[MrBean]

I've been using Alan's PICStep drivers for around 18 months now, and not had any major issues with them. In fact, they've been really good.

I recently changed to the new V2.0 firmware (1/16 stepping), when a glitch started to appear. One of the motors (Z) was behaving erratically. I'm pretty sure I've tracked this down to a heat issue concerning the LMD's.

Is it logical, that running 1/16 microsteping would cause the LMD's to generate more heat than they did at 1/8 microstepping?

I tried to switch back to 1/8 mode for my Z axis, but found that the jumpers on the interface board no longer seem to alter the step mode??

My Setup.......

I'm running the old, non opto interface board from Alan's site (V2.0).
The V 2.0 PICStep driver hardware.
The new V2 firmware (1/16 microstepping)

Has anyone else tried the new firmware with the older interface board and not been able to use the jumpers to change stepping modes?

It looks like I'll have to make up a heatsink and fan assembly to keep them LMD's cooler.

The 1/16 mode does generally seem to make my machine run smoother. Cheers Alan....

Regards
MrBean.

[Garfield2]

Is it logical, that running 1/16 microsteping would cause the LMD's to generate more heat than they did at 1/8 microstepping?

I tried to switch back to 1/8 mode for my Z axis, but found that the jumpers on the interface board no longer seem to alter the step mode??

Ooo really? The pins are the same on both versions, the firmware is universal. Perhaps something else is broken..

Heat is generated with switching so I'd assume with the higher switching rates needed for 1/16 mode the MOSFETs would generate more heat. The LMDs need a fair amount of heatsinking.

Glad the 1/16 mode makes you machine smoother. I'll have a look at that bug shortly.

Alan.

[MrBean]

Hi Alan. I've now got some CPU heat sinks on the LMD pairs. All of the motor glitching is now gone. 1/16 mode is running like a champ. Thanks.

The jumper problem on the interface board for mode switching is odd. After all it just sets the PIC pins high/low, what can be broken? If I get time later I'll flash the old hex file back into one of the pics and see if the mode jumpers work for that axis.

It's no biggie. Now 1/16 mode is running well, I'd like to stick with that anyway.

Right now I'm milling a PCB for my wind turbine. I'll post a picture once it's done.

BTW. The thread with RadekCX and the LMD bug, was quite entertaining. Glad you got it sorted in the end.

Regards Terry.

[MrBean]

Well... I haven't had time to re-program the PICs, but I did get my PCB done. I guess PCB, should really be MCB? (Milled Circuit Board).

I know it's not that interesting, but machine was "Powered by PICStep", so I thought you might like to see the results.
I just need a sticker now saying....... "PICStep Inside"

I'll try using the old firmware regarding the mode jumpers tomorrow.

Regards Terry.

[HayTay]

...but I did get my PCB done. I guess PCB, should really be MCB? (Milled Circuit Board).

I know it's not that interesting, but machine was "Powered by PICStep", so I thought you might like to see the results.

Fantastic result on your MCB!!!

Give us some more details, please (what bit(s), feed rate, DOC, mill time, etc.).

Very, very, nice work. Can't wait to see PICS of the PICStep populated with some components.

[gmfoster]

Man that board does look first class....

Thanks for posting..
Garry

[santiniuk]

Totally awesome board result.

Interesting find on the jumper settings. I haven't had chance to investigate this yet but will be interested to see what Alan finds.

Great work.

[MrBean]

Fantastic result on your MCB!!!

Give us some more details, please (what bit(s), feed rate, DOC, mill time, etc.).

Thanks for the kind comments. I was pretty chuffed at how well the board came out.
Okay........ A few details for those the want them.
I used a 30 degree, single fluted, conical cutter.
Feedrate while cutting was ~300mm/min. These small cutters require low feedrates and high spindle speeds. At over £10 a go, I don't want to break any.
Depth of cut was just enough to get through the copper layer. I'm not sure exactly what the numbers were. It took a couple of false starts to get the depth just right.
Mill time.....Hmmm. The actual isolation traces were finished in around 7 mins, but I chose to run an off-set stepover to remove some of the excess copper around each trace. This step is not neccesary, but I like the look it gives. The more you remove the longer it takes. Figures really!
Drilling cycle was about 15 mins with 5 tool changes. I cheated and used the same size bit for all the holes, but it still lifted the Z and paused each time a new size was required. I just hit cycle start and carried on.

This is only my second attempt at this type of milled board. I still have some stuff to figure out, but it's looking promising so far.

Right.... I'm off to flash me chips, see if I can pin down this PICStep mode select weirdness.

Cheers guys.
Regards Terry.

[PsyKotyk]

The results are very impressive. What kind of spindle speeds are we talking? Like 3000-5000, 10,000-15,000 or higher? What's the size of the mill bit, or since it's conical, is it just as pointy as it can be?

I would be quite happy to get those results everytime.

Thanks for the kind comments. I was pretty chuffed at how well the board came out.
Okay........ A few details for those the want them.
I used a 30 degree, single fluted, conical cutter.
Feedrate while cutting was ~300mm/min. These small cutters require low feedrates and high spindle speeds. At over £10 a go, I don't want to break any.
Depth of cut was just enough to get through the copper layer. I'm not sure exactly what the numbers were. It took a couple of false starts to get the depth just right.
Mill time.....Hmmm. The actual isolation traces were finished in around 7 mins, but I chose to run an off-set stepover to remove some of the excess copper around each trace. This step is not neccesary, but I like the look it gives. The more you remove the longer it takes. Figures really!
Drilling cycle was about 15 mins with 5 tool changes. I cheated and used the same size bit for all the holes, but it still lifted the Z and paused each time a new size was required. I just hit cycle start and carried on.

This is only my second attempt at this type of milled board. I still have some stuff to figure out, but it's looking promising so far.

Right.... I'm off to flash me chips, see if I can pin down this PICStep mode select weirdness.

Cheers guys.
Regards Terry.

[MrBean]

The results are very impressive. What kind of spindle speeds are we talking? Like 3000-5000, 10,000-15,000 or higher? What's the size of the mill bit, or since it's conical, is it just as pointy as it can be?

Hi. I would say the spindle was running in the region of 25,000 rpm
And, yes the cutter is pretty much as pointy as you could make it.
The widest part you can see of the shaft is 3mm, the thinner part is 2mm.

Picture attached below...

[HayTay]

Okay........ A few details for those the want them.
I used a 30 degree, single fluted, conical cutter.
Feedrate while cutting was ~300mm/min. These small cutters require low feedrates and high spindle speeds. At over £10 a go, I don't want to break any.
Depth of cut was just enough to get through the copper layer.

Thanks for the updated info and the photo of the cutter!

I was pretty chuffed at how well the board came out.

Had to consult "The English-to-American Dictionary" to make sure of the meaning of "chuffed".
Yo, dawg, dat board b look'n right nice, yous gots ev'ry right ta b chuffed! Word!

[smarbaga]

what would the machine movement resolution distance be on your finest pass ?
is this a home made machine ?
do u drill at 25k rpm's as well?
a sugestion, if i may, if you drilled first there may not be as much traring.
the board looks great, and you can have one anytime, in your own home.
thanks for sharing , MrBean.

[MrBean]

what would the machine movement resolution distance be on your finest pass ?
is this a home made machine ?
do u drill at 25k rpm's as well?
a sugestion, if i may, if you drilled first there may not be as much traring.
the board looks great, and you can have one anytime, in your own home.
thanks for sharing , MrBean.

Hmmmm. Resoultion? All 3 of my axii (sp?) have different resolutions.
I guess the answer would also depend on if you count microstepping in the calculation? Resolution doesn't always = accuracy or repeatability.
Ha Ha. look.... I'm starting to waffle on now......
This subject has been discussed in depth elsewhere on "The Zone".

I'm not exactly sure what you mean by "movement resolution distance"

But here's some numbers anyway....

The smallest move my X axis can make is 0.0004 Inches
The smallest move my Y axis can make is 0.000027 Inches
The smallest move my Z axis can make is 0.000019 Inches

Yes the machine is homemade. It's just made from stuff I could lay my hands on at the time. I have a second machine about 80% built, and a third machine that I have parts for, but have not started building yet.

Spindle speed when drilling was not so fast. Around 15K rpm.


Regards Terry.

[fyffe555]

Terry,

How did you produce the G-code? How long did it take to produce the board? As the others have said, nice board.

Andrew

[MrBean]

Terry,

How did you produce the G-code?

Andrew

Ah...... Good question.

I didn't......
The computer did it.
Sorry. Just kidding... (wedge)

The board layout was done in Eagle. I'm sure you all know Eagle is free for boards up to 80mm * 100mm.

The clever bit, (G-code generation) is a script for Eagle, written and given away for free by a guy called John Johnson. The script itself is pretty easy to use. In fact, once it's setup, you would hardly know it's there. It even has a graphical front end for configuring set-up files etc. It intergrates itself with Eagle as a ULP. It's a pretty damn impressive piece of scripting.

Probably the worst part is getting all the settings just right and working well together. The design rule checker, net classes, and the settings for the gcode generation script, all have to be pretty good to get good results.

I've only just started using Eagle and the ULP script over the last few days. So really, anyone should be able to achieve these results in that time, or less. Providing you have a CNC that can cope with it.

As for the time it took to produce the board. I'll assume you mean the actual CNC milling of the board.

I didn't time it, but would estimate it to have been around an hour. That did include a lot of unnecessary copper removal though. The minimum track isolation was done in around 7 mins. I just liked the look of the board with more copper removed between the traces.
Of course, you could bring the time down from there. It depends on how fast you dare cut with those tiny cutters.

I hope that answers your questions.

Reagrds Terry.

[fyffe555]

Terry, thanks.. I'm familiar with the ULP... or I thought I was! Is the Eagle Cadsoft ULP version John Johnsons's? I was wondering how did you got it to clear quite so much copper. You prompted me to find JJ's PCB-GCODE yahoo group and the ULP has step feature now. nice, the version I tried a while back didn't. Instead I've been using the old ULP to drill and cut to shape first, then using the holes to align toner transfer and etch on both sides at once. Tedious. Your results are so good I think I'll have to try milling once again. Thanks for the lead..

Andrew

[MrBean]

Andrew. Yeh... If you found the Yahoo group "PCB-Gcode", then you're in the right place.
The step feature is really handy, although the cuts could be optimized for better speed... It seems a bit random, the way it orders the cuts. (Lots of Z and rapid moves). Anyhow, I'm not complaining. The results a way better than I was expecting.

The script can also generate a seperate 'copper removal' gcode file, but this is "Raster" fashion. It tracks accross the board incrementing the other axis each pass. Much like the built in JPG > gcode in Mach3. Using stepover is much better.

If you give the new ULP a go, can you post your results here if possible? It would be good to see how it goes.

Regards Terry...

[tfmacz]

I am trying to V2 working. It is running but the only mode available is the 1/16 step the mode jumpers seem to have no effect????

I've been using Alan's PICStep drivers for around 18 months now, and not had any major issues with them. In fact, they've been really good.

---------SNIP-----------


I tried to switch back to 1/8 mode for my Z axis, but found that the jumpers on the interface board no longer seem to alter the step mode??

----------SNIP----------


The 1/16 mode does generally seem to make my machine run smoother. Cheers Alan....

Regards
MrBean.

[santiniuk]

Weird.....

I am using V2 (Autoshutdown code version) and can confirm that jumpers do work, and change modes.

For the 1/16 step mode All jumpers are on. (On the version I'm using).

Strange thing is that Mr Bean reported problems with the jumpers also ???

Good luck !

[MrBean]

Ooops. I forgot about this thread.

I'm using the NON autoshutdown version. My step mode jumpers don't appear to be working.


tfmacz: Are you running the Auto shutdown version or not??

Regards.
Terry.