What's new
What's new

Federal Mahr EGH-13 precision electronic level test

fobyellow

Plastic
Joined
Mar 9, 2021
I bid an electronic level sensor on ebay, and waited 3months to get the parcel:(:(
The item is announced "Parts only", so i disassembled it first time.

c4851eab89e8ce2b624c7d872aa1e50.jpg

there are some Potentiometer and resistors to fine-tune the "zero point" due to the inconsistency of assemble
2bd44e5e0cc4adedb1f7290c745734e.jpg

the DC-resistance of primary winding is only 14 Ohms from connector, and there are 2 resistors in series in side the case, so the actual winding resistance is only 6 Ohms, thats pretty low compare to the other LVDT probe
90c8942ff7577979528a6a3ed457c18.jpg

since there are no leakage or any sign of disassemble, so i decide to test it first. i dont have the right female connector of ITT MC14H-10-6P, so i solderd the 4 cable from inside.
ea63500911c07a0b36d0f6d2b8ae979.jpg

the result is good. i did 37Hours long period data logging, the total drift is about 0.8um/m.
12.JPG

the short period noise is even better, during the 499 seconds of logging, it shows only 0.05um/m pp-noise, i doubt that my garage(my workshop) contribute half of the total noise, if there are "quieter" places with constant temperature, it could be used to do the geological observation.
13.JPG
 

Halcohead

Stainless
Joined
Apr 10, 2005
Location
Bay Area, Ca
This is super cool, thanks for posting. Did you use your own LVDT amplifier to read the signal?

Also, that 0.8um/m drift is so smooth it looks like it could be thermal (maybe quick high-temp excursions followed by cooldown). Was this reading done on the linear rail calibration rig shown in the photo?

Either way cool work and thanks for sharing!
 

ballen

Titanium
Joined
Sep 25, 2011
Location
Garbsen, Germany
I've done similar experiments with a couple of Talyvel 4 instruments. The 37 hour drift is thermal motion.

The short-term rms noise you are seeing is very low. The Talyvel 4 has a resolution of 0.1 arcsec, which is 0.5 microns/meter.
 

John Garner

Titanium
Joined
Sep 1, 2004
Location
south SF Bay area, California
Those Federal (now Mahr / Federal) electronic levels have been my favorite levels for decades now, and I prefer the Model 432 version that uses a breadbox-sized electronics box using turn-the-knob tuning "pots" and red-LED display to the newer 832 membrane-button tuning.

The tilt-sensor unit was designed in the early 1960s, by Eugene V Grumman; the patent rights were assigned to the Bullard Company of vertical lathe fame. The US Patent number is US3286357.
 

ballen

Titanium
Joined
Sep 25, 2011
Location
Garbsen, Germany
Here is a plot where I am comparing a Talyvel 4 to a bubble vial, which I am reading out with a capacitance to digital converter. The motion is thermal drift of a tabletop in the house, and you can get some idea of the Taylvel 4 rms noise. You can also see the 'stick slip' of the bubble. Note that I am reading the Talyvel data via an RS-232 digital interface. so the raw data that I am plotting has a 0.1 arcsec (0.5 um/m) resolution.


full
 

fobyellow

Plastic
Joined
Mar 9, 2021
This is super cool, thanks for posting. Did you use your own LVDT amplifier to read the signal?

Also, that 0.8um/m drift is so smooth it looks like it could be thermal (maybe quick high-temp excursions followed by cooldown). Was this reading done on the linear rail calibration rig shown in the photo?

Either way cool work and thanks for sharing!
i connect the sensor to a ametek signal-conditioner, the long-term reading was done on the granite surface, the linear-rail calibrator doesnt work good . i'm trying to build a new one using 750mm bridge-straight
WYLER use this, looks like 1000mm long tilt generator
微信图片_20220815115941.png
i'm thinking to follow this idea, change the optimeter with new linear scale
微信图片_20220815115930.jpg

use this 750mm straight
微信图片_20220815115902.jpg
 

fobyellow

Plastic
Joined
Mar 9, 2021
Those Federal (now Mahr / Federal) electronic levels have been my favorite levels for decades now, and I prefer the Model 432 version that uses a breadbox-sized electronics box using turn-the-knob tuning "pots" and red-LED display to the newer 832 membrane-button tuning.

The tilt-sensor unit was designed in the early 1960s, by Eugene V Grumman; the patent rights were assigned to the Bullard Company of vertical lathe fame. The US Patent number is US3286357.
thank you very much, this help a lot to understand how it works
 

fobyellow

Plastic
Joined
Mar 9, 2021
I've done similar experiments with a couple of Talyvel 4 instruments. The 37 hour drift is thermal motion.

The short-term rms noise you are seeing is very low. The Talyvel 4 has a resolution of 0.1 arcsec, which is 0.5 microns/meter.
do you have the inside pics of taly4 sensor? :drool5:
 

ballen

Titanium
Joined
Sep 25, 2011
Location
Garbsen, Germany
do you have the inside pics of taly4 sensor? :drool5:
Yes, and they are here on PM. Just use the search function at the top, with the term "Talyvel".

You can make a nice precision angle generator for calibration and testing in an hour. Use a piece of aluminium box extrusion, say 30 x 40mm (a bit more than 1m long) for the body. At one end, attach (mill slot + epoxy) an 40mm x 8mm or 40mm x 10mm ground dowel pin as the hinge. Then exactly 1m from the center line of this, attach a ball-end micrometer head (electronic with a 1 micron resolution is nice, since you can zero it). You'll find photos in some of my threads under "Talyvel", see links below.

Links:
 
Last edited:

fobyellow

Plastic
Joined
Mar 9, 2021
Yes, and they are here on PM. Just use the search function at the top, with the term "Talyvel".

You can make a nice precision angle generator for calibration and testing in an hour. Use a piece of aluminium box extrusion, say 30 x 40mm (a bit more than 1m long) for the body. At one end, attach (mill slot + epoxy) an 40mm x 8mm or 40mm x 10mm ground dowel pin as the hinge. Then exactly 1m from the center line of this, attach a ball-end micrometer head (electronic with a 1 micron resolution is nice, since you can zero it). You'll find photos in some of my threads under "Talyvel", see links below.

Links:
thanks for reply, i searched talyvel and found several topics with sensor disasemmbly, but all of them are taly1 or taly2, the difference between taly1 and taly2 is the mechanical zero point adjusting parts. the talyvel sensor use 3 sets of parallel suspending wire. i think it's difficult to adjust. so i want wo know how they optimise it in the newer version.

i noticed you have taly4 sensors, so i want to see inside.

i already built an angle generator, but i found the magnification factor varies while calibrating sensors, so i want to built a motorized angle generator , use to calibrate the nonlinearity(i want to change the parallelogram pendulum to simple pendulum, so the movement of the mass would be circular, the disadvantage is nonlinear, but the whole mechanic is much easier to build).
 

ballen

Titanium
Joined
Sep 25, 2011
Location
Garbsen, Germany
The heads for Talyvel 1,2,3,4,5 and 6 are all identical inside. Yes, the Talyvel 1 and 2 have the micrometer adjusters, but the head itself, hanging underneath, is identical. (Well, they changed some BA thread screws to standard metric threads.)
 








 
Top