printing an analogue optical encoder disk

Hi, I have a BLDC motor for wich I am considering a single chanel analogue optical position sensor so it generates a single ramp over

1 whole turn.

Ive made a reflective wheel from some sticky backed paper label, and a carefully pencil drawn ramp and it works well enough to prove its feasable.

the next step is to print it out with my inkjet printer, has anyone done anything like this ?

unless theres already something out there someone knows of I can probably write some software to generate a bitmap image with c# if so il post the program or bitmaps somewhere.

another question is how do I ensure I get the image the exact right size on the paper without doing it by trial and error ?

I also assume black would be the best absorber for infra red but this remains to be seen, or if the some of the other colors are more reflective than the white paper. im not even sure if the color needs to be a linear change, I could use pwm lines thin enough so it doesnt show in the signal.

I have a new epson stylus photo R285. unfortunatly I intended to get one with waterporrof ink but they never make it clear if its not waterproof in the sales blurb of the printer you have to go and delve into the ink itself.

does anyone know how compatable the two different types are ? pigment type seem to be the waterproof ones and dye based arnt. Im not sure a pigment based ink would do anything more than just clog the jets up permanently wich would be a shame. a waterproof coating might interfere to much.

many thanks

Colin =^.^=

Reply to
colin
Loading thread data ...

Someone posted on the same topic here, or an affiliated group, a few months back and I took a note of the URL he gave:

formatting link

... might help.

Chris

Reply to
christofire

christofire wrote:

The original thread:

formatting link
*-*-set-screw-motion+*-running-*-*-*-*-constant-rate+freeware+XP/Vista+centred+zzz+generates-codewheel-images-and-prints-*+build-*-*-optical-encoder+*-*-sound-card-to-digitize-*-*-*-signal+*-message+*-*-shim-would-correct-*+*-*-*-*-signal-processingnews: snipped-for-privacy@4ax.com

Reply to
JeffM

Yes, but not with an inkjet printer. I used an HP2100 laser printer to get enough pigment on the page. I also had accuracy (aspect ratio) problems when I tried it with an inkjet. I used glossy photo paper, clear acetate film, and paper labels in the laser for making encoder disks. They all worked. The HP2100 does 1200dpi which was great for accuracy.

See:

There's a postscript program for generating the various images with links to other sites that are doing the same thing.

You don't use an inkjet printer. They're very accurate across the page, but subject to paper slippage and roller wrap accuracy issues due to the feed rollers. If the paper goes through flat, you have a chance. If it's curled under the inkjet head, forget it.

What I did was to print a grid test page with rulers generated from the dot count of the page. You can find graph paper templates somewhere on the internet or generate your own. See how close you get with a microscope or optical measuring tool. I've found that the laser is very accurate across the page, but varies slightly lengthwise, again due to roller slippage. Cleaning the white clay junk off the rubber rollers made a huge improvement. There was also a problem where the page tended to feed somewhat at an angle. This caused some minor issues when I was trying to maintain registration on multiple passes, but should not be an issue for single pass printing.

Laser toner is black plastic. Nothing goes through it, including IR. No clue on inkjet ink as the formulation can be anything from a dye to various pigments. I suggest you avoid this issue and just stick with the laser printer.

Huh? What kind of sensor are you using that responds to colors? Why do you need a color encoder disk?

Archival inks are pigmented inks.

How the mostly water based inks work:

You should be able to find more on the topic by following the patent references and citations.

Inkjet nozzles are about 5 to 10 microns in diameter. The pigment particles run between 0.02 to 0.30 microns diameter. The "ultra fine" inks tend to be on the low end of the particle diameter range. The "ultra bright" inks on the high end.

One is usually sufficient.

--
Jeff Liebermann     jeffl@cruzio.com
150 Felker St #D    http://www.LearnByDestroying.com
 Click to see the full signature
Reply to
Jeff Liebermann

...

oh well I think you deserve more than one for that reply :)

thanks to the other replies too, actualy I also made a thread about digital encoders too, but my attempts wernt so good for various reasons so i quickly gave up as I found a way to use 2ch comercial ones instead.

however this one is analogue, wich makes a bit of diference. im not sure if there are any analogue generators there, they seem mostly digital but il have better look through.

It was initialy just a rotor reference pulse but as I need to test run the BLDC without the encoders id thought id expand it to be a crude single chanel position encoder as ive used nealry all the available mcu inputs.

it worked surprisingly well with just a hand drawn pencil line so im sure it would be better even with inkjet, ofc once I got it right I could relativly easily get it done on a laser somewhere.

its also a reflective encoder too as well as being analogue wich is why the question about different colours, purley from reflectivity of IR.

I got this inkjet for general use in a hurry becuase my old one decided to go caput 2 days before a presentation and this one becuase it has one of those trays for printing onto CDs so it has a flat path and I imagine registration will be a lot better than thin paper.

I was hoping I would be able to print a panel or something, wich would be dificult in a laser printer, its hard to find a laser with a flat path wich can take something unbendable, plus im not sure the transfer or fusing would work so well. however non waterproof ink isnt very suitable for what I wanted.

however I was mainly not sure as to how control how a bitmap image scaled to the size of say an a4 sheet. but as it seems to automaticaly print a photo image file to fill the entire page I gues I could simply use that as a scale to go by.

Colin =^.^=

Reply to
colin

Regarding print size, check your printer's options to see if you have told it to expand the document to fill the page. I use a couple Epson printers and a HP printer. All are very accurate in printing the exact size of the documents I send to them, whether pictures or graphics.

For a current test, I used a limited use, free trial editor, Eazydraw, to create a 3" circle (measured by it's ruler) inset with a star, and cross-hairs to locate the center. It printed the correct dimensions on the HP. I picked the HP printer cause it's wireless and I'm lazy today. Other graphics editors can do as well.

Before I bought my first "good" printer years ago, I compared the outputs of ink jet printers and postscript printers and found no discernable difference. People who have problems with ink jet printers are usually using cheap quality paper. For graphics I use 94 Brightness, 24lb Weight paper. They also print great on transparency sheets and photo paper.

Reply to
Don Bowey

You'll probably do better to use a quadrature encoder rather than a single analog channel. Analog is going to be subject to all sorts of oddball effects (stray light, aging, temperature -- you name it).

You can extend a quadrature encoder's precision using analog methods (take analog values of the A and B channel, and use them to do interpolation).

--
Tim Wescott
Wescott Design Services
 Click to see the full signature
Reply to
Tim Wescott

I'll add it to my karma collection. I need a karma recharge to compensate for all the evil things I've been doing.

I haven't done an analog control system since college. Far too many complications as others have mentioned. You can also do lots of nifty tricks with digital, like not worrying (much) about critical damping, drift, stopping on a dime, etc.

Did you check it under varying conditions? No load, full load, overload, ambient light, hot, cold, wet, dry, condensing moisture, dust, etc? All of these will affect an analog system.

Dunno offhand. It's probably easier to print a page of color stripes and measure the reflectivity, transparency, absorption, speckle scatter, and edge diffraction, then to guess.

It's not unique to inkjet printers but all electronic devices have a propensity to self destruct just before and during a demonstration or presentation. If that's insufficient, the cords, cables, remotes, and driver disks, tend to hide themselves when needed. There's no way to fight it, so just get used to things blowing up at the point of maximum inconvenience.

Huh? I do it all the time. You can get sheets of sticky back 8.5x11 "label" stock. Print your front panel, cut to size, punch out the holes, and stick to an aluminum front panel. If you want it to look good, use overpriced glossy photo paper. If you want durability, there's paper that has a mylar covering, but requires a thermal printer. You can also use clear transparency acetate paper. Print a reverse pattern, attach to the front panel, and no protection required. Something like this:

Yep. I once tried shoving a sheet of aluminum through the laser printer. It actually printed, but the aluminum panel dissipated the heat from the fuser rollers too quickly causing the toner (plastic dust) to flake off. It also scratches the drum if you forget to deburr the holes. The printer also made some ominous grinding noises which suggested impending disaster. Not recommended.

Transfer works fine, but does not seem to be particularly uniform. There are always a few dots and holes on the media. These can be touched up if necessary.

I use an image editing program such as Irfanview 4.20:

The setup for printing allows resizing to specific number of pixels in both x and y axes. The problem is that if you have a complexicated pattern, you could end up with fractional pixel round off errors, which will deteriorate the quality of the image. Play with it and see how close you get.

--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558            jeffl@comix.santa-cruz.ca.us
 Click to see the full signature
Reply to
Jeff Liebermann

I'm with Tim Westcott on this - you'd be much better off with a two- channel quadrature encoder (which you could buy) generating a digital output.

I'd go further than Tim in suggesting that you feed the digital output into an up-down counter whose output would then drive a D/A converter, which would give you your single ramp. You may need a high resolution encoder and high resolution D/A converter to get a ramp that is as smooth as you would like,but until you tell us more about your application we can't really be all that specific.

-- Bill Sloman,Nijmegen

Reply to
bill.sloman

-- depending on your goals. If you're determined to DIY, and if your microprocessor skills are strong, then I'll stand by a coarse quadrature encoder with interpolation.

But many computers these days come with accessory quadrature encoder hardware. They're these little ivory-colored things, about the size of a pack of cigarettes, with a little ball on the bottom and clicky buttons on top. After you save away the ball to use as slingshot ammunition and discard the case, there'll be two dandy little encoder wheels in there, just waiting to be integrated into your project.

--
Tim Wescott
Wescott Design Services
 Click to see the full signature
Reply to
Tim Wescott

thanks all I appreciate well the pros and cons of all these options however there are actually some exceedingly high resolution

2ch quadrature encoders on the shaft wich are tricky to set up (not to mention theyr pushed way past their ideal max frequency) but as they are 2ch and not fixed I need a reference pulse. the start/end of the ramp will be quite suficient to extract a precise reference. the analogue ramp only needs to be sufficient to enable comutation in the absence of the quadrature signals while setting it all up.

its controlled by a DSP wich at up to 20krpm has plenty of monentum to interpolate well between index pulses and it already monitors phase current to optimise phase angle/voltage, however when the encoders arnt working start up/slow speed and sudden encoder or rotor slip is awkward wich makes it dificult to debug.

it does however cope moderatly well in open loop mode as a synchronous motor, in fact ive tested an induction motor as a good alternative, but size of 3 phase ones is a bit too large.

the DSP also has a QEI input module but im already using another 3ch encoder elsewhere, the addition of several acceleromters and having to double the number of encoders used to increase resolution and other things has used more pins than id iamgined. in fact i have just 1 input for this wich fortunatly is analogue. not to mention its to hard to physically fit more than one sensor or have more wires etc, not to mention the PWM pulses createa a bit of havoc with difital pulses.

the only draw back I see it is it sudenly gets a bit of oil on it :s other analogue problems such as drift can easily be calibrated. otherwise it would be a cheap commercial alternative.

id love to explain the whole project but this is such a small part of what would be quite chalenging project to explain in its entirety.

I have however asked many questions here over the last few years, but often geting stuck and shelving the thing till i have a bright idea, or new devices come out.

I did however acheive almost sub picosecond resolution (with averaging) however it seems this didnt detect what i was looking for and need to improve this by an order of magnitude, I have sevral ideas, but my health is a bit of an issue atm it means I have plenty of time to pursue this but not the ability ... wheres otherwise id probably have the ability but not the time ...

Colin =^.^=

Reply to
colin

....

some people say digital contains a non linear analogue system... digital still can have its complications too.

this is mostly a digital system, as it has a PID control loop done by the DSP, with PWM 3phase drive to the BLDC motor. but with various analogue and digital inputs.

nope, just spun the disc by hand and looked at it on the osciloscope. its posible to see its a ramp with a fast edge. however trying to get it a straight ramp meant adding some more pencil marks here and rubbing some off there wich ended up in a rather jaged ramp. its stored/operated in an environmentaly controlled environment, (very light tight of course) as those things you mentioned severly affect some of the other digital sensors too.

....

thats quite a good idea, however its quite a chalenge to get the glue interface to be perfect so its totaly invisible. Ive previously used back engraved perspex panels for high usage controls. glueing the paper on the the rotor did cause some problems as the glue stained the paper, the sticky labels wernt a very good stick either. the disc also has to have a very large hole in the center wich doesnt help.

thanks again Colin=^.^=

Reply to
colin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.