Will this laser idea work? On topic

So I was just looking at the ED3 display unit documentation. I never used one, so bear that in mind as you read my comments.

The key to getting the resolution you need seems to be the X4 mode. I'd feel better if they said exactly what is different between these two modes. Reading between the lines, the assumption is that X1 gives one count per C of CPR and X4 would give a count for every quadrature transition.

When you are saving configurations into the ED3, seems that you can either save it temporarily or permanently. I guess you know this, but just mentioning it, in case.

Can't remember if you told us -- did you try moving the shaft a full revolution and see if the count makes sense for a full revolution?

So, I'm just speculating, but it seems to me that the ED3 configuration is the most likely place for things to have gone wrong. After a few sanity checks, maybe a call to USD support is in order. I think they are pretty friendly.

Keep us posted, and sorry for my missing the factor of 4 that quadrature gives you in my comments about CPR in the other message.

-Rex

Actually Brian, I want to do both, watch for change AND measure movement. I did not realize the mirror would double the movement until it was pointed out to me. Then it clicked and I remembered. The laser is

O.K., today's silly question- Maybe I'm missing something here but how does the readout know to count up or down with only the raw clocking out from the encoder? Seems to me that without the quadrature info the display won't know whether to increment or decrement-

O.K, so lets say I have missed something in the discussion & the unit is counting in the right direction, if you reverse the encoder's rotation how many pulses does it have to generate before it knows it's going backwards??

Enquiring Minds want To Know-

H.

This won't work with a simple beam. The laser spot will inevitably be many times the size of the motion of the spot you are trying to measure. We have hashed this over and over with people who speculate they can align their round-column mill drills using a laser.

I'm always surprised by people who think laser pointers have near-zero angular diameter. Not the case.

It's inherent in the quadrature signal. If you look at a timing chart of the signals you can see. The A and B channel transitions are offset from each other by 1/4 of a full cycle. Say at some position both A and B channels are in the high state. One or the other will go low if it moves. If A goes low, it is moving one direction. If B goes low it is moving the other direction. Which one means up and which down is a matter for configuration of your appplication, but the direction for every transition can be determined by the previous state and the new state.

According to Eric R Snow :

A nice 'scope.

The main question is whether he can identify the ground clip on the probe as described. Most Tektronix scope probes of relatively recent vintage have a groove about half-way back from the tip to the entrance point of the cable. A clip, sort of like a small version of a hairpin clips into this to connect the ground clip to the scope probe's ground.

Older Tektronix scope probes, as well as ones by Hewlett Packard, had a threaded hole in the side of the probe, into which the end of the ground clip screwed.

Recent Tektronix scope probes have a coaxial tip, with a sleeve surrounding a tiny point. An extension slips onto this for most work, with a spring-loaded hook which is exposed by pulling back on a skirt on the probe tip.

For really high frequency work, an alternative tip slips on, which picks up ground from the sleeve, and has a very short projection for the center conductor.

Your color scheme does not sound like any Tektronix scope probe which I have seen.

Note that any probe needs to be tuned to the scope. This is done by connecting the probe tip to the calibrator output (a square-wave pulse), and adjusting one (or more for higher frequency probes) adjustments to maximize the squareness of the displayed trace at the corners. Adjusted too far in one direction, and the trace will curve gently from the vertical to the horizontal. Adjusted too far the other direction and the trace will overshoot beyond the horizontal, and then return to it -- sometimes with one or two ringing overshoots in both directions.

If there is more than one adjustment for the probe (usually in body of the connector to the scope, but rarely on the scope probe itself), one adjustment will adjust the squareness of the corner, and the next will adjust a portion of the horizontal top or bottom of the square-wave train. I've seen some (high voltage probes, with 10KV maximum voltage) with four or five adjustments in a large box at the connector end.

Some of the older Tektronix (and others) probes will have a knurled knob where the flange of the probe joins the body. Loosen this, and unscrew or screw tighter the tip end to adjust the response of the probe.

Once your probe is tuned to the scope, you should not need to re-adjust it until you move it to another scope (Or for a scope with plug-ins, until you move it to another plug-in).

You probably don't really need to tune it for what you are doing, but if you don't, you may see the pulse height change with speed of motion (as the pulse never really reaches the top of the waveform).

Indeed so.

Excellent scope operation at a distance course.

Enjoy, DoN.

I think most of the important stuff has been covered. From your laser dimensions I'm assuming a HeNe laser and thus a pretty good beam quality, not a diode laser pointer. Check your spot size to be sure it's okay, at your target, after reflection from the mirror, because that spot size will ultimately limit how small a motion you can detect and it will grow proportional to distance so you can't beat it by getting further away. Only point I haven't seen mentioned is that the mirror needs to be mounted on the end of the shaft so that the shaft axis lies in the reflecting plane, and the laser beam needs to be perpendicular to the shaft axis. That way you just get mirror rotation as the shaft rotates. Any distance off-center and you get a mix of translation and rotation and there goes your measurement. I'd suggest drawing two light rays some rotational angle of the shaft apart, both on center and some distance off center, to get an idea of what precision you need for this setup.

-- Regards, Carl Ijames carl.ijames at verizon.net

For low-frequency use, sometimes the ground at the probe is best omitted. On rare occasions the little ground clip has been known to detach from the ground point (in high sproing factor installations) and flip deftly through the air to land on the nearest high voltage terminal point.

With predictably exciting results.

Not that *I've* ever done this. Ahem.

Jim

Howard, The encoders I'm using have a clear disc with a bunch of lines arranged radially and measure about .1" long from the edge toward the center. The two line detectors ( A and B) are arranged in a "phased array". This means that the detectors are placed such that first A detects the leading edge of a line, then B detects this edge, then A detects the trailing edge of the line, then B detects this line. Actually, the line A and B use can be different lines as long as the detectors are placed properly and the lines are evenly spaced (which they are). It works like this: The detector A would see the dark leading edge of the line, then as the disc turns detector B sees this edge but sees it before the disc has turned enough for A to see the trailing edge. Then the disc turns a little more and A sees the trailing edge. Then B sees this edge but sees it before A sees the leading edge of the next line. Kinda wordy explanation. Surely someone here can say it better and clearer. OK, now that we know the pattern of detection we can see that A will lead B when the disc turns one direction and B will lead A when turning the other direction. So that's how the display knows the direction. This detection scheme also allows both single and quadrature pulses to be read by the display. If the display only increments one count on the leading edge of A then the number shown will be the number of lines that have passed A. But if done the way explained above the display is able to count 4 pulses for each line passed. I don't know, but I suppose that the display still uses detectors A and B to tell which direction the disc is turning, even when it is only showing the actual number of lines that have passed one detector. Eric

DoN, Maybe probe is the wrong word. What I have is a coaxial cable with a twist on connector that fits the input on the 'scope. The cable is about 18 inches long. At the other end two wires with clips on the ends are coming out. The red wire is connected to the center wire of the coax. The black to the outside conductor of the coax. It has molded into the strain relief: POMONA ELECTRONICS and the numbers

O.K. I know those, and I would never call them a probe. At least, you don't have to worry about adjusting the vertical V/Div for that, it is a 1:1 connection. And until you get to high frequencies, you probably don't need to tune it for the scope's input characteristics, either.

I probably would just call it a BNC to clip lead cable, FWIW. The twist-on connector is a BNC (Sometimes attributed to "Berkeley Nucleonics Corporation", but I think that the "B" is for "Bayonet" and that it predates the company. The "N" would seem to be for the Type-N connector, except that is a larger connector for a larger cable. But the other features (aside from the Bayonet locking ring) are quite similar to the Type-N scaled down.

There is also a "TNC" connector (Threaded instead of Bayonet), but that is less common, and less convenient for most uses.

You really should get the proper 10X probes for your 'scope, as they have the advantage of not loading down the circuit under test the way the capacitance of the coax cable does. Here is an auction about to close for a nice set:

It has gotten fairly steep already, so you might want to wait for another set at a better price. This set started at $9.99, which would be a nice price, but it is now up to $56.00 with about an hour and a quarter to go.

This auction:

has a buy-it-now, and a more affordable price, but I'm not sure who actually made these probes. I'm pretty sure that they are not by Tektronix.

This one:

is still pretty affordable, and the probes are about the right vintage for your scope, I think. (Quite similar to the ones for my Tektronix

There are several in close sequence starting with:

which are starting at an attractive $1.00 each, but those are 1X probes, not 10X.

My ebay search was a simple:

Tektronix scope probes

and it turned up a lot of scopes as well.

Enjoy, DoN.

(snip)

True, barring exotic lasers and optics. Now having dismissed what won't work ..... let's devise what might work. Same idea, Eric ....optical leverage .... but without the laser. Tape a scale on the distant wall. View the reflection of the scale in the mirror using a powerful (spotting) scope, preferably one with a reticle. The scope can be close to the mirror because it's focussed on the distant scale via the mirror.

Your line of sight via the mirror will move 0.1" at range of 159 inches, as you noted. The distant scale will appear to move that much when the mirror rotates. With a 20X scope, it will be as if you are viewing the scale from a distance of about 8 inches. You can easily discern 0.1" from that distance. At least I hope you can!

Your scale may need to be marked with lines .025" thick for clear visibility in your scope -- depends on the scope. I can very easily discern .025" lines at 100 feet thru my inexpensive spotting scope, though it doesn't have a reticle. A good riflescope of 20X or more magnification, or a transit or theodolite, would work well.

That got a bit confusing at the end, I think.

One cycle as in CPR (cycles per rev) spec'd for the encoder is one full 'line/not-line' sequence for one channel. If it is only counting the lines that pass one detector, it can't turn 2500 CPR into 10000 counts. It has to count all the transitions (on to off, and off to on) in both the A and B channels to get the 4 counts per cycle that make 2500 * 4 =

On 3 Oct 2005 17:18:58 -0700, with neither quill nor qualm, jim rozen quickly quoth:

And what, may I ask, is a LOW sproing factor installation? I've never seen one except when a ground plane hit the edge of a single-sided circuit board. How often do you see those nowadays, hmmm? ;)

Across separate HV and/or ground points together, of course.

The Fry Factor is high.

We've only heard rumors of it being done, right? I've heard of screwdrivers being gently laid upon circuit boards with interesting results as well. And probe tips hitting more than the intended single IC pin during probing, and... Sure glad that wasn't me.

On 4 Oct 2005 02:26:39 GMT, snipped-for-privacy@d-and-d.com (DoN. Nichols) wrote: Thanks DoN. Eric

His 1/10000 revolution is going to give a spot motion of 0.2 inch at a distance of 13 feet. He won't have any trouble seeing this with a halfway-decent laser. If his laser has a divergence of 2 mrad full angle, at 13 feet the spot is only a little more than 0.3 inches diameter, and I'll bet he can see a spot motion of 0.05 inches with a

You're right that people don't generally realize how big and fuzzy the laser spot is going to be, but in this case it should be no problem.

I would use an autocollimator or jig transit or theodolite if I needed to resolve 1 arcsecond or 10 arcseconds, but he has to resolve only 130 arcsec. When I set up an autocollimator measurement I use a diode laser attached to the autocollimator to get it roughly aligned; with the target 13 feet from the mirror, it's easy to measure angular changes quite a bit smaller than 130 arcsec using just the laser and mirror.

D> Eric R Snow writes:

Oops. Range is 159 inches, not 159 feet. You should be able to see an ordinary machinist's scale thru a 20X scope. Not all scopes will focus that close (about 13 feet) but some do. Mine does.

I tried this. From 200 inches I can easily discern the 1/64" gradations on the ruler of a machinist's square when looking at it directly thru the scope at 20X. Resolution with the mirror I had handy was not nearly as good. The mirror would have to be a good quality first-surface mirror -- mine was a crummy inspection mirror. Good first-surface mirrors are easily found at surplus stores for a buck or two.

The "spot" from my $9.99 HF laser level is about .25" dia at about 17 feet -- but there is a central "bright spot" that is probably only about .050" dia. So, with a good first-surface mirror, I think your experiment would work OK with a laser. A small prism may also work well.

All I've had time for today is to make a quick measurement to a closer target. At 75 feet the spot is 1.5". And I calculated that the spot will move .568" at 75 feet. So, the spot is about 3 times as big as the movement of .0001". But, the edge is pretty well defined, so I can see the edges move. If I draw lines at the edges I'll be able to see how much things have moved. I have a couple lenses that I used on a laser pointer to get a smaller spot at 30 feet. They worked pretty well. But laser pointer spots are not round and so any improvement really helps. I was surprised how much the beam expanded. As for the other ideas posted here, they have merit and if I can borrow a 20 or so times scope these will be explored too. It sure is fun to learn new stuff. Thanks, Eric