Sony Magnescale alignment procedure

One axis on my grinder readout is reading erratically and the symptoms
point to the pulse detection/shaping circuit being out of alignment.
The scale indicates approximately the proper distance travelled, but
only counts in one direction. In other words, move forward 1 inch and
the readout will indicate approximately 1.0000; move back to the
original position and instead of returning to 0.0000 it reads, more or
less, 2.0000.
I've swapped scales and the problem follows the scale, not the readout
The scales are Sony p/n SR-2711 and the readout is an LF-200. I
believe they date from around 1985.
I've fixed similar problems in optical encoders, but this is a
Magnescale and most of the circuitry is encapsulated in what appear to
be three mini circuit boards - the long brown devices in the photo. I
haven't been able to make an informed guess as to the function of the
several pots. The optical encoders I've fussed with have one pot for
each channel that tweaks the switching point of a comparator; this is
clearly more complicated.
formatting link

Before I hook this up to a scope and start fiddling with the pots,
does anyone have or know of an aligment procedure for this unit or any
Reply to
Ned Simmons
Loading thread data ...
The way I typically approach faults (without service literature) when there is an identical, known good duplicate item, is to check input/output and various other circuit points on the good one, making notes for comparison when checking the faulty one.
The brown boards, as you probably know, are custom circuits, probably long out of production/NLA.
I'd just naturally suspect the blue tantalum capacitors, just because of their age.
If I found any obvious abnormalities in the signal comparisons, which is very likely, I'd likely replace the 6 capacitors after a close inspection of all the solder connections, as the easiest procedure/age upgrade.
There are a number of silkscreened designations on the board, which is a lot better than none.
It might be worthwhile to check the feedthrough vias (2 visible) for zero resistance top/bottom, or just heat them up and fill with solder, and check all the other soldered connections with a magnifier and reflow any that look suspicious.
The JAE jumper or the main connector could have some oxidation on the contacts.
Marking the pot positions clearly before changing them is a good practice.
The cable lead terminals appear to be soldered instead of crimped, but probably worthwhile to check them.
Reply to
That's a pretty good bet. And Sony is very stingy with information on this stuff.
Easy and cheap to replace. How likely is it that 25 year-old tantalums would be bad? Before the bad channel stopped working entirely it would sometimes not work properly (though the symptoms were different than the current problem) until the readout was powered up for a few minutes, especially if it had not been used for a while. At that point I was suspicious there might be a bad wet electrolytic cap that was reforming under power. Can tantalums exhibit the same behavior?
Yeah, and at least they are in English, though a bit cryptic.
Again, it'd be easy to reflow all the joints. I'll do that.
I pulled it and reinserted it with no change, but I'll do it a few more times. All the connectors are gold plated and appear in good condition, but I know that's no guarantee.
They're spring connectors on header pins. I've pulled and re-seated them, but I'll check more thoroughly.
Thanks, Bill.
Reply to
Ned Simmons
Not necessarily cheap, there are wet slug tantalum capacitors for military/aerospace use in the range of $30 each. Unless there's a damaged case, tantalums fail in short circuit, i.e. if there's voltage across the pins, it's good.
If it counts UP always, the possibility is that there are two channels with crosstalk, in a quadrature counter. Quadrature works by having (for example) a pair of gear tooth sensors offset, so if the gear hits sensor 1 first, then sensor 2, you know it's moving CCW; if it hits sensor 2 first, then sensor 1, it's CW. If both sensors report at the same time, direction gets confused but you KNOW it moved. If there's a way to clean or reregister the sensors and the scale, I'd start there. Then look for wiring common to both sensors (the power supply components in the sense head, for instance).
Reply to
Does this have a smooth rod as the scale and a read head with a hole in it that slides over the rod? If so, the rod is not symmetrical, the magnetic "whatever it is" is polarized on the side of the rod. So, try turning the rod in small increments, like 30 degrees, until it starts to count better. There is a read head floating inside the outer shield box, and it can get knocked out of alignment and ride crooked on the rod. There should be a strip of spring steel that can be bent so the head rides parallel to the rod.
Reply to
Jon Elson
[ ... ]
Hmm ... if those are what they *look* like, I would expect only resistors and capacitors under the coating -- mounted on a ceramic substrate with silk-screened metalization for conductors. And the coating is fired on, like for a ceramic capacitor, so solid state devices (transistors, diodes, etc) would not survive the firing temperatures.
Just how old *is* this thing? The connectors look more modern than the packaged circuits would suggest.
Hmm ... it might be that the comparator is at the other end of the cable, and the circuitry is mostly to match impedance so there is no reflection of pulses.
Looking at the wires, and assuming that ground is carried by the shield, while the red is likely the power to drive the LEDs -- also out of period with the packaged circuits, which were from the 1950s IIRC. Anyway -- that leaves five signal wires -- two for the 'A' sensor (one inverted) and two for the 'B' (also one inverted), which produces the quadrature. That leaves one for the Index (saying that it has reached a certain reference point on the scale. This matches the five pots, which probably set amplitudes for each signal fed downstream to the comparators. Some systems expect to be fed sine waves from the sensors and use intermediate voltages to interpolate extra resolution.
That I don't -- sorry.
I would first try checking the waveform from a constant speed for each of the signals (expecting one of them (the index) to only change once at an extreme end of the scale.) If any one has a lower amplitude sine wave than the others, it is suspect. But it might be the comparators back at the electronics/display box instead.
Good Luck, Don.
Reply to
DoN. Nichols
That's the one.
The scale is very heavily built and has a rather elaborate suspension system for the rod. It'll take a bit more diassembly to see what's required to rotate the rod. Thanks for the tip.
Reply to
Ned Simmons
It's been a number of years since I looked up the working principles of magnetic linear scales, but Jon's recommendations about checking various mechanical parts' fit/alignment/orientation is likely the best place to start looking.
The tantalum cap problems I've encountered have been dead shorts, as Whit commented, and if the current source is sufficient, the outer epoxy layer may develop a pinhole with smoke shooting out of it (if you happen to be watching as the power is applied). Some tantalums seem to just fail abruptly when they get old.
The 3 silvery looking capacitors nearby are poly film types, and if it turns out that there are signal abnormalities not related to physical/mechanical parts alignment, it would probably be a good idea to replace the tantalums and polys just because they're old. Of course, replacing components can require service/calibration adjustments, but having the properly working circuit is an advantage.
One other technique that has proved to be worthwhile, is to just rock the pot positions back 'n forth a little, for a small portion of the rotational range, after they've been clearly marked for their original positions. This can sometimes create a better contact, in the event of a bit of oxidation on the wiper.
If the outcome remains a mystery after making a reasonable amount of comparisons to the working circuit and scale, the only option may be looking for a replacement on eBay or surplus places that sell older encoders, linear scales and similar instrument components. I remember seeing a couple of surplus websites like that, but that would've been over 5 years ago.
I don't believe there is any way to find out what discrete components are inside those custom hybrids, without destroying them.
Reply to
It appears to be an epoxy coating. There are no active components on the circuit board, and I'm pretty sure there are none in the reader head, so there must be some in those potted circuits.
As far as I know, about 1985, maybe as early as 1980.
I don't think so. I'll know better once I hook it up to a 'scope.
No LEDs -- Magnescale = magnetic scales.
I did find a pinout which indicates that the pickup for each channel plus the index uses two wires, so all six leads are accounted for.
I'm guessing something similar. Pots to set the gain on amplifiers that boost the pickup signals, and pots to adjust the comparators. Though that'd imply six pots -- two per channel.
I did find a patent issued to Sony for a later generation magnetic scale with circuits. I haven't read it carefully yet, but it appears there are some clues there.
Reply to
Ned Simmons
Are there any patent numbers on the scale or reader? Looking up the patents can be very useful.
Also, Sony may still have data on the scale.
Joe Gwinn
Reply to
Joseph Gwinn
No patent numbers that I can find. But as I mentioned to DoN, I did find some Sony patents that look to be on point.
Sony is remarkably stingy even with data on current products. But one of my customers uses a slew of the Sony tape scales, so I might call in a favor and see if they can get anything from their distributor.
Reply to
Ned Simmons
I'm reading this with a bit of interest. My bridgeport has what I believe are magnescales. They look old. Blue cold cathode display, some what dim, two joysticks for zeroing, setting presets. Reads out to 0.0001" which is overkill on a mill.
Does this sound like your scales? Just curious, can't remember what model but I found squat about it on the web and I'm pretty good at finding things. Not as good as Iggy, that man is a wizzard from some of the stuff he has drug up.
Part of me wants it to die so I can buy a DRO with features, the other part of me is work with it, more money in the toy budget.
Wes -- "Additionally as a security officer, I carry a gun to protect government officials but my life isn't worth protecting at home in their eyes." Dick Anthony Heller
Reply to
Definitely a different readout. Mine's an LF-200 like these on guess who's page:
formatting link
This is the good scale, still on the grinder:
formatting link
The pic I posted before is the end of the cable that plugs into the readout, with its cover removed. I'm not sure that this interface/connector was the only one used with these particular scales. There may have been others for use with different readouts.
formatting link
Weren't you playing with Alibre a while ago? A good CAD package is much handier than a full featured DRO, IMO. Though a midpoint function and full keypad is nice.
Reply to
Ned Simmons
joysticks for
Yup. Different but I think I saw that style head on a height gage at work. Not sure it is still in use.
Mine don't look like that.
me is work
I'm still using Alibre. Steep learning curve, my cad exprerience started at Autocad 2.6 and stayed 2d until last year. I plan to upgrade my level at the 1 year point when my maintenance runs out. Motion, configurations, and sheetmetal could be handy. The sheetmetal would be handy for projects at work. We turn, mill, have an an Amada cnc turret press along with an Adira CNC press brake. Pretty sweet having the sheet metal stuff when you want to make something out of sheet metal for in house use.
Drawing it using autocad, "not so much" as Latka would say.
The thing I crave is multiple origins. The mill at work, that has a much newer Sony scale system, I indicate the right side of fixed jaw as co-ordinate set 1.
Depending on what I'm making and how many, having a few other origins programed is a time saver.
Bolt circle, no problem, my palm pda has a program for that.
-- "Additionally as a security officer, I carry a gun to protect government officials but my life isn't worth protecting at home in their eyes." Dick Anthony Heller
Reply to
No, these packaged things look so much like Sprague parts, made in the mid-late 60's, just before integrated circuits took over. The rest of that module DOES look newer than that, however.
But, the MAGnescale is magnetic, so there would be no LEDs. It has a smooth stainless rod with SOMETHING inside. There is an excitation coil and two pickup coils, so the signal scheme is exactly like a resolver or inductosyn. You excite with a sine wave at some modest frequency like 400 Hz, and as the head is moved, the signal on each pickup coil rises, falls and then reverses polarity with respect to the excitation. The two coils pick up this signal in quadrature. The converter box interpolates the signal to increase resolution.
I played with a damaged one of these units at a shop in Wichita. I got fairly good signals out of it after repairing some TOTALLY hair-thin wires in the head, but we never got a decent signal out of the converter box.
Reply to
Jon Elson
I think you'll find that the SOMETHING'S inside are precision ground bearing balls.
there is a British manufacturer who still makes them like that. They were fearured in "Model Enginners Workshop" a year or two ago
There is an
Reply to
formatting link
Reply to
David Billington
The patent literature will tell you a lot. Look for patents assigned to "Sony Magnescale". There are ~30 such patents, some with relevant-sounding titles.
Google Patents is a convenient place to search, but be aware that patent drawings on google are often mangled. If so, use to get an unmangled copy.
I don't think that Magnescale and Newell encoders work the same way, even though both are "magnetic".
Joe Gwinn
Reply to
Joseph Gwinn
Hi jon
Did you rewired the coils? I need to rewire the tiny inside the head but the double excitation coil seems impossible to do manualy. Can you help me with some advice?
Thanks Carlos Lisbon, Portugal
Reply to

PolyTech Forum 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.