[marzetti]

Thank you kwackers for sharing this project with all of us!
I have been reading this thread with great interest thinking it may well be the solution I was looking for my application. What I am seeking is a simple single axis controller for positioning a fence on various woodworking machines. The idea is to be able to simply input a distance with the keyboard and move the fence to that position with either a lead screw or timing belt drive. I'm hoping this controller could do this without too much difficulty simply by rewriting the code. Is this doable with this design? (or better yet, has anyone already done it??) I'm prepared to dust off my programming knowledge and tackle it if this is the right platform for the application I have in mind. Ideally other functions such as incremental positioning, preset values, etc. could be implemented also.
Any input or pointers would be welcome and if I can get this project started I'll start a new thread.

[kwackers]

Hi Kwackers,
Thanks to share its ideas.
I have a Vertex of ratio 90:1, but it is possible also to be used with small rotary tables of ratio 36:1. I want to buy one.
Thanks and greetings.
Victor

Hi Victor,
The worm ratio can be set from 1:1 to 999:1 so it will work with your 36:1 no problems.
Obviously you won't get the same resolution as you would with a 90:1 table, since with a 400 step motor the best resolution will be 360 / (36 * 400) = 0.025 degrees (giving an error of half that or 0.0125 degrees).

regards.

[kwackers]

Thank you kwackers for sharing this project with all of us!
I have been reading this thread with great interest thinking it may well be the solution I was looking for my application. What I am seeking is a simple single axis controller for positioning a fence on various woodworking machines. The idea is to be able to simply input a distance with the keyboard and move the fence to that position with either a lead screw or timing belt drive. I'm hoping this controller could do this without too much difficulty simply by rewriting the code. Is this doable with this design? (or better yet, has anyone already done it??) I'm prepared to dust off my programming knowledge and tackle it if this is the right platform for the application I have in mind. Ideally other functions such as incremental positioning, preset values, etc. could be implemented also.
Any input or pointers would be welcome and if I can get this project started I'll start a new thread.

Hi, I've used the current system "as is" for something similar - basically with a custom thread and appropriate settings you can get it so 1 degree = 1mm. The main issue is that it thinks it can 'wrap around' so that 360mm is shown as 0mm. It's not ideal, but it can be done.

Alternatively it would be possible to re-write the software to make it a linear system (originally I started writing this to make 'box joint's). The only issue is one of resolution - the software holds the position as a 16 bit number and this would need extending to 32 bits to hold a decent range.

I've written a new 4 axis engine specifically to do linear stuff. At the moment it's temporarily on hold whilst I decide whether to stay with 8 bit pics or move to 16 (or 32) bit. I'm not sure if the functionality I want can be built into an 8 bit version...

Regards
Steve.

[torquemada]

Hi kwackers,
Thanks for its fast answer and its aid.
Not if it happened with this ratio 36, just like to the Flavioluiz friend with the 72, that it did not show the calculation in some steps well.
Very many thanks, steve, will try when the table arrives.

Regards
Victor

[skullworks]

This looks almost exactly like what I need.

I'm looking to build a rotary table to replace a Haas 5C programmable indexer.

I/O is simple: Cycle start (input) - Cycle complete (output) E-stop as an Aux input.

[marzetti]

Alternatively it would be possible to re-write the software to make it a linear system (originally I started writing this to make 'box joint's). The only issue is one of resolution - the software holds the position as a 16 bit number and this would need extending to 32 bits to hold a decent range.

Is this a limitation of the software or the chip? Would it require a 16 bit PIC to hold the 32 bit number?

[irving2008]

Is this a limitation of the software or the chip? Would it require a 16 bit PIC to hold the 32 bit number?

Software only, it should mainly be redeclaring 'int's as 'long's and playing with some areas of calculation. The length of the machine word has no bearing on the size of the operands it can process, it just takes more machine cycles, which isnt a limitation here...

[mpascino]

First off, I wanted to say I love this design and thanks for sharing. I have assembled all of the components with mild sucess. The problem I am having is I don't get past the "Rotary Table Divider" screen on boot. It just hangs there. I have loaded the key test program and have validated the keypad is functioning. I do not get a beep when I press each key but the screen icons respond appropriately. The passive sounder I am using does not show polarity and operates on 3V P to P. I assume I need to find a straight DC buzzer, correct? I assume this is not related to the freezing issue I have, does anyone have any ideas? The Caps I am using for the crystal are 33pf 500V, might they be too big? Any help is much appreciated.
Thanks,
Mike

[kwackers]

First off, I wanted to say I love this design and thanks for sharing. I have assembled all of the components with mild sucess. The problem I am having is I don't get past the "Rotary Table Divider" screen on boot. It just hangs there. I have loaded the key test program and have validated the keypad is functioning. I do not get a beep when I press each key but the screen icons respond appropriately. The passive sounder I am using does not show polarity and operates on 3V P to P. I assume I need to find a straight DC buzzer, correct? I assume this is not related to the freezing issue I have, does anyone have any ideas? The Caps I am using for the crystal are 33pf 500V, might they be too big? Any help is much appreciated.
Thanks,
Mike

Hi Mike,

It shouldn't make any difference - but just check the limits are pulled high (pin 7 & 8 at 5v), check the sense input (pin 2 is at 0v).

The caps you have around the xtal are fine and should work ok - if it works with the keypad test code then there's no problem.

Worth programming another PIC and trying that - also if you can it's worth trying another programmer. I've had a couple lately that one of my programmers can't program - they verify ok and all looks well but the code doesn't run properly and can hang.
My gut feeling is a PIC/Programmer problem...

Regarding the sounder it sounds like it should be ok - you definitely don't want a buzzer - I can't remember but it may be that the keypad test routine doesn't sound the beeper...

Hope this helps.
Steve.

[Borken]

Hi, and thanks for an outstanding design!
I built one and after struggling a bit with the keyboard it now works properly!

I connected it to an homemade dividing axle with 72 tooth main sprocket and 10 tooth on the stepper motor. It never accured to me that this would be a problem later on when setting the wormratio. The stepper driver is a linistepper capable of halfstep 6:th and 18:th microstep. IS there any way to work around this? Or do i have to change the main sprocket?

[kwackers]

Hi, and thanks for an outstanding design!
I built one and after struggling a bit with the keyboard it now works properly!

I connected it to an homemade dividing axle with 72 tooth main sprocket and 10 tooth on the stepper motor. It never accured to me that this would be a problem later on when setting the wormratio. The stepper driver is a linistepper capable of halfstep 6:th and 18:th microstep. IS there any way to work around this? Or do i have to change the main sprocket?

Hi, if I've understood correctly then I'm afraid you may be out of luck...
Internally the device assumes a 400 step per rev motor (i.e. 200 step half stepping).

So with a 72 tooth sprocket driven from a 10 tooth one you've 7.2 turns per 360 degrees - i.e. a 7.2 x worm.
Even multiplying it up by your 6th and 18th (i.e. x3 and x9 - since we're already assuming half stepping) gives 21.6 and 64.8.

Since the device will only accept an integer value then anything that results in a decimal number won't work.
(It never occurred to me that someone would use a non integer worm ratio!).

You may also wish to lower the gearing - microstepping is (ime) rarely linear, half stepping is about as good in terms of accuracy as it gets, so assuming you're half stepping then you've 7.2 * 400 = 2,880 steps per rev or 360/2880 = 0.125 degrees per step (accuracy +/- 0.0625 deg). Is this accurate enough for your purpose??

Hope this helps.
Steve.

[Borken]

hmm, bad luck.
I have to hunt for another sprocket then, one with even number of teeths.
It would have been accurate enough.
Maybe in some distant future you will put in some decimals in the setup.

Regards
Janne

[kwackers]

hmm, bad luck.
I have to hunt for another sprocket then, one with even number of teeths.
It would have been accurate enough.
Maybe in some distant future you will put in some decimals in the setup.

Regards
Janne

Hi Janne,

It would be a fairly trivial software change to change the "Worm Ratio" section in the setup to "Steps Per RPM".
This would allow any ratio within reason to be entered - from say 360 to 65535 steps per RPM.
The cost is simply readability - I'd assumed most people would use 400 steps per motor rpm and an integer ratio.

If I went this route then you'd have to enter a number which was calculated by: Motor Steps Per RPM X Ratio

So for a 90:1 worm and a 400 steps per rpm motor that would give a value of 36,000

In your case 72/10 * 400 = 2,880 (assuming half step motor).

I'll have a think - if nobody objects I may change it to that...

[Borken]

Would be great!!

/Janne

[kwackers]

Hi Janne,

I've attached a new .hex file, you can try if you like.
As mentioned above I've changed the 'enter worm ratio' into 'enter steps per RPM'.
The number is the number of motor steps required to do 1 complete revolution of the table and is calculated by the worm/gear ratio multiplied by the steps per revolution of the motor. In your case the 72/10 * 400 = 2,880. (If you're half stepping).

I've only given it a rudimentary looking over so there are no guarantees. For this reason I wouldn't recommend anyone else try it.
I don't bounds check the input values - very small numbers are likely to cause odd behaviour.
I've also tweaked the jog code to allow the +/- 0.01 degree jog buttons to jog a minimum of 1 step, this is because with small ratios such as yours the minimum they can move is 0.1 degrees. This way the 0.01 degree button will always move 1 step instead of doing nothing.

Regards
Steve.

[Borken]

Worked like a charm.
Thank you very much!!

/Janne

[TOTALLYRC]

Hi Janne,

I've attached a new .hex file, you can try if you like.
As mentioned above I've changed the 'enter worm ratio' into 'enter steps per RPM'.
The number is the number of motor steps required to do 1 complete revolution of the table and is calculated by the worm/gear ratio multiplied by the steps per revolution of the motor. In your case the 72/10 * 400 = 2,880. (If you're half stepping).

I've only given it a rudimentary looking over so there are no guarantees. For this reason I wouldn't recommend anyone else try it.
I don't bounds check the input values - very small numbers are likely to cause odd behaviour.
I've also tweaked the jog code to allow the +/- 0.01 degree jog buttons to jog a minimum of 1 step, this is because with small ratios such as yours the minimum they can move is 0.1 degrees. This way the 0.01 degree button will always move 1 step instead of doing nothing.

Regards
Steve.

Hi Steve, and not to be picky but shouldn't it have been steps per rev, not steps per rpm.


Nice work by the way.


Mike

[kwackers]

Hi Steve, and not to be picky but shouldn't it have been steps per rev, not steps per rpm.


Nice work by the way.


Mike

Hi Mike,
Had to go away and look - actually it says "Steps per 360".

It did originally say "steps per rev" but I discarded that because it sounds more like it has to do with the motor rather than the system as a whole...
Not sure why I wrote "steps per rpm" - senior moment probably. ;-)

Steve

[Beevo]

A friend and I are building each one of the controllers. When we powered up the first one, we go many question marks. Garbage if you will. On power up, the display seems to change, pressing keys seems to have an effect on the display and we can hear faint key beeps when a key is depressed.
Any ideas appreciated.

http://www.valley-metal.org/scambled%20display_std.jpg

Marty

[kwackers]

Hi Marty,

By your description the device is working - but it sounds like the display isn't being driven correctly.

Check the lines between the PIC and the display, check continuity and also check for short circuits between the lines.

Steve.