What's new
What's new

Talyvel 1 adjustments Pt.1

daryl bane

Titanium
Joined
Mar 12, 2002
Location
East Texas
I am doing this for posterity as detailed info about these units seems to be hard to get. Sorry for the stream of conscience format. This primarily is about the read heads and what I have discovered so far. The unit themselves are really pretty simple. You can access the internal assy. removing four screws on the side of the unit and then both sides come off. There is a pendulum suspended by 5 tiny beryllium wires. These are quite robust and breakage is rare. There is a bar (armature) that sits on the pendulum between two transducer coils. You can check these coils with an ohmmeter, to make sure there is equal continuity between them. Use the smallest voltage on your meter and do it quickly, or you can damage the coils. There are two micrometer knobs that are attached to a spring loaded bar that the whole pendulum assy. is bolted to. You can screw the two knobs all the way up until the bar contacts the two ground pins and can determine zero to the base/frame. The knob scales can be adjusted by loosening the two screws on the top of each knob to adjust the scale. There is a threaded radiused post on the bottom that keeps the unit centered due to being sprung, this also has a locking screw. This post is adjusted that when you screw the micrometer knobs all the way down, the unit rests on two rubber pads and this post, so that the wires are now out of tension and not under pressure from transport. The pendulum unit has a tiny silicon damping blob located on the bottom of the pendulum assy. and this blob is sandwiched between the movable assy. and the mounting base. There is a clear plastic threaded plug that screws into this base, and with a loupe you can see when the radiused front of the plug just touches the silicone blob. There is a copper tension arm that goes on the top of the plug to keep it from moving. There are threaded stops on either side of the assy. to limit back and forth pendulum movement. There are threaded holes on each side of the pendulum assy, that are accessible thru holes in the main housing via rubber plugs. I have not yet done this but I believe the following is to adjust or center the armature bar. You insert a shop made threaded tool into the vacant threaded holes and push the armature bar in or out for adjustment. There is a little arm that sits above the armature bar to do a fine adjust in use. The arm is magnetic and is mounted to an eccentric screw arrangement that is accessible thru a hole in the side of the case. That is all for now.
 
This is pretty interesting. I partly disassembled an early Talyvel years ago. The remark by the OP about there being two coils has me revisiting my understanding of how it worked. Because there didn't appear to be a magnetic core between the two coils like in a typical LVDT or a half-bridge VRDT, I figured the sensor is capacitive. I hadn't considered that the armature connected to the pendulum could produce a differential reluctance in two coils but I bet that's how it works. I never did measure the excitation frequency. Do you know what it is? If it's less than 20 kHz, the sensor is likely a differential transformer. Given the size, if it were capacitive, you'd have to push frequencies on the order of 100 kHz to get the impedance below 100kOhm.
 
I am having a "top man" :) looking at it, but his time is valuable and looks at it when he can. They put it on the oscilloscope and found some electronic issues, but nothing concrete enough to share at this time. Still will try to post some pics.
 
Here are some pics. One is the electronic board and the others are the read head each side. You can see the two coils and the bar in between. This bar is suspended by the 5 wires and the movement between the two coils.
thumbnail.jpgIMG_0042.jpgIMG_0045.jpg
 
Here are some pics. One is the electronic board and the others are the read head each side. You can see the two coils and the bar in between. This bar is suspended by the 5 wires and the movement between the two coils.
View attachment 325076View attachment 325077View attachment 325078

These pics look just like the level I took apart. This reminds me of another detail: What's the purpose of having a 5-wire suspension holding the pendulum? Basic kinematic principles says that each wire *can* constrain one degree of freedom--and we'd like to constrain 5. But all the wires to the pendulum come from above: One purely vertical wire near the center of the pendulum, and then two pairs of wires in a vee-configuration at the ends of the pendulum. This looks overconstrained to me. To my understanding, the wires don't prevent side-to-side motion perpendicular to the intended plane of motion of the pendulum. Basically, I don't get what degree of freedom the 5th wire in the center of the pendulum is intended to constrain. The guys at Taylor Hobson weren't fools, so I'd bet there's something I've overlooked here...
 
What's the purpose of having a 5-wire suspension holding the pendulum?

The patent for this makes interesting reading.

The 5 wire suspension can be thought of as follows. Suppose we are suspending a car rather than the Talyvel pendulum. We suspend the car from three points: one is at the center of the front bumper, the second is at the center of the rear bumper, and the third is on the passenger side, just behind the front door. The first two points support most of the weight, the third point just a bit of it.

Two wires connect to the center of the front bumper, and go upwards to the left and right (as seen by the driver) to attach to the ceiling of the garage. So from the front seat, you see these two wires forming the letter "V".

The two wires connected to the center of the rear bumper are the same form. If you look our the rear view mirror, you see a "V".

That accounts for four of the five wires. The attachment points are actually not in the middle of the bumpers, but just a bit to the driver's side of center. So if you suspend the car like this, it will slowly rotate around the axis that passes through the two attachment points, and tip to the passenger side.

The fifth wire attaches to the car just behind the passenger side front door, and goes to the garage ceiling. This wire prevents the tipping described in the previous paragraph.

The purpose of this? With these five attachment wires, the car can only move (swing) in the back and forth direction. It can not rotate about the vertical axis or either of the horizontal axes, nor can it move in the vertical or left/right direction (as seen by the driver). So five of the six degrees of freedom of a rigid body are constrained.

Here is an interesting article about the inventor, Richard Edmund Reason:
https://homepages.abdn.ac.uk/npmuseum/article/Taly2010.pdf
 
Last edited:
That's a wonderful explanation, Ballen. I didn't notice the asymmetry of the vee-suspension at the ends of the pendulum with respect to the centerline. Could you put up a link to the patent? If the suspension wires were rigid rods, I see how the design constrains 5 degrees of freedom. However, with wire flexures that are rigid only along their length in tension, I still don't see how this design constrains motion along the line connecting the driver and passenger doors. Maybe with gravity loading the pendulum and the flexures being pulled taut, the wires still have a tiny bit of compressive rigidity constraining that DOF, as if they were rigid rods?
 
A full listing of Reason's patents can be found at the end of this short biography. A description of the Talyvel begins at the end of page 451.

https://royalsocietypublishing.org/doi/pdf/10.1098/rsbm.1990.0039

The wires are very fine, and made from beryllium copper. I am not sure how they are heat-treated, but suspect that they are first formed into the V shape and then annealed, hardened, and tempered. If this was a dynamic (moving) system then you would be correct that the wires can not be treated as rigid rods. But in the Talyvel, you move the head to a new position, then wait for it to settle down. Internally there is a drop of high viscosity silicone oil to provide damping. The internal vibrational energy is dissipated by this internal damping in a few seconds. After that, because they are under quite a bit of tension (compared to their breaking strength) and they are no longer vibrating, the wires act like straight rigid rods.

The relevant patents are US3081552 (suspension) and US3160237 (damping):

US3081552A - Suspension devices
US3160237A - Damping device

Cheers,
Bruce

PS: here is an interesting quote from the first reference/link above:
"These elegant solutions of damping and suspension were typical of Reason’s output. Practically he realized that for an instrument to be useful, it had to survive. Talyvel was a unique instrument because it was small and very portable. He insisted that the outer case should be capable of being stood on by an average man: he knew that in practice these things happen! Also he specified that the Talyvel should be capable of sustaining the shock of being knocked from an average size table without losing its zero datum reading. These considerations, when taken with the practical operating convenience of converting angle to an electrical signal so that it could be communicated over long distances, made the instrument unique. It has been used and even built into dam walls, bridges, the Tower of Pisa, used for aligning bearings on all types of engine and so on. This is a perfect example of long term thinking."
 
Last edited:
For calibrating my Talyvel 4, I found it very useful to spend an hour making a "sine bar". For anyone who wants to calibrate a Talyvel, I suggest doing the same. You will save more time in the calibration than it cost to make the sine bar, and you can do the calibration better. Some photos are here: https://www.practicalmachinist.com/...taylor-hobson-talyvel-4-a-366508/#post3393800

I made this from a 105cm length of extruded aluminium rectangular tubing (30 x 40mm 4mm wall). I mounted a 10mm hardened ground dowel pin at one end and a ball-tip micrometer at the other end. The distance between dowel pin and ball-tip was 1000mm to within +-1mm. The Talyvel read head (or heads) are mounted on the top surface, and the entire assembly is then placed on a stable granite surface plate. In the Talyvel gradient mode display, turning the micrometer head 0.01mm corresponds to a change in gradient of 0.01mm/m (2.063 arcseconds) with an accuracy of +-0.1%.

Here is a plot showing the displayed angle versus actual angle for one of the heads, and the error (magnified five times). The spec accuracy +-(0.2 arcsec + 2% of reading) is shown as the green envelope.

attachment.php


I was impressed at how well this 30-year old instrument worked once it was properly adjusted. A very good check is to mount both heads together on the sine bar, then put the instrument in differential A-B mode, and check how closely it tracks zero over the entire +-600 arcsecond range.
 
Last edited:








 
Back
Top