View Full Version : My 2nd Steel Frame Router

12-30-2006, 01:52 AM
As my first router is more or less finished now, I am planning my next machine. First of all let me say this will be a SLOW build, I am basically limited to the electronics for the next three months, as new shed/house, but mainly shed are getting built.

Machine spec's (at this stage ;) ) :

Cutting area : 1200mm by 2400mm (4ft by 8ft)

Cutting depth : 200mm

Motion : Servo's or Steppers with encoder feedback (read on)

Linear movement : Rack and Pinion

No. of Axis : 4 axis... 4th will be angular axis imbedded into table with variable job length up to approx 2000mm

Type of construction : Steel (of course) I intend to make things a little more rigid on this machine, am looking at a base frame made from structual I-beams, with the rolling base consisting of bright steel square bar, using bearings again, but forward planning a little to keep dust/chips away from the surface of the rails. Overhead gantry will probably be I-beam as well, all beams will have gussets welded every so-often to increase rigidity.

Spindle : Interchangable, as in a quick swapover method to change between 1/4 inch and 1/2 inch shaft routers. Also planning for some kind of spindle to be able to mount as well.


12-30-2006, 02:14 AM
So far, I have started on the electronics, here are some pics, but this may and probably will get tossed out if it doesn't work.

My plan is to try to build a microcontroller based driver board, bipolar stepper, and include quadrature feedback to the driver, which will monitor and automatically correct for lost steps, by simply adding extra steps. Of course there will be user selectable error, to allow for accuracy tolerance, if it is out of bounds, stop and check for problems. Hopefully also have realtime display for amount of correction taking place, to keep user informed on whats going on.

Micro I have chosen is the Propeller chip, by Parallax, makers of the Basic Stamp. I have many reasons to go this way, I won't bore you all with too many details, it's enough to say its fast, multitasking, multiple CPU's, its cheap and did I mention fast ? Processor frequency (after multiplication) is at 80MHz, free software with great documentation.

Board was kindly routed by Andrew AKA Babinda01 on his engraver, before my router was built, has been sitting a while but it is finally seeing some use. I still have to populate the board with a few more components, namely the power supply voltage regulator and the output pins, this has been designed as a development platform, and will probably incorporate the rest of the components on one board if it all goes to plan.

I especially like the USB to serial converter you can see on the top right of the populated board... very neat (I didn't make it, it came with the Prop chip)


12-30-2006, 02:22 AM
funny how it goes ,so many guys finish their first machine and imediately start the next build, it'll be interesting to see how your build goes , your last one was a definite success
good luck dude

12-30-2006, 02:26 AM
I know I might be jumping the gun by posting so soon, as I won't really be doing much apart from the microcontroller side of things for the next 3 months, but it should keep me honest and at least make me continue with the build, that's the idea anyway.

Let me say right off I am no electronics whizz, I do a little as a hobby, but by no means an expert. The intention is to make a solid base driver and "tweak" the design in software until it works... or I abandon the lot and go buy some off the shelf components.(chair)


12-30-2006, 02:32 AM
funny how it goes ,so many guys finish their first machine and imediately start the next build, it'll be interesting to see how your build goes , your last one was a definite success
good luck dude

Cheers Dertsap, I think I am the most surprised the first turned out so well, I have a habit of abandoning projects when they get too hard, I think a lot of the credit goes to those that helped out and supported along the way... Cheers all :cheers:

I know that I now want to fix all the little design mistakes I made in the first machine, in the process probably create a whole bag of new mistakes but that's half the fun...


12-30-2006, 04:59 AM
Forgot to mention, the card in the pics with the K142D written in the middle is the same as the driver cards in my first router, it will be the guinea pig for the microcontroller.


12-31-2006, 08:46 AM
There is a saying that whatever you are doing new year's eve you will be doing for the following year. Well right now I am putting the finishing touches on the micro development board, tested the voltage regulator, and am about to start uploading some test programs to check the micro is working.

So today I have spent time with the kids, successfully worked on my micro board, played with CNC routers. Now all that is left is to talk to the wife nicely...

2007 is looking good!!!


12-31-2006, 10:53 PM
O.K. I've decided to keep editing this post to update the electronics' progress, so as not to clog up the thread with the next 3 month's efforts with little milestones of electronics achievements. I don't see any (major) problems with the mechanical build, but the electronics I am flying by the seat of my pants, and it is all I will be doing for the next 3 months pretty much.

For my American and Canadian friends, the date stamp in Australia goes Day/Month/Year, makes sense to me but I thought I might mention it to avoid confusion.

31/12/06 : fully populated micro development board, tested pwr supply and regulator
01/01/07 : tested function of micro, had to re-hash USB to TTL serial convertor connections
01/01/07 : first program download and function !!! very complicated flashing LED :cool:
03/01/07 : Connect VGA output for display.
03/01/07 : Start on software overview.
06/01/07 : First attempt at routing PCB for prototyping...success!!!
06/01/07 : Start PCB layout design and configuration.
08/01/07 : Tidied up PCB layout, created g-code, waiting for some blanks to mill.

01-01-2007, 06:50 AM
I have downloaded some source code for a rotary encoder, here is a snippet of the documentation :

''Example 1: Using a 4 MHz crystal, 8x internal multiplier, 16 encoders where the highest resolution encoders is 1024 pulses per revolution:
'' Max RPM = 4,000,000 * 8 / 894 / 2 / 1024 * 60 = 1,048 RPM
''Example 2: Using same example above, but with only 2 encoders of 128 pulses per revolution:
'' Max RPM = 4,000,000 * 8 / 194 / 2 / 128 * 60 = 38,659 RPM

I'm still a noob to the chip and its programming language, but those numbers are pretty impressive, 1048 RPM doesn't sound like that much but 16 encoders at 1024 CPR !!!

Not really planning to build a 16 axis machine (yet ;) )

So a three axis machine could have a total theoretical stepper speed of around 3842 RPM with a 1024 CPR encoder !!! Not too bad for me. I dont see the need to go so high with the encoder count for a 200 step/revolution stepper though, and I'm thinking of a simpler method, basically once the step pulse is sent, read the encoder and see if the shaft has moved accordingly, if not send another pulse, or add a number to a counter to add it at the end of a move.

Any faster and ther will be some serious torque issues with the motor, even that speed is pushing it. I will build a pretty solid output stage with the hope of a high supply voltage, to try keep the torque up...

This code I have is to keep track of the absolute coordinates calculated from the encoder and is aimed at a servo system, I think I might be able to get away with a stripped down version, utilize the pulse maps to sense motion, and correct if needed. Looking hopeful so far.


01-08-2007, 12:35 AM
Drawn up the micro board, just have to tidy it up a bit, then off to buy some blank PCB board to mill it, I have decided to have a seperate board for the opto's, encoder input's etc, kind of an interface board, seperate from the micro. When I pick out all those components I will draw and mill it prob in a couple of days. Then its on to making a "test" axis, and firing it all up...


01-17-2007, 07:09 AM
Well while milling the boards was fun, I have decided to go the SMD route for them, there are just so many things to fit onto a controller board, if I route it it will be HUGE!!!

I'm looking at double sided direct printing method, have been looking at hacking an Epson printer... like this guy :

Well I'm off to scrounge what I can from ebay, better not hack the home printer, wife might notice and do something similiar to me...lol


01-30-2007, 05:47 AM
Things are moving along nicely, rather than wafting on about the printer "modifying" experiment, I made its own thread, here :


I have decided to use the Allegro A3986 Bipolar Stepper chip, it will make life a lot easier, things like microstepping translation, shoot through protection, decay times and so on are all taken care of by the chip itself, all that is needed is to mount the thing on a board, not an easy task in itself, this thing is tiny, and connect a FET bridge and assorted bits and pieces.

I am sticking with the microcontroller interface for the step/dir pulses, the long term plan is to include encoder feedback so it would be nice to be able to slow down the pulse train in the middle of a g-code line to allow the motor that has missed a step to catch up and hopefully not ruin the part being machined. Some kind of tuning will be needed similar to a servo setup.


07-12-2007, 03:35 AM
Well reading back through this is a little funny, seeing how things change, I have been working on the electronics as mentioned in the first post, but have been looking into the world of EMC, now plan to fit servo's to the machine, with EMC doing the feedback side of things, and using "dumb" H-bridges for the drivers, keeping cost down, and allowing me to keep "spare" driver boards handy for those unexpected occasions.

Machine size has grown a little, I have (hopefully) acquired a bit of a large ballscrew, it is four metres, about 13 feet (I think) in length, and about 25mm(1 inch!) thick, and about 1 TPI. I am not going to go too silly on the Y axis size, still thinking of about 1.5 metres, this should allow me to load a standard sheet of whatever and work on it, and do the occasional bit of (longer) sign work if needed, I would like this machine to pay for itself at least, maybe make a little cash on the side.

I will not be doing much with the microcontroller side of things for the machine, at least not yet, I would love to incorporate a toolchanger, but that is a pipe dream at the moment. No fourth axis, I will build a dedicated lathe for that type of work (CNC of course :))

Our house is built Woo Hoo!!! and now onto the shed to house "my precious" lol so in a few weeks the building should be full steam ahead !!!

Will probably make my own servo drives, especially since they will basically only be H-bridges.