PDA

View Full Version : Trav-A-Dial spinoff



MCS
02-27-2011, 05:55 PM
In the Trav-a-dial thread I had the idea that there had to be a solution somewhere and I found a reasonable one.

Chips from Austriamicrosystems.

First I sampled their state of the art linear high resolution chips. It didn't work with the magnetic tape I purchased from machine-dro.uk. Wrong pole pitch. I am still trying to locate a supplier who can deliver a reasonable quantity for a reasonable price. It's available anyway.

Austriamicrosystems has also chips that match the machine-dro.uk tape so I sampled those along with their rotary encoder chips.

Made pcb's for both and tested them.

The rotary encoder has an absolute output, making them ideal for for example a throttle position sensor. It also has an incremental output, making them suitable for the Trav-A-Dial. Most suitable, because all the encoder needs is a rotating magnet to look at. No additional bearings, no rotating parts.

The low resolution encoder has a resolution of 0.025 mm, which is on the low side for metal working.

The positive side is that there is such a thing as magnetic printable paper.

Used for whiteboard magnets with advertising. I had without knowing 5 of those of my whiteboard. Over short distances they look pretty accurate.

So on normal paper format (long side squared) the DRO can be ghettofied.

Photo:

Upper the rotary encoder, looks through a bore of 18mm, mounted on square 22mm.

Lower the linear encoder with magnetic strip, size 28*11*6 mm.


http://www.mijnalbum.nl/Foto-OWYWOVMZ-D.jpg

MCS
03-03-2011, 04:37 PM
As I have a job coming where I need some more intricate measuring, I decided to go further with the low resolution linear measuring system. By switching pcb and magnetic tape I can always go to high resolution.

But a resolution of 0.025 mm is reasonable for milling or Z of the lathe.

So I milled an reading head (14*14*38 mm) which I can mount on my lathe.

In the past I looked at DRO's for my lathe, but the scales were always too big. This is really tiny.



http://www.mijnalbum.nl/Foto-YDBV4LOK-D.jpg

miker
03-03-2011, 05:54 PM
More please!!!:)

Rgds

MCS
03-05-2011, 03:56 PM
Testing today.

On the photo: left the reading head taped to a magnetic dial holder, the tape taped to the bed. On the right a 0.01 mm dial for checking. On the support the AVR development board. The bare DRO is nothing more than a microprocessor, LCD-display and crystal. It violates my rule not to tie antenna-like cables (2 meter) to the processor.

This 0.025 resolution set is already usefull, did long distances, fast long distance (to check repetition of the software) and small distances for intrapolation tests.

Good response, I feed the LCD-display at ~35 hz, it's dynamic.

All it needs is a little keyboard for reset/preset, feature creep is up to me.

Did also tests on the little pieces of printable magnetic taper.

Magnetism is somewhat weaker, easy adjusted by closing the gap, 1 piece had a deviation of 0.01 mm over 10 mm, the other 0.03. It has no steel backing as the official tape has, but it remains interesting material.




http://www.mijnalbum.nl/Foto-JIRX4WYU-D.jpg

MCS
03-07-2011, 08:10 AM
Latest and last update. I know people hate electronics.

Got today the neodymium magnets for the rotary encoder, all the way from Germany. These are especially sold to use with the chips, on the photo the magnet is sitting on the chip, in the middle of the board.

The magnets are diametrically magnetised, the chip intrapolates North/South to see the position.

The chip has absolute position, to get instant absolute position within one turn and AB incremental output for multiturn use. This can also be combined.

I needed the magnets to establish the gap between the chip and the magnet.

Now that is done I can make parts to replace a Throttle Position Sensor, reducing the original width of the throttle body by 3 centimetres.

Total cost of the encoder as of now (positively reformulated): 0.30 Euro.


http://www.mijnalbum.nl/Foto-AZHVT8FR-D.jpg

MCS
06-29-2011, 05:54 PM
Ever so slowly I'm creeping towards a DRO.

The rotary encoder is a multipurpose thing, so I decided to make that first to use for my mill (beware! EMCO round column). This is the most direct mechanical-electronic coupled device I ever made, so I made it an exercise in concentricity. I hoped it would be a "pencoder", but making things really, really small has its limits. The pcb has a diameter of 20 mm, it is timing belt driven on ball bearings, so that's it.

It has a reasonable resolution (1024 per revolution), so relating it to the real movement (1.50 mm/revolution) is just a simple multiply/divide translation.



http://www.mijnalbum.nl/Foto-TST67PEK-D.jpg

MCS
07-03-2011, 05:45 PM
The encoder mounted (and tested).

The pulley on the spindle is made as an replacement for the dial. I don't want my retrofits to invasive.

I tested with a dial against the milling spindle with increments of 0.25 mm, effectively using 6 points of the encoder in repetition. Encoder and dial readout are spot-on (after you've seen stick-slip and other things that need TLC, like oil), proving the encoder is concentric within limits.

After that I fired up both VFD's, the AVR development board was not impressed, as expected, though antennas don't come much bigger. I think a lot of electronic noise problems is just buggy software.

I'm planning to do this conversion also on my rotary table.

To travel towards a multi-axis DRO is just boring electronics, I will spare you that. For this one I'll cobble up a simple thingie with a reset button for the time being, that will already be an improvement over what was.


http://www.mijnalbum.nl/Foto-KC6CSAYO-D.jpg

Edwin Dirnbeck
07-11-2011, 12:16 AM
What a fantastic project you have created.I have allways liked travadial because they are easy to install and have unlimited travel.Have you considered springloading your encoder directly against your ways just like the travadial? Let me dream further could your encoder comunicate wirelesly with your readout display? Please keep up the good work and keep us informed.thank you, Edwin.

alsinaj
07-11-2011, 10:56 AM
What Edwin said.

MCS
07-11-2011, 03:47 PM
What a fantastic project you have created.I have allways liked travadial because they are easy to install and have unlimited travel.Have you considered springloading your encoder directly against your ways just like the travadial? Let me dream further could your encoder comunicate wirelesly with your readout display? Please keep up the good work and keep us informed.thank you, Edwin.

Thank you.

One should always keep dreaming, but dreams go in different directions.

Assume I would make a travadial. Assume an absolute minimum resolution of 0.025 mm. That would make 40 pulses per millimeter, the diameter would be:

1024 (pulse per revolution) / 40 (pulse per millimeter) / Pi = 8.15 millimeter.

Off course diameter goes up, as resolution is decreased.

The whole travadial idea drives one into the watchmaker range, be it electronic or mechanical. I think it was a good solution in its era, superceded by electronic solutions.

Wireless (or infrared) is also possible, downside is batteries.

The general idea of this solution (DIY encoders) is that everything is that tiny, that you can wiggle your way with cables towards a cablechain without compromising mechanical movements.

Situation as of now:

"Thingie" : Lcd display with processor underneath and update and communication cables.(Reset switch on the back). Now I have to write the communication software (I2C) to have position processors and a display/keyboard processor.

The idea looks more complicated than it is. Cost and function capability of today's processors makes this a logic path, while making it expandable in numbers of encoders and functionality of the separate units themselves.

http://www.mijnalbum.nl/Foto-KOSGAPIQ-D.jpg

MCS
08-09-2011, 05:54 PM
It's beginning to be a case of severe feature creep.

The idea was to make the DRO modular, because counting pulses is a high speed process, and communicating and displaying data a low speed process.

Got that working.

Needed a loop connector and found a self made 7-segment display. Made that working, because the more display's that are in agreement of the position, the better. While developing, that is.

Thought that a 2-axis mini DRO in a single module may be handy and made that working with a rotary and a linear encoder.

Having designed with HCTL-1100 servo chips I suddenly saw the yet-to-program high end, but now affordable, servo system resting on my desk.

The system with linear and rotary feedback. Thought that over, yes, it certainly has it's advantages. Has the potential to screw the screw ball spindles, if you catch my drift. Put more craftmanship in a control till you have a craftsman.

Have already the double digital filter (rotary and linear) working too fast in a too slow processor. The key is interaction of the two and the possibility of different tuning (dynamic and static).

The HCTL-1100 (which can't do the double) is now obsolete, but successed. You can find the datasheet of its successor here:

http://mektronixtech.com/index.html

It's a nice way to control a servo. They have added a few usefull things, like S-curve velocity control and backlash. For myself I can add other features, like handling overload, like pulse-direction commands, for coordinated moves.

So here you see pulse-direction hanging on the other end of the chain.

So it creeps on and on.

MCS
07-30-2012, 07:04 PM
http://i175.photobucket.com/albums/w154/bestwel/Servo.jpg

Have come this far. Interesting journey, these servos. Got it even working with a servo motor. And don't give out code anymore.

macona
07-31-2012, 04:33 AM
Did someone take your code and run?

Servos can be a pain, but they are worth it in the end. I stick with commercial servo controls. Its makes tuning so much easier. I am using Elmo drives in my laser cutter.