Mettler H35 with electronic readout  Latest change 2014-10-21.
Deze pagina is in aanbouw, meer info en foto's volgen. This page is under construction.
Reactions to .... or .... You have to type this, no cut-and-paste

This project is on hold for reasons of priority.

In brief: A Mettler H35 laboratory scale was turned into an electronic scale.
This site shows some pictures of the original instrument. You may find other pictures elsewhere. I did not start making pictures until I had the unit stripped down almost completely.
Required skills: Insight in how to modify / disassemble / remount complex mechanical equipment, Electronics, Software, some Metal working things.

Project status:
Halted for reasons of time / priority.
Expected progress: november 2014 or later.

For months I was considering constructing an electronic scale with a resolution of at least 0.1 gram, in particular for my anodizing activities (Dutch only).
Recently I found a Mettler H35 in a recycle shop for 50 Euros, and that gave me most of the delicate mechanical parts for such a machine.
The original H35 is a very complicated machine which requires a lot of knob-turning before you have a reading. Moreover this machine relays on a lot of adjustement screws, a.o. of mirrors and springs to reach its accuracy. I did not have acces to the calibration procedures -which probably involve special tools I do not have too-, so I decided to stick as close as possible to my original plan, involving a loudspeaker voice coil to counteract, a diode laser, some mirrors and a position sensitive light sensor to measure the angle of the balancing arm. All glued together with some electronic circuitry, a display and a microprocessor.
Below the basic idea.


The loudspeaker voicecoil acts as the counterweight to balance the load. The laser and the position-sensitive light detector provide information about the angle of the balancing arm. The micro controller controls the current through the voicecoil until balance is reached.
As balance is always in the same position of the arm the non linearity of the voicecoil/magnet system is eliminated.

Remind: For this high resolution we need bearings with absolutely no stickyness or hysteresis, and that is one of the most demanding things.

Strip the machine from all parts we do not need, and that is most of the mechanics, the optical train and all knobs, besides the big lever mid-low-front.
Note: some disassembling actions require a particular order. You have to find out yourself. If you do not feel comfortable with such things you probably should not start a project like this.
The pictures below show what is left over. That is of course the case, the balancing arm, the parts which hold the scale and the mechanism to block the balancing arm.
In the mean time I had the opportunity to clean the parts (The machine I bought apparantly had a history of being in a damp place, there was some corrosion, and it had an female inhabitant, a spider found her way out while I was dissassembling the unit) And I also repainted it in my favorite color metal-blue.


Then I had some second thougts about the loudspeaker as a counter-actor.  Waarom?

For a good accuracy it is important that the swivel points of the balancing arm,  load, and counter-load are on one line in the vertical plane. If you do not do this the balancing arm is either sub-balanced (it has a preference position) or instable (it will tumble over to either position.


Some thoughts about a precision balanced system.

Fig 1.
Here we have a theorethical massless beam, where the points of application are on one straight line. Such a system is stable, and has no preference, the beam can be under any angle w.r.t. the horizontal.
This is the design goal for my electronic balance.

Fig 2.
Here the points of application of the load and the counterload are below the application point of the central bearing. Such a system is stable, but has a preference position in which it will eventually come to a stand-stil.
Note that the original M35 used this preference for the sub-gram read-out. The graduation scale on the far end of the beam is optically projected on the window in the front. The calibration of this scale depends directly on the offset between the application points, and of the height of a counterweight, fixed to the beam.

Fig 3.
Here the application point of the central bearing is below the others. Such a system will be unstable. Even if perfectly balanced it will eventually tumble over to one side or the other. This is an unwanted situation.


Fig 4.
In practise we do not have a weightless beam, and in the case of the Mettler the mass of the beam is above the points of application. This has the same effect as if we had the situation of fig 3.
So what we have to do is creating a certain amount of the Fig 2 situation such that it compensates for the weight of the beam. There are two ways to do that, first the point of application of the counterload must be adjusted somewhat below the other two, and second, the Mettler has an adjustable top-weight.
With these two we try to achieve the situation of fig 1 as close as possible.
Start with the top-weight as low as possible. Adjust the point of application of the counterweight such that we just have the preference situation of fig 2. Thne adjust the top-weight in the up-direction until the "no preference" situation of fig 1 is reached. This should be done without using any of the electronic features, and on the location where the unit is to be used, because our counter weight is a magnet an will be influenced by the earth's magnetic field.