Logo
Builders' Plan Gallery  |  Hip Pocket Web Site  |  Contact Forum Admin  |  Contact Global Moderator
May 29, 2017, 04:47:26 PM *
Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with email, password and session length
 
Home Help Search Login Register
Pages: 1 [2]   Go Down
Print
Author Topic: Counter for rubber winder  (Read 2186 times)
0 Members and 1 Guest are viewing this topic.
msc
Bronze Member
***

Kudos: 1
Offline Offline

United States United States

Posts: 20



Ignore
« Reply #25 on: February 20, 2017, 08:18:49 PM »

Thanks for sharing the schematic. It makes it much easier to see how you handled the details. There are a few things that did not come through clear enough to read so I have a couple questions.

What battery type did you use?
Is your VCC voltage 3.3V or 5V?
What regulator chip are you using?
What RTC are you using?

My interest is mostly technical curiosity. A large part of my day job is the design and fabrication of special purpose data recording and control equipment for various research projects. I do a lot of hardware design but have only dabbled a bit with programming.

Mike
Logged
msc
Bronze Member
***

Kudos: 1
Offline Offline

United States United States

Posts: 20



Ignore
« Reply #26 on: February 20, 2017, 08:22:16 PM »

My plan is to use 3 hall sensors for positive tracking of the direction of each rotation. This of course allows automatic counting of the peek turns and the windback turns. If more then a couple minutes go by and then more turns are backed off they are assumed to be the count of remaining turns at the end of a flight.

For any models where the rubber is wound off of the model the torque gauges based on an electronic scale are difficult to improve on. To display or log the torque on the counter display the easy solution is to make a custom electronic module that connects to the same load cell but in addition to displaying at the scale would broadcast to the counter over a bluetooth connection. When doing custom electronics a couple other nice features are easy to add. The auto shutoff time can be set as long as desired and it can be easy to select between In/oz and g/cm for scaling.

When the motor is wound on/in the model it becomes trickier to monitor/log torque but it can be done. It would require a sensor module installed on the output shaft of the winder much like the twist wire torque gauges. It's a much greater challenge to design but should fit on a circuit board about 2 inches in diameter. With a rotating sensor the beam type load cell from a cheap scale would no longer work so a true torque load cell would need to be custom fabricated. Powering and communicating with the sensor is the significant challenge. Slip rings are possible but would have to be very precise to have acceptable drag and be reliable. Battery powered and wireless data link is more practical but who wants another battery that needs to be maintained. Inductive power coupling with Bluetooth data is probably the best approach but adds another system to develop.

It's just a design concept at this point that may or may not ever happen. I am using it as a motivation to do some tinkering with programing micro controllers. I have purchased some sensors to play with and have been re-familiarizing myself with some programming tools.

I'll be starting out with Arduino because I have worked with them in the past and I can get the environment to run on my vintage Mac. There are lots of compatible building block modules with extensive support libraries so it should be straight forward to experiment with bluetooth and SD cards. The initial work will be mostly with available modules tied together with breadboard construction. After getting code running and hardware features finalized circuit boards would be laid out for final packaging. At that point it may be best to ditch the Arduino environment but I'll have to weigh any expected benefits against another learning curve.

If I get a decent start on development I'll start a thread to give updates on progress and seek feedback on features.

Mike
Logged
rgroener
Platinum Member
******

Kudos: 23
Offline Offline

Switzerland Switzerland

Posts: 1,100

Topic starter


Ignore
« Reply #27 on: February 21, 2017, 01:43:42 AM »

Mike, good news that you are on your way.

To your questions.

I use a single cell (1.2V Accu or 1.5V Batty) as power source. The TPS61220 will boost it up to 3.3V. So VCC is 3.3V since the display needs 3.3V, no additional voltage converters are needed.
- TPS boost converter TPS61220 http://www.ti.com/lit/ds/symlink/tps61222.pdf
- The RTC is a MCP7941 http://ww1.microchip.com/downloads/en/DeviceDoc/20002266F.pdf
- Hall sensor is a SS443A http://www.mouser.com/ds/2/187/honeywell-sensing-ss400%20series-product-sheet-00905-947400.pdf

About the detecting of the direction... why do you want to use 3 hall sensors? Wouldn't two be enough?

Good idea about measuring the torque. I never thought on using an electronic scale.
The problem with my ideas on measuring the torque was always the turning part of the winder which would turn the sensor and therefore also the cabling.
Wile reading your post, I maybe have a solution for my problem. I have to think about it and will write more as soon as I see that it could work.
I am still looking forward to your solution.

Arduino is a good solution especially if you used it before. I never used them, but the Arduino Uno board has the same processor I am using. Its also an Atmega328p.
Therefore it would be easy to use it in your follow up project without the arduino hardware.

If you like to skip the Arduino step, I could offer you one of my prints. I would also share my code, including the library for the display. It's just an offer, you can decide what's easier for you.
The programming environment is also available for Mac. An apprentice of mine uses it on his privat mac.
Look here for some install instructions http://www.ladyada.net/learn/avr/setup-mac.html

From my side, you dont have to open a new thread. If you want so, it's fine too. Feel free whatever you want Grin

Roman







Logged
Tapio Linkosalo
Platinum Member
******

Kudos: 26
Offline Offline

Finland Finland

Posts: 1,056



Ignore
« Reply #28 on: February 21, 2017, 01:55:37 AM »


I've been considering the means of measuring torque while winding, and have had rather similar thoughts as Mike. However I have another issue for which I have yet to find any working solution- I would like to measure also the stretch of the rubber motor while winding, but have not figured out any working and yet accurate enough system to find out the distance from winder to torque meter. Laser/light would be accurate, but would need extremely fast time measurement. Ultrasound might work, but would probably need some active system, as the passive sensors are not accurate enough. Mechanical systems - prone to tangle!


-Tapio-
Logged
rgroener
Platinum Member
******

Kudos: 23
Offline Offline

Switzerland Switzerland

Posts: 1,100

Topic starter


Ignore
« Reply #29 on: February 22, 2017, 12:48:46 AM »

Tapio, what do you mean by an active ultrasound system. Ultrasound is always acitve, isn't it? You need a transmitter and a receiver. I would start with something like this.
http://www.micropik.com/PDF/HCSR04.pdf
Maybe you would have to add a plate at your winding stooge to make sure that the ultrasound impuls is reflected nicely.

Roman
Logged
Tapio Linkosalo
Platinum Member
******

Kudos: 26
Offline Offline

Finland Finland

Posts: 1,056



Ignore
« Reply #30 on: February 22, 2017, 01:15:22 AM »


By "active" system I mean a setup  where instead of a passive reflector plate I'd use an active "repeater" to give a strong precise signal from the other (passive) end. In simplicity, this could just a device that is listening to a certain frequency and starts emitting sound when it is receiving sound. Tuning down the receiving sensitivity of the measuring end so that it does not "hear" the reflections, only the actively sending repeater, I could exactly pinpoint the reflection coming from the point I intend. And therefore would get much better accuracy of the distance, just would have to account for difference in sound speed due to temperature and pressure variations. But, this would mean hacking into the systems and writing some specific code for them...

Logged
Andrew Darby
Platinum Member
******

Kudos: 39
Offline Offline

United Kingdom United Kingdom

Posts: 1,698




Ignore
« Reply #31 on: February 22, 2017, 02:26:42 AM »

Since "every action has an equal and opposite reaction" I don't think that you don't need to measure the torque at the winder, simply measure it at the stooge.  So long as the model is not touching anywhere else and the axis of the rubber is at right angles to the measuring "arm" then the torque in the rubber would be the reaction at the rear peg....

Andrew
Logged

Hogwash I tell thee!
Tapio Linkosalo
Platinum Member
******

Kudos: 26
Offline Offline

Finland Finland

Posts: 1,056



Ignore
« Reply #32 on: February 22, 2017, 03:38:10 AM »

Applies for indoor models, or if you are winding the motor in modern fashion outside the model with a half-tube. But not for outdoor winding inside the model, where the wing is impacting the model.

Logged
msc
Bronze Member
***

Kudos: 1
Offline Offline

United States United States

Posts: 20



Ignore
« Reply #33 on: February 23, 2017, 08:47:07 PM »

I believe Art356A is the one who deserves credit for developing torque meters based on the cheap electronic scales. The FliteTork  meter is based on this design concept.

The reason for 3 hall sensors is to eliminate false counts when cranking has stopped but the output shaft may be dithering back and forth a bit and may induce false counts. Two sensors can only positively track all rotations if there is a multi lobe signal and the sensors are positioned to create a quadrature signal. At anything approaching a steady crank speed with two sensors placed close together (and one magnet) you can look at the timing relationship between the two sensors to detect direction but that can fall apart at irregular speeds near zero. I expect you can do good enough for most people with just two sensors but I get fussy about such things.

For a Hall sensor I'm looking at the Diodes Incorporated AH183. It's a better operating voltage match and much lower power consumption then the Honeywell part. It's a slower responding sensor so I need to do some testing to make sure it's fast enough. I will mount a magnet on a 20:1 winder and see what the signal looks like at full speed. I'm curious so I'll try to get to that before long.

Thanks for the link to the AVR-GCC install instructions, it looks promising. I'll still be starting out with Arduino for a quicker start. I expect many of the code blocks are reasonably transportable between Aurduino and the AVR-GCC. If your code is reasonably well commented in English I would love to see it.

I'll end up with a different combination of features and will be laying out my own board design at some point. Thanks for the offer to share the board layout but I don't expect to make use of it.

I would like to measure also the stretch of the rubber motor while winding, but have not figured out any working and yet accurate enough system to find out the distance from winder to torque meter.

What sort of accuracy and resolution are you hopping for on a distance senor? Is this a measurement you think would be useful for winding procedure or is it more of a research tool to help better understand the dynamics of the rubber?

The most practical sensor would be ultrasonic. If you can't get good enough with that I don't think there is a cheap and easy solution. I'm sure there are laser sensors that would give very accurate measurements but the cost of that sort of industrial sensor is not cheap. I would be very impressed if a suitable laser sensor could be sourced for under a grand and I would expect more like $3000+. A string pot could give a good measurement but is probably what you were thinking about potentially getting tangled with the motor. If there is a cleaver low cost hobby solution (other then ultrasonic) I bet it lies in the field of image processing.  A modern smart phone has plenty of imaging and processing power for the task. There might already be an app for that.

Mike
Logged
Tapio Linkosalo
Platinum Member
******

Kudos: 26
Offline Offline

Finland Finland

Posts: 1,056



Ignore
« Reply #34 on: February 24, 2017, 12:56:55 AM »

The reason for 3 hall sensors is to eliminate false counts when cranking has stopped but the output shaft may be dithering back and forth a bit and may induce false counts.

You do not need 3 sensors, just some clever programming. See message #16.


I would like to measure also the stretch of the rubber motor while winding, but have not figured out any working and yet accurate enough system to find out the distance from winder to torque meter.

What sort of accuracy and resolution are you hopping for on a distance senor? Is this a measurement you think would be useful for winding procedure or is it more of a research tool to help better understand the dynamics of the rubber?[/quote]

I think 1cm or better 1/2cm would be good. A distance sensor could be doubled with a force sensor measuring the pull force, so the setup could be used for Pearce-type motor testing. But also I think that with the measures of pull force and distance you could get better understanding of the motor strain, and this could help your motor winding.

Logged
msc
Bronze Member
***

Kudos: 1
Offline Offline

United States United States

Posts: 20



Ignore
« Reply #35 on: February 24, 2017, 09:03:55 PM »

I have a nagging feeling just two sensors can be tricked for a tick or two under the right jittering conditions like reversing exactly on a pulse. It's going to be a year or more of hobby tinkering before I need to settle on the number of sensors. An extra sensor ads $1.00 to $1.50 in parts cost so it's not a big factor.

I had been wondering if tension would be of interest because it's easy to add to my design concept. With minimal mods from what I was already considering it could be turned into a full function rubber motor test stand. Here is how I would go about it.

It would be a simple mod to add a tension load cell to the common digital torque meters. Simply use another of the beam load cells from a digital scale to support a thrust bearing. Make a custom circuit board that reads both load cells drives a large display, doesn't go to sleep before you want it too, gives choice of display units, and broadcast the data to the data logger at the winder.

For a test stand you would want to build a track or rail to support the winder in proper alignment with the load cell unit. Once you have the winder moving on a rail an optical sensor can count lines printed on the rail. With alternating  4mm black and white lines it would be possible to offset 2 photo detectors by 2mm and get a 2mm resolution using quadrature signals for positive counting up and down as the winders dolly is moved. The easy approach to this measurement might be to use the guts of an optical mouse to track the motion.

It would be possible to log torque, tension, length, and turn count at regular intervals. You could then monitor torque as you hold for the amount of time it takes to transfer the motor to the model. After the pause you could display turn rate for feedback as you slowly unwind to simulate flight time while logging both turns and torque.

Mike
Logged
msc
Bronze Member
***

Kudos: 1
Offline Offline

United States United States

Posts: 20



Ignore
« Reply #36 on: March 11, 2017, 09:35:31 AM »

I had been thinking of the sensors only as an independent indication of the magnet passing and as such I'm still fairly certain it takes 3 sensors to avoid the possibility of an occasional false tick or two.

I have come to realize it's possible to position two sensors so the output pulses overlap by about half the pulse width and monitor them as a quadrature signal. This would indeed provide perfect detection of each turn with only two sensors.
 
I did a basic test of the AH183 sensor and it looks like it will work well. I'll have to wire up a pair of them and start learning my way round an encoder library.

I did some reading up on rubber testing to see what others have done and better understand what might be useful.

A couple more thoughts on the length measurement. On a testing fixture the optical encoder reading a pattern on the track would be my top choice.

When winding for flight where you may not want to have the winder restrained or drag the track to the flying site there are a couple options. The transponder idea to improve ultrasonic performance might be the best approach. Rather then sending an echo back, RF or IR could be used to complete the round trip so you only have acoustic errors in one direction. Since I am already planning a torque and/or tension sensor with bluetooth at the far end of the rubber it's simple to add sonar ranging. I'm almost expecting bluetooth to have too much variation in latency to be useful in ping timing but it's worth testing. If the latency is stable to within a few micro seconds the bluetooth would work for echo return.

One thought is to synchronize the clocks of both units. Once the clocks are synchronized the scale unit can send a ping exactly on each second (or some fraction of a second). The winder unit could then tell distance by timing how far past the second a ping is received.

One way of calibrating for both time synchronization and density altitude would be to calibrate the winder with some test pings. Hook the winder hook directly to the torque meter and tell the winder to learn this timing past the second is zero distance. Then install a reference length of cable between the hooks and tell the winder to learn this timing past the second is the reference distance. The clock crystals will not be perfectly matched so there will be some distance drift over time with this approach. If the drift is too great over the course of a flying session a time drift factor could be calibrated in to keep the drift to an acceptable level. A one time drift calibration could be implemented by setting the winder near the scale (pinger) and have the winder listen for a couple hours to learn the exact pace of pings and store a calibration value so it can adjust for drift over time.

Mike
Logged
cman
Bronze Member
***

Kudos: 6
Offline Offline

United States United States

Posts: 67



Ignore
« Reply #37 on: March 12, 2017, 04:27:33 PM »

Hey Raggedflyer, do you have a picture of how your unit actually counts?
The reason I ask is a few months ago I bought a couple of those same counters but I'm not smart enough to figure out how to make it work!

Thanks,
Chris
Logged
raggedflyer
Bronze Member
***

Kudos: 1
Offline Offline

United Kingdom United Kingdom

Posts: 35



Ignore
« Reply #38 on: March 13, 2017, 03:32:38 PM »

Chris,

Give me a few days and I'll add some more information.
Logged
cman
Bronze Member
***

Kudos: 6
Offline Offline

United States United States

Posts: 67



Ignore
« Reply #39 on: March 13, 2017, 05:36:34 PM »

no rush, we're expecting 12" of snow in the next day or two so there won't be any winding going on.

Chris
Logged
Pages: 1 [2]   Go Up
Print
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines Valid XHTML 1.0! Valid CSS!