What's new
What's new

Brown & Sharpe MicroVal Axis Error Fixed

mmcdade

Plastic
Joined
Jul 8, 2009
Location
Denver
Folks,

This is for people who get a 'axis error' message on a Brown
and Sharpe MicroVal machine. I'm posting this for the next
person who has to deal with this - my problem is solved but
reply if you want to ask questions or discuss it. Also, if
anyone knows how to set the dip switches on the controller
board for one of these cmm's, I'm interested.

Mark
----

Adjusting Brown & Sharpe MicroVal CMM Axis Quadrature Signal

11/25/09 - m. mcdade

Each 9-pin axis plug contains a small circuit board with
four pots for adjusting the quadrature signal waveforms.
I did not trace out and understand the exact function of
the circuitry supporting the sensors but at the time of
this work, both the x and y axes functioned well while
the z axis produced a 'axis error' message. Cleaning and
resetting the sensors on the scales helped (several years
ago, cleaning completely eliminated an error on the y axis)
but, eventually, the z axis could not be used any more and
required adjustment.

What I found was that the 4 pots on each little board in an
axis plug are associated with two test points on the counter
board in the cmm control (into which the plugs are inserted).
There are two test points for each axis. To adjust a given
axis, connect a scope probe for the 'upper' waveform (i.e.,
the one associated with the upper two pots when looking at
the plug board) to the upper test point and a scope probe
for the 'lower' waveform to the lower test point for the
axis. Set both channels to 2v/division on the scope and,
in auto mode, set the upper trace to the middle line of the
screen and the lower trace to the bottom line of the screen.
With power on and while moving the axis in question, you
should see sine waves from each channel. The top two pots
are used to adjust the upper signal (on the upper test pin
for the axis of interest). The upper pot of each pair sets
the offset of the waveform and the lower pot of each pair
sets the amplitude of the waveform. The following numbers
were derived by looking at the two good axes and served to
set the third so it now works perfectly. My z-axis voltages
were way off before my resetting and the fact that it
(mostly) worked for a long time indicates to me that these
are not terribly tight tolerance numbers.

top of upper waveform - 7.1 volts
bot of upper waveform - 2.7 volts
top of lower waveform - 3.5 volts
bot of lower waveform - 1.0 volt

---

Each scale unit has four pairs of emitters and sensors that
see out through a pattern of bars which one of the older
user manuals indicates are spaced a quarter space smaller than
the bars on the scale. It is easy to trace the wiring back to
the plugs and see that the pots on the plug are used to adjust
the voltage level to the emitters, apparently one per pot.
What isn't clear is how that translates to the two (unequal)
sine waves at the test points. These seem to form a
quadrature pair (the phase difference is clearly visible on
the scope) but how the circuitry gets from the four physical
signals to two logical ones is not clear to me.
 
Dear Sir!
Have you the same plugs and counter board like in photo? (attach in the post)
 

Attachments

  • B&S counter board.jpg
    B&S counter board.jpg
    100.7 KB · Views: 1,175
  • B&S head connectors.JPG
    B&S head connectors.JPG
    74.5 KB · Views: 980
Display monitoron Microval CMM

Can you tell me what type monitor you are using on the Microval CMM and does it use a DB-9 connection to control box, Also do you have to have en and mouse connected to control box for it to display. Thanks [email protected]


Folks,

This is for people who get a 'axis error' message on a Brown
and Sharpe MicroVal machine. I'm posting this for the next
person who has to deal with this - my problem is solved but
reply if you want to ask questions or discuss it. Also, if
anyone knows how to set the dip switches on the controller
board for one of these cmm's, I'm interested.

Mark
----

Adjusting Brown & Sharpe MicroVal CMM Axis Quadrature Signal

11/25/09 - m. mcdade

Each 9-pin axis plug contains a small circuit board with
four pots for adjusting the quadrature signal waveforms.
I did not trace out and understand the exact function of
the circuitry supporting the sensors but at the time of
this work, both the x and y axes functioned well while
the z axis produced a 'axis error' message. Cleaning and
resetting the sensors on the scales helped (several years
ago, cleaning completely eliminated an error on the y axis)
but, eventually, the z axis could not be used any more and
required adjustment.

What I found was that the 4 pots on each little board in an
axis plug are associated with two test points on the counter
board in the cmm control (into which the plugs are inserted).
There are two test points for each axis. To adjust a given
axis, connect a scope probe for the 'upper' waveform (i.e.,
the one associated with the upper two pots when looking at
the plug board) to the upper test point and a scope probe
for the 'lower' waveform to the lower test point for the
axis. Set both channels to 2v/division on the scope and,
in auto mode, set the upper trace to the middle line of the
screen and the lower trace to the bottom line of the screen.
With power on and while moving the axis in question, you
should see sine waves from each channel. The top two pots
are used to adjust the upper signal (on the upper test pin
for the axis of interest). The upper pot of each pair sets
the offset of the waveform and the lower pot of each pair
sets the amplitude of the waveform. The following numbers
were derived by looking at the two good axes and served to
set the third so it now works perfectly. My z-axis voltages
were way off before my resetting and the fact that it
(mostly) worked for a long time indicates to me that these
are not terribly tight tolerance numbers.

top of upper waveform - 7.1 volts
bot of upper waveform - 2.7 volts
top of lower waveform - 3.5 volts
bot of lower waveform - 1.0 volt

---

Each scale unit has four pairs of emitters and sensors that
see out through a pattern of bars which one of the older
user manuals indicates are spaced a quarter space smaller than
the bars on the scale. It is easy to trace the wiring back to
the plugs and see that the pots on the plug are used to adjust
the voltage level to the emitters, apparently one per pot.
What isn't clear is how that translates to the two (unequal)
sine waves at the test points. These seem to form a
quadrature pair (the phase difference is clearly visible on
the scope) but how the circuitry gets from the four physical
signals to two logical ones is not clear to me.
 
All,
We are putting a B&S Microval back on the road after being on blocks for a long time. Mechanically it is "fine" and I believe the scales were OK.

I am using the Deva037 USB Interface, along with a Deva018 Interpolator and Deva023 Pre-amp on each axis. The Pre-amp takes the place of the original pre-amp and features the 4 adjustment Pots just like in the above original posts.
Deva Electronics - Products

The Deva Documentation does not seem to offer any guidance on tuning the pots, or what we should be looking for on the scope waveforms.
The X and Z read OK on the interface side, the Y jitters when moved but isn't reading correctly. I hope this can be solved by tuning the pre-amp.

The above post is the most helpful, and only info I have been able to dig up as to what we are looking for when tuning. Its a starting place but I was hoping for a more definitive resource.

Any help would be appreciated. I have reached out to Deva but not heard back yet.

Thanks
 
Wery intresting thread..

Detailed Pictures of the internals (both side of the pcb’s) of the Deva023 and Deva 018 units would be much apreciated…

Considering geting these for my Tesa Micro-MS
 
Pot adjustments

From my understanding of it, (while trying to figuring out my missing electronics) The potentiometers adjust the current through the individual LED’s in the read head. Thereby changing the amplitude of the sine waves at the four outputs of the readhead when it moves across the scales.
The four potentiometers controlls the intensity of the four individual light beams in the scales, each representing the 0, 90, 180, 270 degree outputs of scale. In the end the 0 and 180 outputs are used as a differential pair, and the 90 and 270 is another pair. These signals represent the sin/cos output.

The pots shall be adjusted so that the pairs are balanced, ie identical amplitude of the positive and negative signal output of each pairs. The pairs shall also be balanced relative to eachother.
Note that the read head has to move to generate the sinus cosinus signals.

Each of the signals in one pair shall have aproximately 0,5v peak to peak amplitude, and about +2,5v offset. (To comply with the normal 1v sin/cos interface) When these signals are received by a balanced receiver the offsets will be canceled out, the sinus shaped signals will add upp. So you will get a 1Vp-p signal with no offset. And due to the balancing, induced cable noice will also be canceled out.
Then it is up to the interpolator to find subdivisions of the signals by using trigonometry, using the sin/cos signals to get a position between the scale lines… one increment of the scale (movement of one line) equals one period of the sin/cos signal).

By trimming the potentiometers, and establish identical amplitude and offsets, we get the best linearity and precission wich minimizes the errors in the trigonometry functions of the interpolator.

The offset error is the primary aspect to correct, as it will have the biggest influence in the trigonometry math functions.
 
Added picture

I am in the process of bringing an older Microval back to life. When I got it, nothing worked. I replaced all the hoses to get it functioning mechanically. Then I was able to get the hard drive from the old 386PC working and copied the MM3 software onto a new PC running windows 10. Next step was to install a serial card and get DOSBOX working to run the software. Lastly, I had to replace the power supply in the large interface box. After all that, the X and Z axes work, but not the Y. I swapped the X and Y cables and the issue followed the read head/cable. I bought a new/used read head and replaced it. Still had the problem. I suspect the issue is with the small circuit board in the connector. I read the description on how to adjust them, but it is not clear to me where the "test points" are on the counter board.

I've added a picture of the board I have. Can someone provide a description of where the test points are located?
 

Attachments

  • 20211024_163643.jpg
    20211024_163643.jpg
    98.4 KB · Views: 40
Last edited:
I found the test points and included a picture below for anyone who comes across this in the future. I connected an o-scope to the good axis and saw nice wave forms (also shown in pic). I then connected to the bad axis and didn't get anything like a wave form. The signal did move, but it was more of a tilt than a sine wave (also shown in pic). Any thoughts on what the issue is?
 

Attachments

  • 20211201_200407.jpg
    20211201_200407.jpg
    84.9 KB · Views: 27
  • 20211201_200107.jpg
    20211201_200107.jpg
    89.2 KB · Views: 31
  • 20211201_200923.jpg
    20211201_200923.jpg
    93.9 KB · Views: 30
Where are the potentiometers located? Separate units?

The read head gets its current through the potentiometers… if I got it right…
The potentiometers controls the current through the LED part of the LED/Phototransistor bridges… and the input to the interface electronic is the phototransistor side… and this shall be modulated by the interaction of the gratings and diffraction pattern on the glass scale and the filter on the read head.
Does the output signal of the faulty channel have an obvious dc-offset.? (Similar to the functioning channels)

(Not related, but valuable to me to build my own…what is the resistor values of the potentiometers and the resistors in that module?)

This is how I believe the read head and its interface is connected..

IMG_1128.jpg
 
Last edited:
The potentiometers are located at the end of the cord attached to the read head. I've attached a picture below. To get the resistance of the potentiometers, can I read them in place across the board?


The signals from the bad channel had a slight offset, but the main problem was they didn't have any waveform when I moved the axis. The picture below shows a side by side comparison of the good and bad axes.


Thanks for your help on this!!
 

Attachments

  • CMM waveform comparison.jpg
    CMM waveform comparison.jpg
    74.6 KB · Views: 33
  • 20211130_202436.jpg
    20211130_202436.jpg
    87 KB · Views: 31
Aha… it seams to be quite many variants on how this interface is built. But the circutry seams to be the same… there are pictures in other forum threads where this is a separate module with connectors on both sides.
The component values can be read in place. on the sides of the potentiometers there is probobly something like 1kohm written. Or only 1k… the same goes for the yellow box which is a resistor pack… i.e. multiple resistors in one package. Pictures of their markings would be much appreciated.

Regarding the first picture it seams like one channel has an offset, and the other not… but i suspect that one of the oscilloscope channels was in ac-coupled and the other dc-coupled.. the second picture indicates that it is not modulated… but the offsets seam to be similar to the first… the modulation is produced by the etched lines on the fixed glass scale together with a matching ‘filter’ on the read head. The filter has the same lines that the glass scales has, but in four sections, where each section is offseted in length by one quarter of the scale division (normally 10 or 20 mictometers or similar) the effect is that the light is blocked in some places and you get various amounts of light transmitted to the four detectors depending on the position of the read. A movement of the read head by one period (10/20micrometers) causes the four sensors to cycle through one period, where the four channels varies one time from lite to dark to lite again… (or similar wherever they start in that sequence).
It seams like you don’t have that modulation… might there be an missalignment between the readhead and the scales, so that it misses the grating lines..? Or might the filter be missing? Or might somebody changed the glass scale to a pice with the wrong spacing between the lines.? Can the read-head be of the wrong type.. i.e.. wrong filter line spacing…?

Thats my initial thoughts..

Does varying the pot setting a small amount change the behaviour of the outputs… i.e offset level… that would indicate that the read head electronics is working…

Do all four signals from the read head look the same?
(My writing above assumes that the signals you measure is before the diff-amp and that there should be four test points for each read head. I.e before the signals are combined by a diff-amp to create the pure sin/cos signals to the interpolator circuitry)

Edit, also a picture of the underside of your cable interface would be nice for my replication of the interface)
 
Last edited:
For anyone interested, I have the manual and some documents for a mid-1990's Microval and have scanned them into several PDF files. PM me with your email address if you want access to them on Dropbox. In aggregate the files are too large to email.

They cover the Micromeasure II Plus software as well as the hardware and mechanics. Sorry, but there's really not much info on the internal electronics other than an over view of the connections.

Mike
 
Mike, I want to say thanks again for sending me the manual. It helped a bunch to figure out how to route the hoses as they were all rotted away and I had no way of figuring out how they were routed. You can feel good that your help has gotten me 90% of the way to a working machine! Just need to get the Y axis reading.
 
I have that fun job ahead of me... :)
I also got the pdf's from Mike... Thanks!
Have not figured out what tube to use, neighter their dimensions...

My first goal is to build interfaces for my scales... :)
(I have no electronics at all at this moment)
 
I checked the resistance values of the pots. They ranged from 8 to 400ohms. But there really wasn't much different between the working and bad axes. So I figured I would take the next step and put the read head from the working axis on the non-working axis. The o-scope showed a nice waveform, so this proved the problem is with the read head. I didn't expect this because I had bought another read head from ebay and swapped it. The problem persisted with it. I assumed the probability of someone selling a bad head was low.

So the question is why doesn't the read head work? I have a couple of theories.

First, the original "bad" head and the one from ebay do not have what looks like a glass cover over the read head. I suspect this is to protect the LEDs and sensors from debris and damage. I investigated the original bad head closer and it appears the cover was removed because there is residual glue. My suspicion is the axis went bad and someone tried to diagnose/repair it. Based on this, I think not having the cover is likely not the cause.

The next thing to investigate is if the LEDs are working. I checked the good head and it does not produce visible light. I remembered my camcorder can detect non-visible light, so I used it to look at the read heads(see pic below). I found that the LEDs on both the good and bad heads are working.

So now I am on to trying to determine which sensor is bad. My thought is to measure the voltage coming off each sensor to see if one of them doesn't output a signal.

What I need to figure out is how to accomplish this. Where can I "tap" in to look at the signal on the o-scope? It is getting late, and I've made enough progress for tonight. Will investigate this tomorrow and let everyone know what I've found.
 

Attachments

  • 20211202_212241.jpg
    20211202_212241.jpg
    85.5 KB · Views: 25
  • 20211202_212212.jpg
    20211202_212212.jpg
    80.2 KB · Views: 25
  • 20211202_212806.jpg
    20211202_212806.jpg
    79.3 KB · Views: 31
I have not removed and looked at my read heads..
There really shall be a filter there. But how this looks like is unknown to me..
Maybe its placed closer to the phototransistors...

The outputs would be the test pins you already measured... allthough this might be after buffer amplifiers...

I tested my phototransistors by measure their resistance while shining a flashlight onto the readhed (mounted on the scales) i could then see some change in the resistance of them.
One multimeter lead was on the +v signal and the other on the respective output leads.. se my schematic above. I beleeve the pinning shall match your cables.. +v is pin 11. And the phototransistor outputs are on pin 1,6,2,7 respectivly... you can measure this with the connector loosen from the other electronics... no power needed, just a multimeter to measure resistance... a couple of 100k ohm up to some Mega ohm was my result, but it depends on the amount of light hitting the sensor.

If the filter of the read head was missing, I think you would get the results you have sowwd us....the filter is half of the magic...

Your picture showing the read heads really looks like it got grating lines upon the glass cover at the exact right place...
I really think you should test the head with this 'cover' back on...
 
I checked a couple of things.

First measured the voltage coming from each of the detectors with the machine static. I observed ~1.14v coming from each of the sensors. I also saw 10v going to each LED. I did this with the good and bad heads and got the same answers.

Next I hooked up an o-scope and measured the signal coming from the sensors while moving the axis. I thought I would see something like a square wave. I did not see anything other than noise. The result was the same with good and bad heads.

Any thoughts?
 








 
Back
Top