Position Encoding.

When I was involved, the methods were equivalent to using the I output to clock a counter and the Q output to control its direction.

Yes. A state machine was the high-tech hardware fix. I showed how to do it with two XORs

Jerry

--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

in most applications I've never used an encoder that would suffer much from that effect. Usually the flutter is at a minimum if you're using this for high speed synchronizing.

Even so, the hardware would have to be in bad shape to have it troublesome to the point where it is inhibiting performance.

I suppose if you are dealing with high speed movements this could be an issue, in my case, it was quit the opposite and even so, the hardware was indicated and constructed properly

It helps to have precision machinist and tools at your disposal. :)

Jamie

How does speed affect position error?

Come now! Doesn't that depend on what performance is required?

Again, how is speed involved?

Our machine tools were equipped with interferometers and accurate to a quarter wave of the mercury green (587.2 nm) line. In fact, I first developed the quadrature tracking circuits to make those interferometers stable in the face of vibration. How accurate was yours?

Jerry

--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

[snip]

It's how the unskilled pat themselves on the back and think they know something useful.

Also how they "cope" ;-)

Unfortunately they are a society-damaging crowd... they're the kind who've given us climate change.

There should be NO publicly-supported universities. Then the crap would end.

Absolutely!

...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |
             
Annoy a Liberal: Stand Them Up to a Blackboard and Ask Them to 
Show Their Math That Balances the Budget, Even After Taxing/Taking 
Everything the Rich Make... Then Punish Their Ignorance :-)

The fastest that 6 inch disc moved was most likely no more than 20 RPMS. I never measured accuracy to that level, didn't need too, the machines these units were mounted to developed some very highly accurate products.

This has to be at least 20 plus years ago now. Also around that time, I worked with a group using camera's to inspect and reject items off a line if for example, the label on the bottle was not correct in some manner. The camera was accurate to .001, 100 feet away. Yes, it was a very expensive camera. The code done in a dedicated Z80 box using Forth.

If any one is familiar with the Herseys chocolate squeeze bottles, that is one, and next to me was another project doing hole drilling inspections via a camera in the gyro mechanics of missiles.

So, I did the encoder scheme back in that irra

I just don't see what all this fuss is, it was done, used flawlessly and we moved on to the next adventure.

I don't know why you are taking this so personally, I have no interest in designing encoders and I really wasn't that interested back then, it was just given to me to convert some old production equipment over for better through put, which is what happened.

You can fight and argue all day long. Seeing it was operations with out problems is all that mattered.

Accuracy was checked as I said, with a CNC DRO and hand tools via the machinist. The Film shop we had do the work for us used some sort of shadow graphing to get the aspect correct for the photo disc. I am not a photographer so I don't fully understand how that process went off but it did and worked out fine. That is why you some times need to pay people to do certain aspects of your project.

Jamie

Among themselves.

It's still a status competition inside a closed society.

Reality is designing and building electronics that people want. It's called productivity. You wouldn't understand.

How's that oscillator coming along?

John

The "Occupy" thing is hilarious. A major part is kids who want jobs and want their student loans forgiven. It's funny to hear what their majors are: sociology, psychology, art, music, English, theatre, political science. No engineers, no physicists, no mathematicians.

John

(snip, someone wrote)

Last time I was working with one of these, I looked up some references.

If you do it that way, then you lose three quarters of the resolution.

The one I found uses a clock that must be faster than the possible rate at which it can go from one transition to a different transition. Then, latch the inputs on each clock cycle, compare to the previously latched values, and increment or decrement if different. The only complication was getting the counter to wrap right, as the period wan't a power of two.

It can sit right on one transition, and even change faster than the clock, without any problem. (It might miss a transition and then back again in that case.)

I didn't think of this one before, but if the clock isn't fast enough the effect is the same as aliasing sampled ADC data. When I was doing that one, I had a 66MHz clock which I believe is faster than you can physically turn the device it was attached to, so I wasn't worried at all.

-- glen

latching readings from the disc isn't so much of a problem, it is how fast you can relay those serial bits error free to the device that needs the information.

As for the encoder, we used 2 8 bit parallel to serial encoder chips that had a latch function in them. all you needed to do was to activate the DTR line and this would allow the encoder to latch and transmit its data. The software in the computer had it all calculated on speed moves because if there was a great distance to travel, it didn't matter if you didn't get all the readings between, you just needed one here and there so that it knew ~ when it was approaching its destination, then the drive controls would slow down to get single readings to lock on position.

You only needed to make sure you were able to take readings before a full turn was completed since the last reading. This way we could do the binary roll over math for the wrap around. We never ran the machines fast enough for that encoder to have refresh problems to the CNC systems.

With today's fiber band width, that would no longer be a problem with

1:1 readings at high speed movements.

Jamie

You can always use the I and Q outputs themselves for another two bits of resolution. In fact, if the raw outputs are available (not Schmitt trigger) you can synthesize a double resolution by appropriate adds and inversions.

Jerry

--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

Like everything else, lefties want someone else to pay for their wishes. What about your congresscritter and her socialized medicine program so someone doesn't "have to work" to pay their medical insurance and can become an artist, or whatever. They're all smoking OPiuM.

...

When "plenty good enough" comes easily, that's the way to go.

I have no personal investment here. I am surprised by glib assumptions about accuracy and precision.

Once it works to spec, move on. That's much easier to do if the spec is loose.

One problem with photographic disks is centering them. (Another is stability of the substrate.) We made our finest disks by engraving on quartz or borosilicate glass. We put up to 10,000 lines in a 5" disk.

A friend manufactured "web riders" to measure the speed and length of cardboard corrugator output using photographic encoders. I mounted them using my cheap home lathe with extra attachments that I built. The best I could do for finding the center was about .0002" runout. Eccentricity is, of course, half that. I can see .0001" runout, but I can't do anything about it.

Precision is relative. A neighbor, hearing that I owned a lathe, asked if I would drill a hole in the center of an aluminum cup that he had had made. It was a cylinder 3" OD, 3" high on the outside, with a bottom. Bottom and side wall were 1/4" thick. My three-jaw chuck was good to about .003". Shimming one jaw (which one depends on the size of the work) easily beings that down to about .001. All of that is easy. With the four-jaw chuck, a tool-post microscope, and a fair bit of time, I can get down to .0002", as I mentioned above. So I asked him how accurate he needed it. His response amused me much as yours did. He said that it has to be absolutely accurate, with no room for error at all. That's why he couldn't use a drill press. I told him that if it were made last week, it didn't have a center because it wasn't round any more. Internal stresses are relieved when metal is machines, and what's left gradually relaxes to a new shape. Very accurate work is brought nearly to size, temperature cycled several times, given time to relax, and then finished. I told him that if it were round enough to suit his needs, I could manage the tolerances I mentioned above: 1 mil easily and

2 tenths with effort. His answer was revealing, "How much is a mil?" I answered, "About a third the thickness of newsprint." He looked shocked and said, "Hell! I can't even see that small!" So I used the three-jaw chuck and drilled and reamed it. (For better accuracy, I could have bored it.) Evidently, "no room for error" depends on the size of the error (and on the requirements of the job.)

Jerry

--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

...

yup. even with perfectly-positioned pickups, the speed of the "larger" portion of the disk (opposite of where the hole is) must be faster and the pitch is higher (unlike v^2/r, it's omega*r so the farther out goes faster).

and i think that the error in position of the sensors might be more comparable to "flutter".

no, i got it, i thinks. of course, a wow error is indistinguishable from a periodic component in the the true angular position of the shaft.

i think i understand Tim's alg, where there are three tracks, two that are quadrature (the 2 LSBs of Gray code), and a third that is the this MLS sequence with an extra 0 slipped in. so a minimum of 3 sensors are needed, whereas for N sensors and Gray code, you can always know, right from bootup, what the position is down to an error below the LSB. but, for where wires matter, Tim's method would be preferable but requires a turn of N bits and you get it with 3 pickups.

how does this centration error thingie relate to Tim's alg thingie?

--

r b-j                  rbj@audioimagination.com

"Imagination is more important than knowledge."

--
You pretend to be clever but, in the end, everything has to be
explained to you as if you were a slow child.

most

l

se an

lute

nded.

a

w the

ng

e

Jamie is a nitwit, and krw is another. If I casn lure them into posting something, I can usually find a way of ridiculing them.

I shouldn't have to explain to you what's going on - I've done it to you often enough.

-- Bill Sloman, Nijmegen

--
And you're the keeper of the keys?

most

an

te

ed.

the

ng

e

I wonder why Jamie thinks that? The top ranks of academic life - people like Feynman and Einstein - always had high citation rates, because they wrote classic papers, which pretty much everybody has to cite. Feynman was a famously good teacher, and the skills involved in setting out an idea in a way that students find easy to understand aren't all that different from those involved in setting out an idea in a paper in a way that other researchers find easy to understand.

The people who get stuck with teaching mentally incapacitated free- loading students aren't the ones whose papers get cited a lot.

A trifle ironic, since you clearly failed to get the point.

-- Bill Sloman, Nijmegen

Not at all. They cope by teaching students, doing research and writing papers based on that research. Counting citations is what university administrators do to find out how well their staff are doing. The other researchers read the papers and talk to other people who have read the papers and have a much more direct idea of who is good and who isn't, but citations do roughly track the real reputation.

Jim-out-of-touch-with-reality-Thompson doesn't appreciate that the people who are giving us climate change are the people who are digging up fossil carbon and selling it as fuel - who aren't academics.

Academics were the first to notice that climate change was happening, and they did it early enough that we could have cleaned up our act and limited to global warming to a degree Celcius or two if anybody had taken them seriously. Pity about that.

Right-wing nitwits do like this idea. It means that a hell of a lot of bright kids whose parents have less than average incomes won't go to university, which means that any country silly enough to adopt this idea is going to crash and burn in a couple of decades due to a shortage of the educated people who do the kind of work we do in electronics and a whole lot of other high-tech fields.

The US is - at the moment - importing quite a few educated people from other countries, but not as many as it used to. The rest of the advanced industrial world has caught up, and offers comparable living standards and better health care. It is fishing in the same pool, and if university education gets to be more expensive in the US than it is elsewhere in the advanced industrial world, the US is going to be even less attractive to potential immigrants.

Jamie can't think staight, and Jim is too out of touch with reality to notice.

-- Bill Sloman, Nijmegen

I can't help it if that free loading education didn't teach you the skills to comprehend common sense! Otherwise, there would be no point to get to. And you wouldn't be reading this, while scratching your head like your brothers in the jungle, oops, did I give your family secret away?

Jamie