Hi
I'm currently building a mowrobot and there is only 1 major electronic component missing. It is a sonar.
I actually never used a sonar so I only lately started reading up on sonars and I was kind of disappointed. I thought I could have a sonar that could have view about 80cm wide. I thought so because the big mow and the park mow have a sonar that is mounted directly on the robot (not using a servo but fixed liked in bolted to the chassis)
But searching for sonars I don't find any that have this wide view and can be used outdoors 24/7
Can anyone direct me to a sonar that fits my requirements? Or a document that can help me understand sonar better so I can make a sonar that leads to a solution of my requirements?
By their very nature, ultrasonic transducers aren't going to have a wide "field-of-view"; they produce a narrow "beam", that widens out with increasing distance. If you need to cover a wider area, you either need to scan the beam (using some kind of actuator), or you need to use multiple transducers.
While you could (in theory) repurpose something cheaper, and perhaps protect it from the elements and dirt, it makes far more sense to spend the money up front for a sensor that is already professionally designed for such environments. Another supplier of "robust ultrasonics":
cr0sh
Thanks for the feedback. I've looked into the sites and the closest I found to what I need is: HX40STRC Aluminium Can or HX40STRE British pipe threads on HX19 Ultrasonic Positioning System
I think these are senders and receivers in one box. The beam pattern is nearly 180° which is great but the sender has only 12° transmitting. Which means that if I want to scan in front of a 50 cm wide robot I need a 2.38 meter long robot (50/(2*tang(6°))
Am I missing something?
cr0sh:
While you could (in theory) repurpose something cheaper, and perhaps protect it from the elements and dirt, it makes far more sense to spend the money up front for a sensor that is already professionally designed for such environments. Another supplier of "robust ultrasonics":
I fully agree. This is why I prefer buying working stuff that fit my requirements even if it cost more. This because it saves me time and frustration.
Thinking about this I was wondering.
Would it be a good idea to use a car sonar? You know the sonars used to safely reverse a car. They are placed in the bumper. I mean you have typically 4 for a car which is about 2 meters so that makes at least 50cm for each sonar. Which is exactly what I need.
These sonars should be about 50cm above ground to avoid ground interference which is feasible for me.
They are designed for outdoor use.
There are however 2 things that bother me.
They normally work when reversing. Will they last long when used full time?
Jantje:
cr0sh
Thanks for the feedback. I've looked into the sites and the closest I found to what I need is: HX40STRC Aluminium Can or HX40STRE British pipe threads on HX19 Ultrasonic Positioning System
I think these are senders and receivers in one box. The beam pattern is nearly 180° which is great but the sender has only 12° transmitting. Which means that if I want to scan in front of a 50 cm wide robot I need a 2.38 meter long robot (50/(2*tang(6°))
Am I missing something?
It's +/- 12 degrees - so a beam width (center lobe) of 24 degrees total; rerun your calcs I think - you'll probably end up somewhat smaller.
Note that these devices are single transducers tuned to allow sending and receiving - you must build your own transmitter/receiver circuit to drive them (fortunately, hexamite gives examples, including one for the ATMega - which will be OK for the Arduino).
Is there any reason why you couldn't mechanically scan the sensor in front of the robot (or use multiple devices)?
Jantje:
Would it be a good idea to use a car sonar?
...
They normally work when reversing. Will they last long when used full time?
How can I connect them?
It would be worth trying; if you are simply trying to detect something in the way (vs actual distance), then using such a device might prove simpler than other means (you could just tap into LED outputs or similar from the device). I don't see why they wouldn't last/work just as well all the time vs intermittently. There's also the fact that it's already designed for outdoor use...
Crosh
I found this page on using Bosh car sensors. The guy hacked the protocol between the sensor and the UAP (ultrasonic park assist). If I understand it correctly rx tx can be used to communicate at 9600 baud. http://ph-elec.com/archives/ultrasonic-sensor-hacking-step-1/ The first 7 pages are on hacking the UAP. You can easily skip them and go to page 8 directly.
Seems very interesting for my project. Surely as these sensors seem to be very common in cars (GM) so I could get a set easily.
Not sure I fully understand the impact of the multiple echo's and the rx tx stuff. (I'm not an electronics guy)
It's +/- 12 degrees - so a beam width (center lobe) of 24 degrees total;
Indeed I get to 1.18meters. Having about 50cm of robot 2 could do.
Is there any reason why you couldn't mechanically scan the sensor in front of the robot (or use multiple devices)?
I don't understand what you mean with "mechanically scan the sensor in front of the robot "
I could use multiple devices but each device means a construction to position and a hole in the cover plus additional complexity in the software.
As I only want to slow down the robot when there is something ahead and stop based on bumper contact; I would prefer 1 sonar scanning the front plus some part of the side of the robot.
I have build an object avoidance robot in the past based on a ir on a servo. That worked great but then I realized a mow bot is not an object avoidance robot. A lawn mow robot needs to go straight until it "hits" something and then take another direction. So the sonar is only there to avoid the impact when you reach "something". In other words I'm not interested in scanning the environment. I'm only interested in adopting the speed based on the "chance of having an impact".
Best regards
Jantje
Jantje:
Crosh
I found this page on using Bosh car sensors. The guy hacked the protocol between the sensor and the UAP (ultrasonic park assist). If I understand it correctly rx tx can be used to communicate at 9600 baud. http://ph-elec.com/archives/ultrasonic-sensor-hacking-step-1/ The first 7 pages are on hacking the UAP. You can easily skip them and go to page 8 directly.
Seems very interesting for my project. Surely as these sensors seem to be very common in cars (GM) so I could get a set easily.
Not sure I fully understand the impact of the multiple echo's and the rx tx stuff. (I'm not an electronics guy)
Interesting - I don't know how "common in cars" ultrasonic parking assistance systems are (certainly not available on my 96 GM truck!); the greater issue would be if you hacked one, got it working - everything was ok for a few years, then it failed in some way that you had to get another, and you find they changed the system (so you would either have to re-hack it in a different manner, or do something completely different). At least with the hexamite or other ultrasonic sensors, you wouldn't have to worry about that issue.
Jantje:
I don't understand what you mean with "mechanically scan the sensor in front of the robot "
I could use multiple devices but each device means a construction to position and a hole in the cover plus additional complexity in the software.
What I mean by "mechanical scan" is to mount the sensor on some kind of rotational (yaw) platform or such that allows you to "ping", get result, rotate the sensor so many degrees, "ping" again, and so forth (just like a radar works); its not as fast, but it does work ok.
That in mind - have you thought about or tried a Sharp IR sensor (you'd definitely have to scan with one of those)?
Jantje:
As I only want to slow down the robot when there is something ahead and stop based on bumper contact; I would prefer 1 sonar scanning the front plus some part of the side of the robot.
I have build an object avoidance robot in the past based on a ir on a servo. That worked great but then I realized a mow bot is not an object avoidance robot. A lawn mow robot needs to go straight until it "hits" something and then take another direction. So the sonar is only there to avoid the impact when you reach "something". In other words I'm not interested in scanning the environment. I'm only interested in adopting the speed based on the "chance of having an impact".
Another possibility might be to use/build a high-frequency "chirp" sound sensor; basically like an ultrasonic sensor, but more in the audible range (20-30 kHz?) - that would probably spread the beam out further, but you would need a fairly good quality microphone to detect the return...
CrOsh
To be honest I was pretty skeptical when I read your proposal.
I have done some reading up and it looks like a good idea right now.
I have all electronics in a watertight box. There should be no problem having 2 extra connectors for the US1440.
And these are wide beamed outdoors running at the same frequency
I'll look for a place to order them for Belgium
Best regards
Jantje
Crosh
I completely missed your post of 27 Feb.
Thanks for the extensive answer.
As to the In 5 years problem. I don't think is really is a problem. These sensors have been used for years. Reading through the aritcle it seems to cost a fortune to make and test them. So it is unlikely they will disappear soon. I however completely agree that the "non documented interface" and the "not comon usage" of a "component of a solution" can have serious drawbacks at he level of sustainability.
As to the "mechanical scan". I have a indoors robot with a IR distance sensor running. I have bought several of these sensors so I have them lying around. I was originally planning to use them and I probably still will do for the finer work (mowing next to walls; finding the recharge bay;...) However I have been told that IR scanners can have problems with sun light.
Next to that having a "mechanical scan" means having a servo (or other type of motor) at the top of my robot which does not really comply with outdoors.
As to the chirp. That might be a great idea but I really wouldn't know how to start it. I prefer buy and build (a shield gets a high score) because I have to little electronics experience and to little time. It is getting spring
Jantje:
Crosh
Could I use Distance Sensors with the Ultrasonic Sensors (US1440) sensors.
This would have the benefit of only one provider
Best regards
Jantje
Yeah - quite possibly (the only problem you would probably face on such a retrofit is whether the pin spacing of the transducers will match up; if not, you'll need to extend some short leads); something else to keep in mind - check this datasheet (for the part right below the US1440:
Apparently, that particular part (and likely the 1440 as well - not sure) can be used as a "singular" sensor - instead of needing two of them with separate send/receive sections; of course, building such a thing would require you to find all the other components, as well as have at hand an oscilloscope (you can't really successfully do this kind of thing without one; indeed, even modding one of the boards with different sensors - it may be good to have a scope handy).
Cr0sh
I'm planning to use leads anyway. To get the electronics inside a box and the waterproof sensors on the correct place on the robot.
During my study I bumped upon some threads on using the 1440. It basically states: you buy a set. 1 is a sender the other is a receiver. If you switch them it wont work.
Anyway I'm ordering tonight 8)
The dice has been thrown
Thanks and I'll keep you posted
Best regards
Jantje
The package arrived today.
I'm all happy. I tried the Arduino ping example with the non modified sensor and it works (it needs some tuning to get correct measurements but it works)
So I'm heating my solder to replace the sensors. Waiting for the solder to heat I unpack a sensor and I notice that 1 lead is longer than the other
Does that mean anything?
Can I connect them wrongly?
Best regards
Jantje
ps
I also noticed that the sensors came as 6 seperate packages (I ordered 3 sets) so I think the comment below is wrong.
Jantje:
It basically states: you buy a set. 1 is a sender the other is a receiver. If you switch them it wont work.
I replaced the sensors with the outdoor sensors.
No results so far. The receiver seems to work as playing with my keys provides different reads.
But it seems dead.
I will try to do some more tests tomorrow.
Best regards
Jantje
Hi
I did some more test today and it kind of works now.
What happened is that I spoiled the pcb of the first sensor card when I removed the sensors. Therefore all my tests were negative. I found that one out because it didn't work with the original sensors.
I have removed the sensors from my second pcb (destructive for the sensors to safe the PCB) and tried with the original sensors of the first card. That worked fine.
Then I tried with the original transmitter sensor and 4 (I unpacked 4, assuming if there are 3 transmitters and 3 receivers there should be at least one of each in the set of 4) outdoor sensors.
All worked (tough not so great as with the original sensors).
Then I tried with a original receiver and the same 4 as transmitter. All worked (tough not so great as with the original sensors).
But when I place the outdoor transmitter and the outdoor receiver I get "128in, 329cm" all the time.
If I ring my keys I do get different measurements.
This sounds like a 'insufficient power issue" to me. However I have no clue whether this is correct. And if it is correct what do I do? Is arduino not capable to deliver the power? Should I add a capacitor? Do I provide a higer voltage to the card?
Any input is welcomed.
Best regards
Jantje
