The idea I had for these was to place them inside of a rubber or foam bumper of a tracked or wheeled robot, and act as a bumper sensor. When the bumper touches something like a door frame when the robot is turning, the flexing of the sensors would would trigger a script to stop motors, reverse, and re-track to avoid the obstical with a script.

These apparently work with Arduino, so they should work with an EZ-B.

I know it will work on an ADC port, but not sure how to hook up or what value to use for the pull-up resistor.

I think you just connect between signal and ground with a resistor on the signal side but not completely sure.

@jeramie? Are you reading this?

Alan

You won't need to use a resistor with this flex sensor as it is basically one big variable resistor... It simply connects to the analog signal pin and ground pin of the ezb... When you flex it you will see the analog value change in proportion to the amount of bend...

Not sure why you would want to use it as a crash sensor, though... This is cheaper and has a built in pull up resistor to make it plug and play to the ezb's digital ports...Crash Sensor

Thanks for the replies Alan & Richard.

I was thinking of using these to put inside of K-9's side bumpers. The front ping sensor works well for front object avoidence, but there has been a couple of times where when he takes a corner turning in to another room, as he turns he sometimes hits his side panel on the doorframe, and has cracked one side of the blue acrylic panel.

Although a good idea of using the micro switches Richard, I would have to use a row of them (maybe 5 or 6 per side), but using these flex strips I would only need one per side that could monitor the whole length of the bumper. That's the idea anyway.

I forgot to say in the last post in regards to using resistors, in the item description it says about using a "pull up" resistor, but in a link explaining how to connect it, it uses a force sensing resistor as an example and says to use a "pull down" resistor which contradict each other.

In both discriptions, it assumes using a 5v Arduino analog connection so I was thinking what Richard said about it not needing a resistor on the v4's 3.3v signal pins. You can see why I'm a little confused. I have ordered 3 of them so I guess a little testing will be called for.

SpectraSymbol has some real nifty stuff. I'm using their round Softpot on my B9 arm to send position and speed info to my Kangaroo/Sabertooth. It works great and I was a true skeptic at first. I hooked it up to the ADC port on the EZB and when I would run my finger around the pad I could watch the reading change in ARCs ADC monitor control. Very cool and they are rated to last for over a million uses.

If you want to use this sensor you should really read some tutorials and reviews on how people are using them and about any problems they are having. I'm sure Richard R has good intention with his advice but I think is a bit off. If you go to Sparkfun they always have good feedback and real use advice on the products they sell. I was looking over a tutorial there on this sensor and they recommend using a resistor and give a good explanation of why. It looks like you will need to feed the sensor 5vdc through a Pull Up resistor and tie it in between the top pin of the sensor and the ADC sensor pin of EZB. The bottom pin of the sensor goes to ground. Here's a short sample of that tutorial:

The simplest way to incorporate this sensor into your project is by using it in a voltage divider. This circuit requires one resistor. Many values from 10K to 100K will work, but we'll use a 10K resistor here (SparkFun part number COM-08374, also supplied in our Inventor's Kits and locally at Radio Shack, etc.). Connect the flex sensor to your microcontroller using the following circuit:

Here's how to hook it up to the ADC port of an Arduino useing a Pull Down risistor but it's the same on the EZB:

So go over there and read all about using this sensor. There's lots of information. Here's some links to make it easy for you. Make sure to read all the reviews. Thats where you're going to find most of the practical information on how to use this sensor and how to avoid or overcome any problems.:

www.sparkfun.com/products/8606

www.sparkfun.com/tutorials/270

Datasheet

Here's a tutorial on how to hook it up to the ADC port of an Arduino useing a Pull Down resistor but it's the same on the EZB:

Tutorial on hooking up to an Arduino

Thanks for the info and links Dave. I'll check them out.

I edited my post with more info. Have a look now.

Have fun, Dave

@Steve, You could easily use the crash sensor Richard linked to and you would only have to use one or two per side. However you would have to place them under a toggling panel with spring clips under each side to keep it level and armed. Very much like a Roomba uses with it's front panels. That way when K9 turned a corner and bumped the crash panel at any point along his side it would push in that panel and activate the switch.

That is using it with a pull-up resistor... It will work without one however (I have used them without a resistor so I can confirm that it does work sans resistor)... Remember they are basically a variable resistor all by themselves... Since you were going to use it in a go/ no go (on/off) situation I figured you weren't looking to use them linearly..... However, a 10k resistor will give it more stable (linear) readings. I am not sure if pin "float" happens on the analog ports of the ezb, but I do know it happens on the digital ports... A pull-up resistor will control any pin "float"...

That's great to know. Thanks Richard. So a little trial and error to help dial the readings in to what I need, and all should be well. I have a few 10k resistors knocking around should I need to use them for reading stability as you mentioned.

Thanks again.

@Dave depicts 2 different ways to use the flex resistor... The first on has it using a 10K resistor to pull high. The second picture with the arduino shows a 22K resistor and the flex being pulled low... You'll have to see which one works better.... One way may work better than the other... I have only used these with my old V3 so I am curious how the 3.3v ezb4 analog port will affect how they preform....

Humm, good to know these can be used without a resister.

Yes, I'm also looking forward to seeing how this works for you also. There are lots of uses for these. I've read that the readings from them are a bit hard to read if you use them for linear readings. Seems they read slowly at first and then when they get to the last 20% of the bend they really take off.

Readings from the ADC ports on the EZB do float badly. I'm using all of mine as an on/off (high/low) switch so it's not an issue with me. I've been wanting to add a pull up resistor to one just to see how well it stabilizes it.

That's cool. I'll have a play around when I receive them and post the results. Over the next week or two I'm going to strip out K-9's entire electrical system to re-wire/retrace everything, add additional sensors (hence the flex sensors), replace the damaged acrylic panel, and give him an Acer W3 tab for his display screen.

Lucky dog!

Tell me about it, lol. Well he has been a good boy so he deserves it.

A bit of an update in regards to the flex sensors. After a couple of weeks break after receiving some very bad news regarding family issues, I've got back to robots to take my mind of things. Anyway, I got home today and found that the sensors had arrived. But unfortunately the happy feeling soon went away after I hooked them up to my v4.

Using an ADC reader set at 100ms update time, I first tried wiring one up just using signal and ground pins on an analog port. With the sensor flat, the noise reading fluctuated quite a bit between 0 and 7, and bending the sensors didn't make any difference.

I then tried the way that was documented on the website using a 10k pull up resistor (as they recommended). I ran the Vcc through a 5v regulator to a digital pin, and tied in the 10k between the ground and signal wires and ran them to the ground and signal pins on an analog port. It reduced the noise (readings between 0 and 4, mostly between 0 and 2), but flexing the sensor made little to no difference whatsoever.

So unless anyone has any other ideas, I think I just blew $60 (3 sensors, shipping, import tax) on something I can't use. So I might look in to one of the ideas Richard or Dave mentioned. Kinda sucks though, not just for the waste of cash, but these would have been ideal with easy fitting to the bumpers. mad

Hmmm @Steve G you should definitely get some response on the ADC, instead of 5V I would tie the 10kohm pullup resistor to 3.3V. The ADC pins are 5V tolerant but can only read a voltage between 0-3.3V.

Reading the specs it looks like the resistance changes from 10kohm to 20kohm when flexed. So initially if you are using a 10kohm pullup you should see a value of 1.65V on the ADC port and when flexed the voltage should go up to ~2.5V.

Please check if your solder joints are good on the sensor itself or double check if the correct ADC port is selected.

@Jeremie.

Thanks for the reply. I can confirm that the solder points are clean and well attached using a solder/flux mix (which I use elsewhere with good results). And I defiantly had the correct port selected from the config menu. The only voltage it was reading was flutuating between 0 and 0.2v at the most when either flexed or straight. I even tried different ADC ports just to be sure. I can try using the 10k on 3.3, but the documentation was using an Arduino on 5v on their example so I would have expected it to work.

@Steve G, Can you try reading the output from the sensor with a Voltmeter?

Will do. I'll try it tomorrow morning and report back.

Try swapping the leads on the sensor.

@Steve G... Another idea I tried for my robot Questor was using pieces of a metal measuring tape. It seemed to work fairly well but I never got around to actually installing it before I took Questor apart. What was nice was that it makes contact no matter where it is pressed along its entire length. The spring steel and specific shape of the tape measure allows it to spring back to normal after the robot moves away from the object it came into contact with. See post # 17

A further update on these sensors.

Thanks for the advice Jeremie and Dave. I tried hooking up the sensor again using Jeremie's idea of not using the regulator, and running all the wires straight off a 3.3 analog port with the 10k. No change.

Then I tried Dave's idea of swapping the wires around which made a very slight difference as the noise values were reduced between 0 and 2, so bending the strip changed the value to around 4 to 5 briefly, but it was too unstable to use because of the noise fluctuation causing false positives.

Then tried a combination of both Jeremie and Dave's ideas, and the reading shot up to 255. So I added back the 5v regulator on a digital power pin and tried again... SUCCESS. With the sensor strip flat, the reading sat around 185/190. With about a 5 degree bend the value dropped to 170. 20 degree bend, 150... and so on. Definitely a hugh difference and very much usable now. So now to tidy up the wiring and install the strips inside of the foam bumpers.

Thanks for all the advice Alan, Richard, Dave and Jeremie. I'm very grateful for the help. And to rgordon as well for your input. Where were you when I needed you? You could have saved me a few beer tokens . Seriously though, I had a look at your thread using the measuring tape idea. That's a pretty cool idea and may try that some time on another project. Thanks for the link.

Cheers everyone.

Steve.

Outstanding! Team work! So just make sure we feed it 5v useing the risistor your useing and make sure the leads are attached properly to the strip?

If you have time please do a how to and a video.

Great @Steve G, glad you got it working!

I don't mean to be a buzz kill but a flex sensor is essentially just a variable resistor, changing the leads around won't change anything in terms of resistance, remember there are only two leads and resistors aren't polarity sensitive. I can almost guarantee that the flex sensor can work on any voltage regardless if it's 3.3V, 5V, or otherwise.

My theory is that since the flex sensor has crimp connections onto the resistive substrate the plastic may have melted a little during soldering process and provided intermittent contact. After a few more tries it may have made contact again.

If the sensor starts reading really low again, it's likely the this is the case.

@Dave.

No problem. I'll try to knock something up. Great teamwork. High 5's all round.

@Jeremie.

Truth be told, I didn't think the polarity would make a difference either. I was careful when soldering and can't see any visible damage/melted plastic as I had already tinned the wires so the soldering was very quick (there was a warning on the website saying to take care when soldering on to these sensors). I've got another one to do for the other bumper so I'll have another play around with that one and post back. BTW, you're not being a buzz kill dude. Any info you supply is always good and always appreciated.

@Jeremie.

Maybe you were right (I didn't doubt it for a second). I wired up the second sensor, this time without a regulator, and it worked. I did the same with a third strip I had and it worked too . The readings now are 138/140 straight, then 120 and 10 degree bend, and so on. And the first one I hooked up now works without a regulator too. So it is possible I did do a little heat damage I can't see.

@Dave.

I took a couple of pics and posted a video demo of the second sensor I just wired up for you. It's pretty straight forward so you'll get this straight away. But if anyone else coming across this who wants to do the same, I wrote a quick tutorial as well. So here's how I did it...

1.) Using a servo extension cable, snip off the male connector and strip the white and red wires. Then tin the ends with solder.

2.) In the middle of the White and black wires, strip back the outer casing to bare the wire, and tin with some solder.

3.) Solder a 10 Kohm pull up resistor to the tinned White and black wires mentioned in step #2.

4.) Carefully and quickly (as has been mentioned in above posts ), solder the White and red wires to the sensor connections.

5.) Add some heat shrink to insulate the solder joins.

6.) Connect the servo wire plug to an ADC port, wires matching the pins.

7.) load up an ADC meter in ARC, set the port the sensor is connected to, and bend away. As mentioned above, you should see a static reading of about 140, then the further the sensor is bent, the lower this value will drop (as seen in video below).

NOTE: For a little added stability, apply a small drop of clear silicone adhesive over the crimp connectors on the flex sensor, as thay are a bit fragile and this will give added strength.

Awesome @Steve G!

Thank you for sharing your findings and creating a great tutorial!

No problem at all. Just happy I can share and give back to the community.

Sweet. Thanks.

No problem Dave.

Hope it helps.

Love the tutorial vid and that makes sense for specific animatronic things. However, and I apologize if I missed this in another thread, I was hoping to utilize the ADC ports to operate flex sensors in this sort of function:

http://www.instructables.com/id/Arduino-Wireless-Animatronic-Hand-1/

I tested it out with the Arduino and it does what I want it to do, I just dont know how to "map" in the script the adc flex sensor input to the real time output of servo control.

My hope is to use ADC 0-4 to control servos on D0-4 repsectively on the EZ-B4. The other issue I am having is using the MPU6050 variables to control servos as well.

I have some ability with circuits and assembly, but I have the programming skills equivalent to just learning that the strings on my shoes keep them on, next year I hope to learn to tie them...

Again I apologize if these question have been answered elsewhere.

That's easy - if you know how to do it in complicated C++ Arduino, then you can certainly do it even easier in ARC . Here's how...

1. Add a new EZ-Script control

• Project
• Add Control
• Scripting
• EZ-Script

2. Paste in this code

:loop

Servo(d0, getadc(adc0))
Servo(d1, getadc(adc1))
Servo(d2, getadc(adc2))
Servo(d3, getadc(adc3))
Servo(d4, getadc(adc4))

sleep(100)

goto(loop)

*Note: you may have to add a multiplier of sort to the GetADC() for each of the positions to dial in your accuracy. As you may have seen with your Arduino setup, the resister values on the flex sensor determine the accuracy and voltage being returned.

Thank you DJ! I am very proficient with copy and paste This makes me so happy, I've spent months reading how to books in arduino trying to figure out this I2c and wifi thing to no real avail. I can get the circuits to read right with meters but seldom get the code to operate correctly. (impatient with paying work getting in the way I guess), then I saw your creation and fell in love with it. A heartfelt thanks to you and your team.

What would the multiplier you speak of look like?

@artchas... analog values are from 1 to 255...servo positions are from 1 to 180, so you would use a "multiplier" to convert them to a servo position... The multiplier in this case is 0.7058... You would multiply all your analog readings by .7058 and that would covert them to a servo position from 1 to 180....

:loop
$adc0= Round(getadc(adc0) *.7058,0)
$adc1= Round(getadc(adc1) *.7058,0)
$adc2= Round(getadc(adc3) *.7058,0)
$adc3= Round(getadc(adc3) *.7058,0)
$adc4= Round(getadc(adc4) *.7058,0)

Servo(d0, $adc0)
Servo(d1, $adc1)
Servo(d2, $adc2)
Servo(d3, $adc3)
Servo(d4, $adc4)

sleep(100)

goto(loop)

Thank you Richard for the explanation!

So literally a mathematical function is in place. I see. Hypothetically I could then use this same concept to acquire and apply reading from a accelerometer for servo control as well? I will look at the SC32 already in the EZB script and see how that is called for reading when I get time and see if I can figure that out. I am guessing I would need to use the add control function for the MPU6050, then set up a script that tells it to read that (similar to above with the exception that I would use the Hex address X or Y variables) and then assign that variable to a servo...? Am I in the ball park?

I see that the MMA7455 under I2C is already set to control servos in the settings. I will look at that too.

Thank you again