking the tooling. When the action is occuring it will lock one of the axes and the other one can travel for a long ways. I have seen anywhere for inches to feet. This has happened with both the SG7.2 gearbox and the 7.2 belt reduction.

The benefit to the belt reduction is the lack of backlash.

Ross,
I don't worry so much about speed. I spoke with a gentleman from Texas a while back. As he said, "You are not in a race with the guy down the road, it is the quality and care put into the part you make that will make the difference not that you were done cutting 30 minutes faster." This statement rang the Bell of Truth to me. I also agreed with him that when you start giving a little on accuracy, both accuracy and precision drop. For anyone reading this, here is wikipedia's explanation of them. Once you go down the path of giving up something, it will increase at an exponential rate and your end result will ultimately be affected. I try to strive for perfection but realize it is a goal few if any ever achieve.

Two thoughts: with greater Accel, you may be passing through a portion of the torque curve that has problems more quickly. Also, you might try metering your power supply to see if there is a dip that correlates with your stalls. I'm poking at insufficient capacitance on that one.

Thanks, Brad. I will check that as soon as I finish this last sign. It may be a couple days but I will post what I find.

Pete,
Ross & Brad answers most part of your questions. & I agree with the the Texas gentleman Chopper spoke to.
but again, on general terms,
Power = Angular speed x Torque, you can only go so fast with the available motor torque & power.
the tricky part with stepper motor is its torque Vs rpm characteristics, its not linear & you gotta work within the motors' capacity.
Other than the torque, power, speed, voltage, impedance & current rating. The rotor inertia is also important factor in motor selection. General speaking, when the rotor inertial is too large, it can not stop fast enough (Muscle car), when the inertial is too small it can not rev up fast enough (drifter)...
With your lighter motor, I don't see why it can not work, as long as you play with the speed+Voltage combo which the motor is happy with.

In general, if the motor runs cool, push the voltage up (limited by what the drive will handle). Yes, that does need a different transformer, but it is well worth it to get the higher performance.

Gerald,
My motors do run cool. The room they are in never get above 60°F so that adds to the lesser heat of the motors. I may have to move up in the size of the power supply

What voltage are you running on your motors Pete

35vdc

Pete

That seems small. I have a 56 VDC.

Nils,
I often wondered but it ran fine until I changed the settings from being the same. I was debating about a 35 to 42 in the beginning but opted for the 35. I spent time reading the pros and cons about voltage and really enjoyed reading the debates between Richards (Mike) and Gerald about the reason why one and why not the other. I opted for the lower based on conversations with Mike. I cannot complain. I am sort of tied up with finishing a sign but I am hoping Saturday to try and isolate it.

According to calculations I am pushing harder than the calculated maximum rpm but I had done that before with many different size steppers at work. But they have had a distinct purpose, this position to that. But I have also seen running maximum voltage on a stepper cause some of our 42's to be so hot it seems you can fry an egg on them and most definitely burn your hand.

It may be a bug with Mach and it may not. Bugs are like recalls to me, affects some people but not all. And some never take in for the recall and never have a problem. It is definitely interesting but it is working now at the current settings. I would like to change the settings back to 500 and 50 this weekend and dry run the problem programs and see if the problem still exists. If it does not, I will upgrade to Mach 57 and try again. If it still does not happen, I will lower the Z setting and see if it rears again.

In the end I may need to upgrade the power supply, only time, testing and further investigation will prove this out.

Pete,

I have a background in software. We had to test stuff all the time to find the solution. That is my current state. I have finally gotten close to fixing a noise problem with my spindle. Then today, I cut a piece of 0.5" Acrylic with lots of pocketing. None of it turned out right. Pockets had high points and low points and it was a mess. Fortunately the customer could use what I gave him but it will cost me.

Later today, I created a cut file with some pocketed circles. I used a scrap piece of plywood, I cut the file in four different places. Each one of the pockets came out correctly. I was thinking that my new Z was not square to the table but after cutting the circles, I don't think that is the problem.

Tomorrow, I will cut part of the customer job out of plywood and see what kind of results I end up with. For me it is the only way to begin to understand why things don't cut right.

Good luck with you problem.