Interesting discussion. I still think that the easiest solution would be to instrument the handle; the differences in tension between the up and down lines would provide an indication of the torque being applied to the bellcrank and with a little math an idea of the load applied to the pushrod. The other thing to keep in mind is that we are measuring small loads and the portion of the control system to be instrumented will have to be designed accordingly (i.e. be flexible enough to give the gauges something to measure). If instrumenting the handle the load measuring clips would have to be relatively thin or maybe be U-shaped. I do not think that line stretch would affect the results since in the end, the input is manual which itself carries a lot of variation.
Many years ago I read an article on an old magazine where they were trying to measure the loads needed to move a control surface; this was in the early days of radio. So they ended up mounting a section of a wing on a car or pickup, deflecting the surface a known amount (in degrees) and then driving at different speeds to see how much force was needed to keep the surface deflected. Compared to a R/C model our planes fly in a relatively narrow speed band so the testing would be a little simpler.
You could do the same today using a servo to deflect the surface and measuring the current draw to estimate the load needed to deflect the surface a certain amount. You would have to do a little testing ahead of time so that you can correlate the current draw to a given load but it is doable without being overly complicated (as in recording data using a pencil and a note pad). Anything we do will be an approximation but then again, we are not trying to land a rover on Mars....
Teo
Anything worth doing is worth doing well as they say. No we aren't going to Mars but it is just as easy to make good accurate data as it is to make garbage or questionable data. I am going to make some data. The following is me sorting through the options and discussing my decision process to that end. The only relevant question is whether to instrument the push rod or make a torque meter or a combination of both. The data collection package is the same regardless of where it is located, handle or airplane.
Well, do you know what the contribution of the drag is on the wires well enough to be able to accurately calculate the small amount of tension required to move the controls to overcome the hinge moment out of the combined influences of flying in the elements? Yeah, me neither. My educated guess, having done that calculation for solid wires and realizing that I haven't a good idea of what the drag coefficients are for stranded wires, I'm gonna guess my level of confidence in that calculation, is plus or minus 25% at best. I could refine that by carefully reviewing the fuse cam videos, carefully measuring the line sweep from a less than optimal perspective and reduce that error by 10% ish. And I'd still be thinking about how to determine the contribution portion of those numbers from the elevator and from the flaps. I've just illustrated that there are three unknowns in this measurement method which require assumptions and calculations to estimate the levels from an alternate location when we could measure them directly with the same level of effort and not have to make any corrections.
I cannot talk to the other efforts but this is the kind of thing that requires some basic knowledge of instrumentation. I'm by no means the expert but I have spent a significant amount of my engineering years instrumenting engines and airplanes for flight test. This is the kind of thing that I do, make data. Good data, like Howards, is a tough one to get. In the years gone by they didn't have the same level of miniaturized ICs we have today. We have a computer hundreds of times more powerful than the best of the computers the space shuttle flew with in our cell phones.
The instrumentation difficulty is the same on either end of the control lines with the exception that, on the airplane side, you have to be able to provision access to a place to mount the equipment. On both ends you need strain gages or load cells (an off the shelf strain gage in a package), an amplifier, a data logger and a battery. If you can't find a load cell that will "clip in" you'll have to make one or two. 10, 20 or 30 years ago they didn't have this little beasty:
https://www.adafruit.com/product/2796 , a microcomputer with data logging capability. If the ESC could work on 3.3V logic this thing would be the only processor you need on the airplane to do everything you'd like to do. It could connect to an IMU, a servo pot, a load cell amplifier, an ESC and the data could be extracted as easily as pulling the SD card out and plugging it in to your computer.
It is a fairly straight forward task to get the small signal out of the differential tension across either side handle to bellcrank. In both instances a load cell or place to mount strain gages is necessary. I searched for an hour or so and could not find anything suitable for this purpose which, a) easily fit between the handle and lines or b) cost less than $350. So, creating something is required. it's not too hard to do and calibration isn't difficult either. So, purchasing and using strain gages is necessary. These range all over the place in price and the price is function of sensitivity. But I can purchase strain gages sufficient for the job $37/5 enough to instrument a bellcrank or a handle or a pair of pushrods.
Picking off the differential tension between the lines on either the handle or the bellcrank is straight forward. Each link would be configured in a half bridge and each would receive the excitation voltage across the pair. Each pair is wired as a voltage divider on the link. The differential signal then would be pulled out between the two cells on each branch and compared. Simple enough and it would certainly make some interesting data.
The alternatives to that are to instrument the push rod or the control horns. Neither one is any more difficult than creating the tension links before the model is assembled or if it is a take apart airplane. Carbon fiber is very stiff and we use it primarily in the control system because it is easy to get carbon fiber tubes and make pushrods with them. We do that more because of the bulk modulus of the section which prevents the PR from bowing. Stiffness isn't the primary problem. We get a bonus of an easy to bond very stiff material when we select CF as the material for a push rod. But the requirement doesn't force the use of CF for the pushrod and aluminum will suffice for load and modulus wise. Bonding the aluminum reliably is a bit more difficult than CF by a factor of about two requiring adhesives such as 3M DP460. What this means is, that if a good strain signal can't be gotten because the loads are so low, a material change in the PR could be substituted with less stiffness and the impact to the airplane would be minimal. I have 7075 arrow shafts, I used to make tail booms with from the ones I damaged shooting them at stuff. The primary advantages of instrumenting the push rod is that it can be accomplished in-situ and we don't have to perform any compensation calculations. We're directly measuring the force required to create the hinge moment. Calibration wouldn't be all that hard either.
In an ideal world where we are wishing to know the hinge moment we'd measure it directly. This is the last on my list of options and I would venture a guess that the data quality difference of this method to be insignificantly small when compared to using push rods. However this is an option to consider during the construction of an airplane which is being built as a flying test bed. Such as my FLCL model in the other thread. Yes, this diatribe is me thinking out loud on a thread. I am going to instrument that airplane for hinge moment measurement. The only question is how and where. Creating a torque meter is no more difficult than making a load cell link. It involves pretty much the same elements in construction. The only real challenge is in attaching the strain gages to a current control horn. A torque tube could be made in the same way as a push rod. My only reservations with this method is the pivots and the bonds. A pivot on the tube could be accomplished by...
--- dang, I hate when this happens ... I had considered an alternate torque tube flap drive early on the in the design and decided against it. Now, I'm thinking a minor reboot as I just figured out how to do it and a good reason to proceed. ---
... The torque tube pivots would be easily done by making an insert which adapts a rod which is essentially a control horn cut in half. The torque tube would then be inserted between the two halves. This would result in the flap control link being moved outward to about the MAC of the flap This would allow for enough wind up to create a good signal. This is the most complicated method but most likely the one with the highest accuracy. I'm going to spend a little time on this in the CAD. Stay tuned.