New robot article in MAY issue of SERVO magazine
rdo...@bizserve.com
For those of you that subscribe to SERVO magazine keep an eye out for the new MAY issue that is starting to be mailed out now. I have an article in there on creating a small robot based on an in expensive trash can that I picked up at target. The robot body is approximately 12" tall. You can basically install the same brains and drive that many of you already have bolted onto a flat chassis. The article covers how you can take all that and install it in the can (under $8) to make a nice finished robot. I used the $2 sonar sensors and some a couple small 9G servos to control the servo pan and also automate the lid movement. The response has been good so far and I hope you enjoy it.
Robert
Keith Mc
> I used the $2 sonar sensors and some a couple small 9G servos to
> control the servo pan and also automate the lid movement.
Which MPU? How much processor overhead does it/they take?
(If you are using a Propeller, how many COGs does it/they eat up?)
I'll start with some of the options I've seen and tinkered with, but
much of my experience is from many years ago. Things have
changed a lot in the interim, and I'm curious as to see what
other people (and the market) has/have done with them LATELY.
Who here has used any ultrasonic sensors lately? Which one(s)?
What is your experience with the current crop of them, esp
WRT process overhead, sensitivity, reliability, et al?
Has anyone arrayed a LOT of the HC-SR04s together?
If so, have any good code for running a bunch of them?
(eg I don't see a multiplex driver for them on OBEX yet).
- Keith Mc.
------------------------------------------------------------------------
My experiences, and some of the options I've seen:
(Please add to this list!)
1) DIY, with raw sensors from old Ultrasonic (Us) remotes, piezo speakers, et al.
How many started in US sensing years ago. Versatile, as you can
drive them any way you wish, and MAYBE being able to try to use
them for low res data exchange as well as ranging. (but I never heard
many successes with data exchange, other than with X-10 remote controls)
The raw transducers aren't cheap to buy, but I do hear of people cannibalizing
something newer like an SR04. Some designs did require funny HV
power supplies, and without proper docs for the wafers they were often
a pain to tinker with. They also had a tendency to fail suddenly from
environmental exposure, handling and mounting issues, HV arcs, and
in general were often "POM Dependent" (Phase Of Moon... {wink}
IOW who knew WHY it failed, IF it WOULD work, etc. :-)
2) Polaroid SX-70 ranger hacks (the Polaroid 6000 transducer).
Harder to find now, with Polaroid's downfall (due to the digicam revolution).
Many of the hacks sometimes had a tendency to fry themselves, but if
you got one working right some swore by them. With Polaroid stopping
production of cameras and the SR04 around now you don't see them
as often anymore. I'm not sure if the sensors are even still being made.
(Can anyone comment? ARE they still being made now?)
3) Parallax 28015 "PING)))", at $30. (combined trigger and data lead, you do timing)
That always seemed a bit high for a sensor since you must do the timing,
AND you are multiplexing the trigger and return signal lead..
Sure it is less pins, but you still have to dedicate some resources into
running it.
4) The ubiquitous HC-SR04 - ~$2 (separate trigger and rtn data)
This seems to be the cheapest of them all, and is now WILDLY popular.
Separate data & trigger helps vs PING))), but you are still responsible
for real time processing of the return pulse. However, at $2, it
is definitely the choice for experimenters who want cheap hardware!
With MPUs becoming cheap, dedicating a second MPU to
handling a bunch of these is definitely an option.
5) Maxbotix.com "EZ" series - list ~$50 (but really more like $20-25 each)
This one has an on-board MPU doing ALL of the processing for you.
It also has multiple interface pins, and many versions with different
beam widths. Talking to it is EASY. They give you both pulse I/O and
"analog out" pins with a distance value on it that is constantly being updated.
They are also easy to daisy chain with a single lead from one to the
next (eg trigger ONE, and then each one fires the next one in a ring :-).
This frees your processor up vs many other solutions.
Available in quantity purchases at a lower price.
My favorites:
The EZs work well, and are easy to use. But at $20+ per copy,
I'd definitely use them for industrial and other "funded" robots, like FIRST.
I'm not THAT rich to buy a large quantity of them for tinkering, or
creating huge arrays of them on EVERY robot for servo-less operation.
(I have a FEW of them... I do like 'em...)
I haven't worked with one yet, but at $2/copy that HC-SR04 is starting
to look *mighty darn attractive* for larger arrays and multiple bots
(eg my RA...). I'm just curious to hear about any experimental results
of attempting an array of them, or what happens when you have a
population of them in an area.. I've only seen limited use on a single
robot, and/or ONE on a pan/tilt steering servo.
I'm curious to see IF an array on multiple bots in an area CAN be useful...
Next: Are there any issues with making a wide angle SR04 array work?
24 * $2 is still under $50 for a 360 deg FIXED ring array (they are
~15 deg each). Simply mux a few data/trigger pins, and you have it.
BTW, I've seen HORRIBLE results with US in general, in FIRST contests.
Multiple bounces, "invisible objects", "bank shots" making walls invisible,
crowd noise, robot bumpers being "soft", and other bots interfering
because THEIR US systems are in operation all make it questionable
for using as a reliable Roaming Object Sensor.
IR seems to work MUCH better for most object sensing in FIRST.
(For example - a Sharp distance sensor is especially useful for
determining wall distance in Autonomous Mode operation.)
That's my input. What have YOU used?
How well did it work? What issues have you encountered with US sensing?
- Keith
rdo...@bizserve.com
Hello Keith,
I am using the HC-SR04 modules. I made a pair of these robots and use a Propeller QuickStart as the brain for one and an Arduino as the brain of the other. I'm only using a single HC-SR04 module since I have a servo on the top for panning and also one on the lid that enabled tilting the sonar up and down. Eventually I may add some sharp IR or other sensors on the body.
There are examples out there for the Arduino. I started to modify the Ping object for the Propeller but found that someone on the forum already did it. The changes were pretty trivial and mainly consisted of separating the I/O pins (instead of sharing it) and you can leave out the code to toggle the direction of the I/O pin. I believe some people out there are multiplexing some of these to reduce overall pin count. For the projects where I've been using them it wasn't an issue so I didn't bother with that. If I had a lot of them then I would have looked into it. These sonars act similar to the SRF04 modules which I believe they named them close to get people to try the cheap ones. I also used this with the LEAF project board which was setup for sonars with separate trigger and echo pins. On my LEAF based robot I have two real SRF04 sonars (front and rear) and have four of the cheaper HC-SR04 modules (on each side and on the front at 45 degree angles)
As far as overhead there isn't much. Using the hacked PING object it doesn't even required a dedicated COG. Very easy to use. They seem to perform ok and I think the issues I ran into using them on the small trash can robot has more to do with where they were mounted. That is covered in the article.
Reliability varies. I've heard of people that had a couple modules DOA. So far I've gone through a bunch of them and had only one bad one. There was one other sensor that didn't work but I found that it was my fault. When I mounted it the screw knocked loose a surface mount cap. Once I resoldered it the module started working again. If you buy them in quantity 10 you can get them for around $1.70 so I can live with a small percentage of bad ones.
If you find a good way to multiplex a bunch so you could have a whole ring of them around your robot I'm sure everyone would be interested in that as well as an object for it.
Robert
--
You received this message because you are subscribed to the Google Groups "MiRobotClub" group.
To unsubscribe from this group and stop receiving emails from it, send an email to MiRobotClub...@googlegroups.com.
To post to this group, send email to MiRob...@googlegroups.com.
Visit this group at http://groups.google.com/group/MiRobotClub?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.