Please welcome... "Contraption 1"

[JRoque]

Hello all. It's been a few months since any update to this project so here's a pic of the new control box. I will be taking more pictures of the remainder in the next couple of days.

I started with an old PC mini-tower case. It looked the right size and certainly larger than the previous one I had but it got crowded in there very quickly. Take a look at the picture below and follow the numbers for a description of the parts:

1. Stepper drives from Applied Motion, 3 of them side by side. I run these at 36V.

2. Case fan PWM switch. This plugs to the microcontroller (#5) and there's another one by the drives to switch the spindle fan as well. The MOSFETs can switch up to 60V@4A. Both circuits are optoisolated.

3. Case fan, 12V from an old Compaq server. The spindle fan is the same model as well.

4. USB, PS/2 and RJ45 connectors salvaged from this same PC. The front panel (on a separate pic) show these connectors

5. Microcontroller. This is an Atmel Mega8 chip running at 16Mhz. It controls the speed of the fans using temperature feedback from a couple of 1-Wire DS1822 sensors. The temperature threshold of when the fan start spinning is programmable. The speed of the fan is set according to how far off the threshold is the current temperature. Current temp is read every 15 seconds. The fans, once started, will remain on for at least 3 mins to prevent frequent start/stop conditions while floating around the threshold temp. There's also an alarm output to a front panel LED to warn about over temperature; a buzzer will be added later. This threshold is also programmable. All programming is done via USB port and a terminal program. USB facilities are via SiLabs USB interface chip. This will also serve as the interface to a pendant in the works.

6. Parallel port breakout. The parallel cable is too long now and will be replaced later on

7. A/C outlet, not currently connected. Possibly switched with SCR or relay - still thinking about it. Underneath this plug are the 4 plugs to the axes and spindle. More on this on a separate pic

8. 250 watt PC power supply. Source for 5V and 12V

9. Optoisolator I/F to the spindle controller. This takes the PWM signal from Mach2 to set the spindle controller speed.

10. Minarik spindle motor controller. Spec'd output is 130VDC @ 10 A

11. Linear voltage regulator for 3.3V LED used on spindle. The LED is a 10MM, 35000 MCD white LED.

JR

[JRoque]

Here's a couple more pics. The first one is the front panel. I didn't have any other material so I used 6061-T6 .125" thick.. a bit too much for panels. It has switches for the main high-voltage DC supply and a switch to power each drive individually. This is handy for those times when you need to push one of the axes by hand. There's also puch-out ports for USB, RJ-45, PS/2 and a couple more switches. The engraving was done with a .0625" ballnose for aluminum.

Since this picture, I've added the front panel USB and PS/2 ports that connect to the microcontroller inside (see above posting). The USB port plugs to the control PC and the PS/2 port (or maybe the RJ-45) will be used to talk to the Mach2/3 pendant I'm building.

The second picture is the back panel. It has Switchcraft 9 pin plugs for all three drives plus a 4th one for the sensors and gadgets on Z. On the Z cable I run the spindle and stepper motor drives. On the X and Y cables I run both the stepper and limit switches. I was anticipating some kind of interaction between the stepper drive and the low-voltage sensors and sure enough, a nice 23Khz pulse showed on the limit switch lines. This spike tripped the limit switch sensor on Mach2, though it didn't register in the "diagnostics" screen since it was such a short and fast pulse. A simple RC, low-pass filter cured it. The original plan was to run the sensors separately from the drive cable but I couldn't justify adding another 9 wire cable, plug and receptacle for just 1 sensor (I ran out of inputs on Mach2).

There's also an A/C outlet to be switched with either SCR or mechanical relay - still thinking about that. Not shown on the pic are all of the ground screws for the shields, etc. Finally, the parallel port connector that runs to the control PC and the EPO switch input.

JR

[yukonho]

Great thread.
Any word on that macro for the spindle touch off?
colin

[JRoque]

Hey Colin,

Here's the macro I'm using. I made this before I even had Mach2 so it's rough and unfinished but it works. Now that I know Mach2 a bit better I can see that it can be made differently but this one works for me. Proceed with caution.

Connect a "probe" to input2 on Mach2 and make sure the spindle/tool is grounded. For a probe, I used a piece of thin copper. Its important you confirm the tool is grounded by touching the probe to it manually. I have ceramic ball bearings and a plastic coupler on my spindle which isolates the tool from ground. This had to be fixed before the macro would worked. Check the "Diagnostics" screen while setting this up.

Cheers,
JR



'******this macro auto zeros Z axis using external trigger
'****** Inch version

'****** JRoque, July 11, 2004


'********************************
'Declarations

Public RunawayZ



'*********Get current environment:


Startfeed = GetDRO (18) 'Get starting feedrate



'********Set new environment:

ZZFeedRate = 10 'Sets Z feedrate
ZZCurrentTool = 255
ZHome = 1.0 'Set Z home to +1.000" off auto-zero
ZProbe = 0.025 'Thickness of probe
ZApproach = 0.01 'Feedrate on Z while moving down, towards the probe
ZRetract = 0.001 'Z feedrate while moving up once it has touched the probe
RunawayZVal = -2 'How far Z will travel down (inches) before the system asks: "Where do you think you're going, Mr?"


'*************************************************************************
'*************************************************************************
'*********************** *********************************
'*********************** MAIN SUB *********************************
'*********************** *********************************
'*************************************************************************
'*************************************************************************


SetCurrentTool(ZZCurrentTool)

Code "F" & ZZFeedRate


Call MoveZ

If RunawayZ = False Then
Call SetDRO (2, ZProbe) 'Make the DRO = to the thickness of the probe
RestoreSub 'Reset the feedrate to prior value
Code "G0Z" & ZHome 'Move Z to it's new "home" position
End If




'*************************************************************************
'*************************************************************************
'*************************************************************************
'*************************************************************************
'Move Z

Sub MoveZ

While IsActive(19) = False '19 means that probe is connected to input2
'18 = Input1
'20 = Input3
'21 = Input4

'Lower Z
CurrentZ = GetDRO( 2 )

'Subtract (ie: move towards probe) ZApproach to current Z position
Code "G0Z" & (CurrentZ - ZAPproach)

'Loop here while Z is moving
While IsMoving()

Wend

'Runaway safety
If CurrentZ <= RunAwayZVal Then
RunawayZ = True
Exit Sub
End If


Wend

'Z has touched probe
Runaway = False
Call Backout

End Sub
'*************************************************************************
'*************************************************************************
Sub Backout

While IsActive(19)


'Slowly backout up until the probe disconnects
CurrentZ2 = GetDRO( 2 )

'Add (ie: move away from probe) the ZRetract value to current Z position
Code "G0Z" & (CurrentZ2 + ZRetract)

'Loop while Z is moving
While IsMoving()
Wend

WEnd

End Sub


'*************************************************************************
'*************************************************************************

Sub RestoreSub

'*********Restore environment
Code "F" & Startfeed

End Sub

[JRoque]

Ok, here's a couple more pics and a link to the full description of my new spindle.

As seen from the log pictures above, I tried a Precise S65 spindle first and had ok results. The spindle is a great machine running at 45K RPM and very low runout. The problem was that it's only 1/2 HP and it was choking on the stuff I wanted to cut. I then moved to a new Porter Cable 7518 but was disappointed with it's performance. It's loud, hot and, at least mine, has .002" runout. The PC is fine for what it was designed for but it didn't work for me in my application.

So, I decided to brew my own. I started by purchasing the best and largest bearings I could afford. I then went shopping for a cylinder where to mount the bearings and ended up with a raw piece of stainless steel. After reading all I could on different tool systems, etc, I decided on a design and started cutting. The project took about 5 or 6 weeks end to end and it ran about $150 in parts.

You can read the details on this project and see how it built up to the finish product. Dimensions and other specs are also there. This is a short video clip showing the runout

Current challenge is adding a set of three wheels to control the machine manually. This has taken longer than I thought mostly due to coding issues on accurately reading the mechanical rotary encoders. More on this later and hopefully more pics on the entire machine as is now.

JR

[JavaDog]

Wow! That is pretty impressive!

SS is supposed to be a real bear to work with. That run-out looks pretty good to me, but what do I know IANAME...

[DieGuy]

Pretty cool, Where did you get the bearings?

[tauscnc]

WOW also,

Great work. I have tired many times to make percision parts and it is tough! Great job. What is going to be running it?

taus

[JRoque]

JavaD, IMHO, SS is one of the best materials to work with. Aside from the hot and stringy 'chips' that I got, it's absolutely predictable and machines great. Of course, I don't know what I'm talking about since the spindle was my second lathe experience ever - but I just love SS.

DieGuy, eBay baby, eBay 8 - ) You can still see the bid on eBay for the next few days. I later contacted the guy and bought two more.

Taus, it took a LOT of patience and help from this forum. Its definitely not perfect and I'm sure luck played a big role (judging by the runout) but it was worth it. Check out the full details on the link above and you'll see I basically kicked, bit and spat it to make it work so I'm not claiming any talent here.

Later,
JR

Oops forgot: It's being turned by a threadmill DC motor. Pretty good so far actually.