Looking for suggestions building an encoded unpowered cart

Community General Discussions
1

The cart would be lowered down the side of a wall and measure where it is as it goes. Resolution would be 0.00306 inches/pulse.

The four wheels are staggered so that no two wheels hit a seam in the wall at the same point.

There would be a cable to my datalogger from the cart.

Two questions.

  1. Can it be done with Lynxmotion parts?

  2. What would be a good four wheel robot kit to buy to scavenge parts from? (I don’t want to waste a bunch of time shipping if I don’t get enough parts the first time.)

To those who want to know more: I am attempting to replace an $19K bot with a unpowered cart. It will have an ultrasonic transducer to measure the wall thickness. Perhaps some day motors will be added, but in the years I have been doing this they have never been needed. Sometimes though, it would have been nice so that I would not have to climb in a windstorm. There is something about never needing to actually use the climbing helmet… You guessed right, I am repelling down the wall and holding the transducer. One of my clients complained. Perhaps he got scared? So, now to make him happy I’m trying this.

Thanks in advance for the comments.

2

If you are wanting .003" position resolution from a rolling cart, then I think this is probably not possible with normal hobby parts.

3

The encoder on the hand held cart is gear driven - off of two of the wheels which are on a common axel. Is the gear what I need or the encoder? I would guess the ratio is about 12:1.

The wheels are in line and everytime they hit a weld it pops off unless I’m holding it.

The logger software is very flexible and I can change the encoder settings but I want to have at least ~10/inch (0.100"), the more the better.

4

I would tend to agree with zoomkat on this however I am curious… why .00306 inches?

Also when you say lowered down a wall… what is doing the lowering, a winch of some sort? Would that be on the “cart” or at the top of said wall?

If you are lowering a load under constant tension (gravity) on a winch then it may be possible to index the cable over a pulley with a high resolution shaft encoder.

How do you plan to control the pressure and coplanarity of the transducer to the surface being examined? This seems like it would be easy being held by hand but remote and mechanically where there might not be much down force between the cart and the wall being examined that might be more of a trick.

5

I suggest you put the gizmo on a cable/rope and hand winch it up and down the wall from a small hoist setup at the top of the wall edge. For verticle position measurement use a fiberglass measuring tape like below attached to the gizmo. Benchmark the tape reading and the gizmo position at the top of the wall and go from there. Use a video camera and portable DVR (or just a digital camcorder) to record the tape reading at each test point (you can also narrate what you are doing to the camcorder for later data consolidation).

harborfreight.com/cpi/ctaf/d … mber=35712

6

Thanks everyone for the thoughts! Here are some more details:

There is (are) 200 lbs of magnetic force holding the cart to the tank.

There will be 20 measurements (or that is what I have been using) per inch and 30 feet of tank to keep track of so a tape measure/camera… would cost more time than the project would permit. However, the logger is quite able to record the measurements and the location together.

Lazyness is the reason for 0.00306 inches as it is the default of the logger.

Pullys are nice in theory but these tanks are dirty and the wind and rain make everything unstable. Including me at times. 8-0 :open_mouth:

Ah, coplanarity check out the Silverwing website they have “rubber” tired transducers. Very cool and clean too. Between the 200 lb pull and a spring we should be good! Actually the RScan unit I am using now works well as long as I am pressing it into the tank but that tends to twist me on the ropes so I end up dragging one knee down the side leaving a shiny stripe on the chalked paint. (And paint on my coveralls)

Imagine this:
One hand on the cart
One on the datalogger
Watching the cart
Watching my ropes
Watching the welds
Watching the ground
Watching the pipes
Watching the conduits
One finger on the Pizel decender
One toe on the tank
One knee on the tank

Guard: “Hey, Mister what are you doing?”

Me: “Ask me in ten minutes when I’m on the ground.”

My preference would be to climb to the top, secure my safety line, walk to the edge and lower the cart/transducer down while sitting on the edge or lying on my chest and watching the logger’s screen. Typically for a small tank it is done 8 times.

Richer companies use remote controlled crawlers and stay on the ground. However they are $19K and tend to fall off from time to time. Then the hard hat is all you have to protect you from three pounds of metal and plastic falling.

(Back to the tape measure camera . . . I once, no twice used a laser rangefinder, tape recorder and walkie talkies to record one mile of data from the shore, while a small boat carried the datalogger out to a pier while measuring potentials on a metal tunnel. That took hours to convert into location/potential pairs!)

7

“My preference would be to climb to the top, secure my safety line, walk to the edge and lower the cart/transducer down while sitting on the edge or lying on my chest and watching the logger’s screen”

Just add a 5th encoder wheel to the front of the cart/transducer you have now. When accurate distance testing is done with cars, you see what looks like a bicycle wheel attached to the rear that is making precision distance measurements. Add something similar to your current setup. Should be cheap and fairly easy to do, and you can get actual experience on how well an encoder setup will work.

8

The cart I have now will not cross a weld because two wheels hit at the same time. It has an encoder which works off of two of the wheels. Perhaps I need to mention that I must build another cart as I cannot afford to buy the better cart (remote controlled) for $19K.

I need to know if Linxmotion parts will work to build the new cart.

9

Well, since you have not offered any pictures of your current cart (is it 3"x3" or 3’x3’, who knows?) or a design drawing of what you want, what you expect to build is still some what of a mystery. It is not uncommon for people to come to this forum expecting the forum members to design them a robotic device for less than $1k to replace commercial versions that may cost between $5k and $50k. You need to understand that the Lynxmotion products are generally hobby/educational grade and not usually suited for serious commercial applications. For the $$$ you expect to save you need to do your homework by reading and understanding the specifications of the products you intend to use. You need to be very specific as to the issue you are trying to solve. Off setting the wheels on your cart may or may not solve the apparent issue you have with rolling over surface welds. You may be better off to make a new frame for your current cart that will have the current wheels off set. Problem solved! $19k saved! Just asking a general “will your parts work” is a little simplistic and may indicate that your constructing your own cart might be a challange.

10

I agree completely.
My plan is to design based on the parts not the other way around. Sort of the Erector set way. What I need to know is if the parts have enough rigidity and mounting points to build something that will hold up to the rigors of daily use. You folks have used the parts and may well have a better idea than I.
I would love to post a design and some pictures but the problem is that I am not yet permitted to post links. However, you can see the cart in use now at the Silverwing site. Now how to avoid a link: there are three w’s a dot, the name and uk, a dot com, then forward slash utsystems dot html look at the RScan Lite.
The new one would imitate the Scorpion above it but with the wheel layout of the www dot ndtint dot com slash crawl dot htm site.
The ultimate upgrade (besides motors) would be if the encoder could be controlled by a tiny plastic coated stainless steel cable. The cable could be attached to the structure and wheel slippage would be eliminated. Hopefully, if the cable is thin enough and has enough tension on it wind will not be a serious factor.

What do you think?

11

Below seems to be the referenced sites. Where can one find a picture of your current gizmo?

silverwinguk.com/utsystems.html
ndtint.com/crawl.htm

12

Silverwing R-scan Lite (in the middle of the page) is the one used now. It fits in the palm of the hand. The magnetic wheels are only able to pull it against the tank if it is not on paint or a weld. The cable is so thick and stiff that it will twist the cart off of the tank as well.

The NDT International unit has four 50 lb pull magnetic wheels. They should do the job of holding the cart to the tank. Now to locate wheels like they use… kjmagnetics dot com slash proddetail dot asp?prod=RX4C2 is almost 60 pound pull. Wow! They have stronger too.

Potential Parts list:

frame - able to withstand 200 lbs of stress when putting on and pulling off of the tank

axels - 50 lbs force each

springs - to allow movement of the frame over welds

encoder - to count wheel turns or cable length

gears - to increase the resolution of the encoder

spacers - to allow things to fit together

bolts - to keep them together

nuts - for the bolts :slight_smile:

and lots more which you know far better than I

(transducer, cables, magnetic wheels, terminals, and so forth by others)

Well, with all that said, will Lynxmotion parts be suitable for building such a device?

13

To be on the safe side I’d say NO. All you have done is presented a general idea and a list of general hardware items. You need to look at specific lynxmotion parts and design yourself a cart that you think will meet your needs. Once you get a design on paper and present it to the forum, then you may get specific answers. These vague design questions often appear and the final result is sometimes not so good. What happens is somebody says “yes” to the vague request, then the requester wants the person to explain in detail how to use the components and how to construct the gizmo. The requester in effect tries to have the forum members design and then direct the construction of the gizmo. You need to take pictures of your device and post them (I couldn’t make much from the posted link). The gizmo can be built, but my ideas are probably significantly different than yours.

14

Thanks for all of the comments and questions. Reading this has clarified many of the design factors. Apparently there really is no way to determine if the parts are strong enough or have a sufficient number of mounting positions without actually building something with them.

Since no one has built anything similar to a magnetic wheeled cart they really cannot give an opinion as to the sufficiency of the parts to do what is needed here.

As I will not be checking this forum, if you want to contact me please email to: emc-2 (at) verizon dot net

Thanks again everybody!

SparKs