DIY magnetic DRO

[ChrisBParticipant @chrisb35596]

Posted by Ian P on 21/11/2019 15:48:10:

As I understood it, whilst the tape has a magnetic pattern in it, the amount of magnetic force is so low its unlikely any swarf or chips would stick to it, maybe ferrous dust might accumulate but with the protective stainless foil over, it would be easy to wipe off once every few years.

You'll be surprised! Large pieces of swarf will not stick, but on a milling you'll be producing lots of small chips, and those will stick to the magnetic tape with ease. My reasoning goes: better have the tape visible and accessible where I can easily give it a wipe.

Chris

[Jed MartensParticipant @jedmartens56976]

I've taken the plunge and ordered two read heads and tape, which I'll use for X/Y on my mill. If it's a success I'll roll it out to the mill Z axis, and maybe the lathe too.

I'm still pondering the pros and cons of the various ways to mount the tape to the table, thanks for the feed-back above. I like the simplicity of mounting the tape directly to the side of the table, and see the advantages of accessibility, but it makes mounting the read-head more challenging…

I've ordered the aluminium to make the tape bracket, should I chose to go down that path, as it was cheap, but I'll probably wait until I have the read-head and tape before I decide.

For the electronics I'm planning to go full DIY. Something like…

* RS-422 receivers to convert differential signal from read head to 3.3v single-ended.

* Small programmable-logic card to count the pulses and present the data on an SPI interface. This is timing-critical "real time" stuff, which I don't like leaving to software.

* Raspberry Pi with touch-screen, reading SPI data and running a DRO GUI. The software will be written in Qt (a cross-platform C++ IDE with decent gui/widget support)

I've used similar set-ups for other DIY projects so I'm not having to figure much out from scratch.

On a slightly different topic, I note that RLS also make rotary magnetic tapes, which are used in conjunction with the same read heads. Given that electronic lead-screws are en vogue, has anyone used a magnetic encoder instead of rotary optical encoders to track the spindle position? I imagine that the rotary tape could be fitted directly to the back of the spindle, and monitored with no physical interface, removing the timing belts and pulleys that I've seen the guys on youtube use.

[Jed MartensParticipant @jedmartens56976]

I decided to go ahead with my first idea. But only for the Y axis for now, as it's the shortest and if things don't work out, it's easy to replace the tape.

I machined the read head enclosure out of aluminium. I started with a drawing but ended up free-styling a lot of it.

The bracket for the tape is pretty simple…

I ordered some armoured cables from Machine Dro. I removed the DB9 from one end and drilled a hole to accept the cable sleeve in the read head body (10mm). This is retained with a M3 grub screw. The cable quality is very good – all connections individually sleeved, and good screening.

Next, I bodged together the electronics with parts I scavenged from the left-overs bin in the office. It ain't pretty…

But it works!

Ok, only the axis labelled "X" is live in the above GUI pic.I'm just moving the read head against the tape bracket by hand, but it increments in one direction, decrements in the other, and it changes by roughly what you would expect.

To be honest, it's all gone too well, so I'll stop now before something blows up. Next step is to fit it to the mill, and to consider how to tidy up all the electronics and mount it in some kind of enclosure.

Edited By Jed Martens on 18/12/2019 22:44:04

Edited By Jed Martens on 18/12/2019 22:51:02

[Jed MartensParticipant @jedmartens56976]

And here it is fitted to the mill…

[ChrisBParticipant @chrisb35596]

Well done! Nicely made Jed, glad you managed to make it work.

Chris

[Jed MartensParticipant @jedmartens56976]

Continuing with this project…

I've put together my own DRO display, based around a 10" touch-screen and a Raspberry pi…

Starting from bottom left, there are 4 DB9 connectors that go to the scales. The red PCBs on the left are the RS485 differential signal converters. The output from those goes to the PCB at the top of the box, which is an FPGA card that does all the heavy lifting of quadrature decoding, counting, etc. It conveys the axis data via SPI to the Raspberry Pi (centre). This drives the display/touch-screen, runs the GUI, and has an Ethernet connection to the outside world. The board to the left of the Pi is touchscreen driver. The PSU is on the right.

Here's the case from the outside. I made it from aluminium, at just the right size to fit the screen…

I've started playing around with the GUI. Here's an example of one idea, where a bolt hole circle is plotted graphically.The wee cross-hair moves as you move the table, the plotted holes stay where they are.

I've separated Z axis and quill (which I've labelled Q, but I have no idea if that is correct – what is the quill position normally called?) and plan to allow them to be optionally summed together. 

Edited By Jed Martens on 07/06/2020 17:08:28

Edited By Jed Martens on 07/06/2020 17:13:53

[Ian PParticipant @ianp]

That is quite an impressive project!

Ian P

[Nick Hall 1Participant @nickhall1]

It's been a while now chaps, how are your DIY DRO's holding up ?, still in good calibration ?

Nick

[ChrisBParticipant @chrisb35596]

Posted by Nick Hall 1 on 29/06/2021 17:30:19:

It's been a while now chaps, how are your DIY DRO's holding up ?, still in good calibration ?

Nick

The DROs are holding well after 2 years in use. No issues to report about so far. Been well worth my time and saved me some money as well.

[Author]

Posts