Newbie- Design Stage Looking for Info

[RCXPLANES]

Hi All,
I am in the design stage of a CNC router table that will use a Porter Cable 7310 and would like to find tech drawings of that router for its mounting without buying one initially to measure.

I want to build a 4 axis system that would be belt drive with IGUS linear motion and 8020 framing. Simple and available parts. The mechanics are no problem for me but I need help with the electronic packaging and drive belt ratios.

I am thinking about the HobbyCNC PRO Drive 4 axis package with the 305 oz, steppers or even the Stepper World FET 4 Dynamo system but the motors are a bit weaker at 190 oz. I would like at least the X axis motor to be double shafted. Any better solutions for performance and budget?

Another question I have at this point is with the belt drive pully ratios and micro stepping systems. I am not well versed in stepper motor rotation from input. I would like some links of threads that discuss and educate one in these fundamentals.
For instance I own an Emission Technologies laser kit that works great. It is very accurate. It has a specific motor to drive belt system ratio but I am not certian what determined that ratio. I would sacrifice speed for accuracy and I think it has to do with microstepping at the controler but don't know for sure. I want to work in inch rather than mm as well.

Thanks in advance.
Bud

[johnmac]

I used a Stepper World FET3 board on my first machine. I think its a good value for the money, but you get what you pay for. The transistors that control the motors have a slow switching rate. That means the maximum pulses per second it can handle is lower than other boards. It can do full, half and wave stepping but no micro stepping. If you choose to use one, a setup that yields 1 inch of travel to about 4 turns of the motor at full steps would be a good starting point for good feed speeds.

John

[RCXPLANES]

John,
Thanks for the input. I am still looking at a lot of information concerning the stepping of the motor and the controlers function in that.
So far all I have really learned is that the motor itself (KL23H276-30-8B) moves 1.8 deg per step which means that 200 steps are required for one revolution. What I haven't put together yet is the relation between the CAM software G code telling the system to move say an inch and the setting of steps to call for from the controller to the motor. Still reading while working with the mechanical CAD portion of the project.

I know this is one of those ask ten people get ten answers kind of things but.... With your comments about the transistor switching rate, would you have any system recomendations at all.

Thanks,
Bud

[ger21]

The Hobby CNC board is a much better choice than the FET, and has been used by many members here with very good results. Very hard to beat for the price. However, if you want better performance, look at the Gecko G540.

Yes, most steppers are 200 steps per inch. However, most good drives use some form of microstepping, usually 1/8 or 1/10, which will give you 8 or 10 times as many steps per revolution.

G-code is basically just a set of coordinates for the machine to follow. In the control software that you use (Mach3?), you input how many steps it takes to move one inch. The control software will read the g-code, and send the appropriate number of steps to move to the commanded position.

[RCXPLANES]

Originally Posted by ger21 The Hobby CNC board is a much better choice than the FET, and has been used by many members here with very good results. Very hard to beat for the price. However, if you want better performance, look at the Gecko G540. Yes, most steppers are 200 steps per inch. However, most good drives use some form of microstepping, usually 1/8 or 1/10, which will give you 8 or 10 times as many steps per revolution. G-code is basically just a set of coordinates for the machine to follow. In the control software that you use (Mach3?), you input how many steps it takes to move one inch. The control software will read the g-code, and send the appropriate number of steps to move to the commanded position.

Gerry,
This is the kind of information I am searching for here. Experience is priceless. The control part that I was mentally missing was the setting in the software that tells the controler how many steps to take to resolve a particular distance. (Post Processor??)
I design in both Solidworks and BobCAD/CAM but have never used BobCAM to control anything. I use DeskNC for my laser and that portion was already set for the machine. It sounds like Mach 3 is popular piece of the control puzzle from what I am reading here.
I will definitely look into the Gecko because I really only want to build these systems once.

Thanks,
Bud

[RCXPLANES]

Originally Posted by johnmac ..........If you choose to use one, a setup that yields 1 inch of travel to about 4 turns of the motor at full steps would be a good starting point for good feed speeds. John

Ok so I've been studying the belt drive option and using a 10T XL pully and belt as the final linear drive, 1/2 turn is very close to an inch of travel using the pitch diameter specs. Based on available parts I have found this would require a compound drive reduction of 8:1 to get 4 motor turns to equal an inch of linear travel.

I am curious if you would still use the 4 motor turns to an inch of linear travel if you could use 1/2, 1/4, 1/8 or some other micro step?
This question comes about from observing my laser and seeing that it works very acurately at 1.5 motor turns (300 steps) per inch of linear motion.
I know there is virtually no load on the linears for a laser so I'm not sugesting that 1.5 motor turns per inch is appropriate, I'm just trying to sort it out using available parts and attempting to KISS the linear drive solution.

Bud

[johnmac]

First off, I'm not a CNC expert. I'm a DIY'er like many of us. I'm basing my recommendations on experience from my first build. I used 1/4"x20 threaded rod for lead screws and the FET3 board from Stepper World.

The 1/4"x20 lead screws have way too much resolution and the FET3 board has a low limit to the pulses per second it can handle. The result is a top speed of about 8 inches per minute. That is using full steps. If I use half steps, its even worse!

If I were to replace the lead screws with a .25 lead (4 turns per inch) I would get about 40" per minute.

If I were to replace the driver with a better quality (more expensive) stepper driver, I could get more ppi, but the screw would need to spin fast, resulting in possible screw whip.

Also keep in mind steppers have most power when they are running slow. They will stall if run too fast.

And last, If you 1/2, 1/4 or micro step, you will be increasing the resolution. In order to regain the faster feed and travel speeds, the computer and stepper driver need to be able to handle the higher ppi.

All this is where I ended up with 4 (or 5) turns per inch of travel. Even at full steps, it gives good resolution for most of us while allowing fractional stepping without loosing too much top speed. Acme 1/2x8 or 1/2x10 with 2 starts will do the job. A lot of ball screws on ebay seem to fall into this category.

Hope this helps,

John

[RCXPLANES]

John,
Thanks for sharing why your choice would be 4 to 5 turns per inch.
I have been reading a lot of information out there today. I am getting some of it to stick so thats good. Seems that if you run 4 turns per inch that only 400 motor RPM would give a respectable 100 IPM feed rate. Not certian yet if 400 RPM is a reasonable expectation so I'm still reading.

I think that I am settled on the Gecko G540 if I stay under 3.5 amps for the motors. It seems like a very well thought out 4 axis drive solution. Even though there are less expensive products available, I haven't really read about one that seems to have been as well thought out yet.

Bud

[johnmac]

I guess I should add that I'm putting my "money where my mouth is", so to speak. I have almost finished the frame work on my steel framed 36"x42"x7" router frame. It has ball screws with 2 1/2 turns per inch connected to the motors with toothed pulleys and belt for a 2:1 ratio. That's 5 turns per inch.

I also watched a lot of You Tube videos of CNC routers. Everything form MDF first time builds to some really nice machines. Some moved real slow, others nice and fast. The faster machines looked like they were "working like they should".

John

[ger21]

2-4 turns per inch is even better for a router. With a Gecko G540 and 5 start 1/2-10 acme screws, you should be able to get close to 300ipm, if not more.

[RCXPLANES]

Thanks guys,
The majority of what I'll be doing is wood or even foam with the ocasional aluminum bit here and there. While accuracy is very important, when cutting foam, you need to get some speed for sure or it just melts rather than cuts. Consider also that there will be no lead screws on this machine as it will be completely belt driven from motors to linears.
I think my laser demonstrates that darn good (0.001") repeatable accuracy can be had at 1.5 turns per inch of travel resolution with great speed too (300 IPM). Again there is only the small load of the linears for the system to overcome on that machine. It cuts very acurately at 15 IPM to 200 IPM depending on material and engraves acurately at 300 IPM all day. Clearly Robert has limited the speed in the software to stay below 300 IPM at 1.5 turns to an inch. That is only 450 motor RPM.

In reading the Geckodrive FAQ, 1200 RPM isn't out of the question for a stepper motor with a top speed threshold of around 1350 RPM so at 4 turns to 1 inch, 300 IPM is possible at 1200 RPM in rapids. Whether it's advisable or efficent is another story. it seems that steppers have a resonance problem between 300 and 900 RPM that Gecko have incorporated a damper in their drive to counter also. Another thing I can think of in the feed speed question is the acceleration / deceleration capabilities of the motors and drives.

What does all this mean to me at this point? So far there is a fairly strong case for fewer motor turns per inch to retain the motors torque and make rapid and cutting speeds higher as well as a better ability to change or stop direction of travel. Also a motor to shaft belt drive ratio of 4:1 puts the linear ratio at 2 turns per inch and removes the need for a compound reduction setup so less complexity there.

Bud

[RCXPLANES]

Well I have been busy drawing up what I think is a viable 600mm x 600mm x 100mm cutting area router frame concept. The X and Y axis are fully expandable in future tables. I started out simple but like a lot of CAD projects, with the ability to draw just about anything and import files of available products, the table and motion became a more engineered and machined part table. This seemed to come about as problems were solved.
We have a great CNC machine shop in town so I will get the machine parts quoted out this week along with necessary machining of existing available parts. I have two Igus 700 mm "W" linear assemblies so I will need to get quotes on the two remaining linears and the EZ Chain.

Here are some pictures of what I have come up with so far. It is not 100% complete but all of the very sticky opperational problems are solved and just a few details need ironing out.

Bud