Sync via RF.

Go To Last Post
19 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello again:

As lazy as always, I want some ideas about a hobby project I have in mind right now. Since I'm a noob with this, some of the RF guys here could give me some hints about the simple thing I want to do.

Mainly, what I want to achieve is to syncrhonize few different ATmega64's by RF means with another uC (probably AT91SAM7S64 or SAM7X256). There would be a single Txmitter (the ARM or whatever), and few receivers.

The M64's are currently soldered to a PCB, and have the IRQ0 and IRQ1 pins free, so I would use this as the trigger to Sync the whole bunch, attaching a simple receiver to that pin.

So one possible idea is to generat a few MHz signal with the xmitter, amplify it, apply that signal to an antenna, and use some NE567 or whatever to detect the signal, triggering the interrupt at the receivers

The intended precission would be arround tens of us, if possible, below 0.1 ms. Range <100m in open space, line of sight. Hobbyst purposes, no FCC, or any other regulation.

But I don't know if this would work, if there are better solutions, or other means to achieve the same. If it works, I don't have any idea about the amplifier to be used, neither the antenna layout (probably a simple loop would do for such low frequencies), the receiver or any RF stuff.

Can some point me in the rigth direction?

If it works, I probably will post results and some related stuff, but don't expect anything soon.

TIA,

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Oh, I forget. This is a localization system for hobbyst robotics, using ultrasonic multilateration. The xmitter would be mounted on the mobile robot (running at a low speed), together with a simple ultrasonic xmitter, and there would be few ultrasonic receivers (the ATmega64 units) that will report the time of flight back to a PC. The receivers are already build, since they are obsoleted commercial products.

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Cool project.

How do the remote sensors connect to the PC right now? Is there a wired or wiress link already in place for that part? Assuming the remotes have to talk to the PC anyways, the PC could certainly send the synch signal to the remotes. This would mean that only the robot to master PC would need a new wireless link to carry the synch signal.

Depending on your budget, and your ability to synch otherwise, another (expensive) option would be to put a small GPS receiver on each unit with a 1 pulse per second output pulse. Each remote sensor would then send to the PC its data which would be Seconds.xxxx seconds into the current minute. The GPS would be used to provide a very accurate, synchronized, time signal, not for gps positioning.

Remember that in RF the antenna is the most important part of the system!

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The first thing that comes to mind is TR1001.
http://www.rfm.com/products/data...
In 115kbps mode it should give you less than 8usec error for a one way synchronization, and 1-2us for two way.

Building a receiver takes quite a bit more effort than a tone detector. At 13.56 or 27.12MHz you need to build a proper tuned RF stage, mixer, and detector in order to function. The transmitter likewise really needs the tuned output.
The 13 and 27MHz ISM bands are suggested since you can legally build, test, and operate a transmitter in those bands without a license. Hobbyist purposes or not, I can't assist you in committing a crime.
http://www.erodocdb.dk/docs/doc9...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

DocJC:

Right now, the communication is trhough RS-232 using the ISP header. I'm developing a PCB that will do RS-485 attached to the same header, that I plan do make publicly available. So comms will be wired and with quite standard ModBus protocol, that I already know, and I'm currently working on it.

Since that means to run wires, wiring all of them together to the PC or another means is a possibility. But the mobile platform wouldn't support it.

KKP:
I also prefer to use ISM bands, like the ones you suggest. I also didn't want to commit any crime ;), and I also want to do something that could be shared by other hobbists. TR1001 is an interesting approach, but I think it would be a little difficult to solder for many hobbists. But that remembers me another 'cheap' solution that could be XBee modules, or even WirelessUSB modules from Cypress, but I'm not shure if all of them would fit the bill.

Anyway, this suggestion could be an interesting starting point. I will investigate along that lines.

BTW, I have a Raven Kit, that I would use for both comms and synchronizing, but I think I only could assume the first part.

Thanks for your help and suggestions. I keep you informed.

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Guillem i already work with Radiometrix's modules, tx and rx 173.250-10(long distance, choose other for short distance), you could make a simple protocol to sync yours uC once that this modules doesn´t modulate the digital signal, i mean they transmit as you send.

Other option i saw today in the saelig website:
http://www.saelig.com/miva/merch...

But i think that is just for sync yours uC the radiometrix could be a low-cost approach.

Brunomusw

Regards,

Bruno Muswieck

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks, Brunomusw. Can you give me an idea about it's cost?

Right now I'm reading Cypress datasheet for their Wireless module, that is really cheap (about 10€ at Farnell), and my first option.

Thanks again.

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Guillem Planisi wrote:
This is a localization system for hobbyst robotics, using ultrasonic multilateration.

I was thinking of building a similar system, but was put off by the attenuation of ultrasonic frequencies in air. According to this tool:

http://www.csgnetwork.com/atmossndabsorbcalc.html

20 kHz sound attenuates at 52 dB/100m in sea-level air, not including the attenuation due to distance. I was thinking the situation might be improved by using a "chirp"--that is, a tone which increases in frequency over the duration of the tone--instead of a single-frequency tone. It might be easier to pick this out of the background using correlation.

Have you run any numbers on the strength of transmitter and sensitivity of receiver needed?

Michael

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I use LINX modules with great success. If all you want to do is send an interrupt sync signal, you can find a whole load of simple RF units in Digikey.com If the shipping costs are out of your price range from them do waht I have done in the past....Paypal me the cost of the devices, and hteir shiping to my home. I will then send them to you via USPS. Done it for a few folks and it works great.

Jim
P.S.... When you comming to NYC? :D

I would rather attempt something great and fail, than attempt nothing and succeed - Fortune Cookie

 

"The critical shortage here is not stuff, but time." - Johan Ekdahl

 

"If you want a career with a known path - become an undertaker. Dead people don't sue!" - Kartman

"Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?"  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Well, although I had made some numbers, that I don't have at hand, I had done many tests to the comercial product, that detects a wall at 12m without problems. Since that means receive an echo, the attenuation is much bigger than for direct receiving (fourth power of the distance instead of quadratic power). That insures me that I could achieve more than 10m. In fact, I would expect more than 20m, but that is such a big distance that I probably wouldn't need at all.

Of course, if you want ranges about 100m, this could be a serious drawback. Anyway, I would not recommend sound in any form to do this in open space, since wind and weather could affect it a lot.

Chirp modulation is a really good idea, that military radars had been using it for some time, but it has one drawback: ultrasonic xduzers are usually narrowband, thus this modulation is severely restricted in bandwidth.

BTW, this systems are working at 41.6KHz

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Jim, congrats ;) (I refer to your post at John/AIIN contest)

I will check Lynx modules, as well ad Digikey, but Cypress modules offer an interesting approach. Specially because there are some network project, open source if I recall correctly, to use them with AVR's. So firmware would be more or less written.

Anyway, it wouldn't do any harm to me to check your suggestions.

BTW, I will came to NYC as soon as my wife decides to do so. Unfortunately, I don't think this gonna happen soon :(

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Oh, I forget, Jim: thanks for your offer. Luckily I can afford them, and I had ordered other things from Digikey in the past, for personal uses also.

The 'money issue' is for other hobbists that would use them, since I plan to do it publicly available. After all, I'm the only 'AVR defender' in this 'PIC land'.

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

given 1nSec per foot, multilateration for robot position tracking will be tough to do at low cost

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Can you give me an idea about it's cost?

Guillem, the one that i use is for long range, so you need to look for a short range module and look for a seller in spain or around.

Sorry..

Brunomusw

Regards,

Bruno Muswieck

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello Guillem.

Apart from the attenuation problems of using ultrasonic transmissions for time of travel calculations, you have mentioned weather and wind having an effect. These effects are in fact the basis of the operational concept of the ultrasonic anemometer. Time of travel is a function of the separation distance, the wind vector, the air temperature of the path and humidity (a secondary effect). Unless you know the air temperature and humidity, I doubt that you will be able to achieve any accuracy in measuring distance. Just my $0.02 ....

Cheers,

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Je, je. Yes Ross, I know perfectly the problems related with wind and so on. In fact, my intended target is a precission of about 2-3 cm (one inch would be perfect) or less in indoor applications, where wind and other meterological effects will be minimized.

I had done tests with the commercial products, and I had an accuracy better than the expected when aiming to walls or big objects, and this means strongest attenuations due the echolocalization technique, that is less precise and more sensitive to environmental effects.

And about ultrasonic anemometer (that also measures temperature), Monsieur Sylvain Bisonette has one developed in his web page. Quite a nice project. And he uses the same ultrasonic xduzers I will be using. So I was aware of this.

Anyway, thanks for your hints. IIRC, you work for metereological insitutes, so your advice about this issue is important to me.

Brunomus, thanks anyway, I will try to contact spanish distributors. I had to this anyway ;)

Stevech, I plan to measure the distance by means of speed of sound, not lightspeed. As you say, an 1ns per foot would be anything but cheap. I will only use the RF to sinchronize different crhonometers that will measure the time of flight of the ultrasonic wave emitted at the same time than the RF pulse, but receive much later by the commercial products. At about one foot per ms, with 0.1ms accuracy I would be more or less happy.

The commercial products (receivers) will take into account the attenuation due distance (that is proportional to the time of flight) and the temperature (they have NTC sensors that do the thermal compensation). So this effects will be more or less compensated. Directivity would be another issue, but I'm pretty shure that I could work around it.

Thank you all for your advice. I will update the progress on this.

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Guillem Planisi wrote:
Je, je. Yes Ross, ...

And about ultrasonic anemometer (that also measures temperature), Monsieur Sylvain Bisonette has one developed in his web page. Quite a nice project. And he uses the same ultrasonic xduzers I will be using. So I was aware of this.

Yes Sylvain does good work. I designed my first one in 1979 using a 6809. It also measured temperature, wind speed and direction. It assumed relative humidity because it was to be used on a drifting sea buoy 1 metre above the sea surface. This way the air temperature "sensor" never got wet and therefore did not become a "wet bulb temperature" sensor.

Guillem Planisi wrote:
Anyway, thanks for your hints. IIRC, you work for metereological insitutes, so your advice about this issue is important to me.

Past tense ... worked for 20+ years there off and on.

Guillem Planisi wrote:
Thank you all for your advice. I will update the progress on this.

Also happy to "wax lyrical" ...

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

That was a nice project, Ross. For shure not easy, since the worst problems would be the environmental situation. A sensor that would work on the sea should be rugged and well designed. The little aluminum xduzers that I have will be corroded really fast under that circumstances. And that's not to mention that usual xduzers are 'open face', so saly water could go inside and do a mess.

Well, 20+ years of experience is a valuable knowledge, so I will reinforce my statement that your opinon about this kind of things should be taken into account ;)

Luckily, I have 1+ years of experience with ultrasonics, so I think I could finish this project succesfully.

Guillem.
"Common sense is the least common of the senses" Anonymous.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

After reading Cypress datasheet, I had decided to go on with it's WirelesUSB modules. They are cheap, FCC compliant, easy to find, easy to interface, they fullfill my requirements, and IIRC, there is some comms protocol OpenSource or similar, developed for AVR's.

The other alternative, that I don't discard, is the XBee modules, that probably I will also use. But since they probably have less accuracy to assert any IRQ, I feel that they don't fill my timming requirements. Instead of this, other communications need that I have would be 'transparent', thus the reason why I will probably use them too.

Guillem.
"Common sense is the least common of the senses" Anonymous.