[abfa9358]

And so the long process of CNC conversion starts....
After the success of moving and installing the machine (http://www.cnczone.com/forums/showth...ht=BF30&page=2) I have made a start on the conversion to CNC. I suspect that this will be a sloooow process due to family / re-decorating commitments but in true Shawshank style, I'll just chip a little away each day.
Heres a picture of the X axis mount plate I finished last night with a pic of the general 3D CAD. I've ordered the X and Y ball screws and have the stock material to make the ball nut housings. I'll post more when I get to an interesting point.

[forrest39]

Nice drawing!
What motors are you going to mount?

[Arich0908]

Looking good already . Im taking my seat.

[abfa9358]

Motors will be NEMA 34's from my existing mill. Looking at the work of Longrat(http://www.cnczone.com/forums/showth...highlight=BF30) I think my X and Y motors will be fine. I suspect that the Z axis will need some more torque. Optimum have around 17Nm on the axis in their official conversion kit - however I think this is aimed at drving the conventional lead screws and is just supplied for the ball screw conversion as is.

Katran (http://www.cnczone.com/forums/showth...highlight=BF30) added gas struts to his BF30 due to weight of the BF30 head. I note his comments and now have experienced how heavy the head actually is but I plan to see how things work and fix what doesn't. I built my whole system so I'm not too worried about changing things around.

[LongRat]

Good to see it is under way.
Please remind me, are you getting your screws from Linearmotionbearings too?
I recently got a Wabeco 1410LF High Speed for work (unfortunately not home) and it is absolutely beautiful. You really get what you pay for - which is a lot. With all the options it was about £12k I think.

[abfa9358]

Screws are coming from Zapp Automation. I've asked for C7 grade, preloaded using oversize balls. Wabeco 1410LF - Mmmm that sounds very nice, what line of work are you in?

[LongRat]

Look forward to hearing how good those screws are.
I am an R&D engineer, designing and building machinery for milling dental crowns and bridges. This mill will be used for prototyping.

[abfa9358]

That sounds like an interesting and challenging job, I am a non-linear FE guy with a large Aerospace company.

Anyway back to the BF30. I’ve had a delay on the motor mounts due to waiting for some materials, so I decided to swap over my home brew DRO unit from the X2 mill. I bought two new scales for the X and Z axes and will swap the Z scale from the X2 mill on to the Y axis of the BF30. The other night I wired up one of the new scales to the DRO reader to check that it could read the scale correctly. It did and so I don’t have to change the PIC firmware, which is good.

Over the last few nights I’ve been mounting the X axis scale and building a swarf / coolant protector for it. Flood coolant is now installed on the mill and makes machining steel a much easier task. I’ve used the re-circulating pump from my old mash brewing setup as a coolant pump and all seems to work fine. I get coolant from a few drips/minute up to complete coverage of me and the entire workshop, it does however probably spell the end of using the pump for brewing purposes – hay hoy.

I’ve attached a few pics of progress.

Next move will be to mount the Y and Z axis scales and get the DRO system back working. Its surprising how much I took it for granted on the X2. Then I’ll get back on the X and Y motor mounts.

[the4thseal]

what is the resolution of you scales? and the box reader? the stuff i find is only resolving to .001 or 0005. yours seem better? what are they and where did you get them?

[abfa9358]

I built the scale reader myself following this resource http://www.shumatech.com/support/chinese_scales.htm. I used a PIC 18F252 for no other reason that I had it and it could do the job.

The scale output is two 24 bit words sent LSB first giving relative and absolute position. The scales measure 20480 positions / inch. In practice, I found that quoting position to 0.0001" was the best that I could do in terms of reading jitter and to be fair I'm not a 1/10000" machinist. So my reader decodes the 48 bit data streams from each axis scale, gives the option of mm or inch and measures the spindle rpm. I did find that triple sample / majority averaging of each data pulse had a big effect on the stability of the readings. All scale run in slow mode as the jitter gets worse if I run them in fast read mode.

I think that commercial scales only read to 0.0005" because of the jitter issue, I find my readings have +/-0.0003" of jitter when on the mill - a bit less when on the electronics bench inside.

I was thinking of moving to commercial glass scale DROs but the cost was around £500-600 pounds and not really worth it for me since the DRO reader has worked fine for around 5 years now.

[abfa9358]

Well the X and Y axis ball screws arrived today - all looks well and nicely machined. I'll post up some pics tomorrow. Attached is the Z axis scale protection scheme. Basically, the link between the reader head and the Z table will run between two clamped brushes (draft excluder strips) the digital scale will mount on top of the hexagon via the more central hole. This should prevent chips and fluid from reaching the reading head whilst allowing free movement.

[abfa9358]

A bit more progress has been made on the DROs. The X and Z axes are now inplace and waiting for connection to the reader. The reader has also been mounted to the mill. Here are some pics of the Z axis scale and jig used to drill the mount holes. The face to face brushes seem to work well. Also posted are pics of the ball screws from Zapp Automation. I will continue with the Y axis scale over the coming days. Progress is slow but on a bright note, the loft conversion is now fully decorated and waiting to be filled with child related clutter.

[photomankc]

Do you have a schematics on the reader box? I ask because I am thinking of doing a very similar homebrew remote display but using a Parallax Propeller chip to read the scales and do the display. My main question revolves around level conversion between the ~1.5v to 3.3V signals and the most effective way to do that. I really just need to pick up a scale a start experimenting.

Since the Propeller is an eight core processor I figure 4 can be devoted to collecting the data from the scales in assembly while the others drive the display, monitor buttons, and maybe do some calculations in the interpreted language.

[abfa9358]

I don't have schematics for my reader(I'm not that organised), I have the PCB layout and probably can find the component values used. I built my reader based on the information here http://www.compendiumarcana.com/caliper/ and here http://www.shumatech.com/support/chinese_scales.htm. It uses a PIC18F252 clocked at 20MHz which spends most of its time doing not much. My code occupies about 14% of the available code space.

Best way is definitely to buy a scale and start to investigate. I built everything up on a plug board to start with until I was sure the circuit was ok. A digital oscilloscope is really useful too. My reader does nothing clever in terms of DRO functionality. The scales operate in slow mode (~3Hz) and this has been fine for my use. I built a version which switched to fast mode (~50Hz) but found the jitter increased quite a bit. After that I think I got hooked converting my X2 mill to CNC so never progressed the reader as it work ok for me.

[abfa9358]

Well after some time there has been a bit more progress. I've pretty much finished the X axis motor mount(see pics). I'm quite pleased with it and its been good experience in terms of getting used to the new mill - which is turning out to be a very capable machine. I've also finished re-fitting my DRO system after a false start with one of the scale units.

I am now making the Y and Z motor spacer plates and two pairs of side plates to bolt directly to the mill. The Y and Z axes are less complicted than the X axis in terms of motor mounts.

I've also had to rebuild my control PC which runs EMC2 as the (almost new) hard disk failed. I put in a 30Gb solid state drive and in future I think I'll change over to an ITX motherboard if I can find one that runs EMC2 with a reasonably low jitter.

[LongRat]

Looks like your motor mounts are made of steel? Hardcore!

[abfa9358]

Hi Longrat,
Yes mounts are in steel, side plates and main plate are 1/2" and mount plate for the motor is 1/4". The mill chews through it no problem now that I've setup a flood (the workshop) coolant system. All big holes were bored out on the lathe.

[grf]

do you by any chance have any cad drawings for those plates? i have a bf-30 too and just received some servos and want to start conversion. thanks

[abfa9358]

grf
The whole motor mount assembly is modelled in 3D using Turbo CAD 17. I am quite happy to share this with you, need your email address and a file format you can handle. A few points to note:
Initially, I used my old X2 and my DRO system to measure the profile of the X axis plate on the mill. When I transposed this data in to TurboCAD I found that the errors in the measured data were unacceptable. To resolve this I made a number of suitably sized bushes, clamped the plate down on to the 1/2" steel plate blank and then spot drilled all the holes (3 x M8, 2 x Taper pin and one for the lead screw). I then used my centred the mill on the lead screw pilot hole and then pitched off for the 6 x M6 mount hole for the motor brackets.

The two 45mm deep side plates need reliefs adding to provide a bit of clearance for the two outer M8 fastener heads. This can be done with an end mill or a boring head but is not shown in the CAD. Without relief all fits together but there isn't much movement left for the bearing plate to float around.

Can I ask why servos on a machine this size?
Regards

[grf]

so if i got it right the drawings are not accurate enough

the servos i got are yaskawa 400w
i bought from fa-parts. i can tell you their service is second to none.
progress is very slow but at the moment i have all the electronic parts and going to start testing with mach3 (i would like to give the emc2 a try but scared i ll be lost). i am thinking of 3-1 reduction for x-y and 4-1 for z.
let me know if you need any more info