Giving PCB/Controller Wireless Capabilities

[gomidas2007]

Hi all,
I'm a Mechanical Engineer but I'm going into uncharted territory both for my company and myself so I could use some help. We currently have a line of trash compactors, of various sizes, which are used in commercial applications from restaurants, to airports, to buisnesses and are controlled by a standard PCB. The controller operates the machine; determines if it is full, if there is an error, etc. We are now trying to integrate Wi-Fi capabilities to these machines so they can be accessed from a single location. Ideally, this would allow the company, or the customer, to network the machines to allow for easier maintenance, i.e. which machines are full/ready to be emptied, if there are any error messages, etc. Ideally, this would be something that could be optional for customers, but it could also be a permanent feature. I really have no idea where to start so any feedback would be much appreciated!

Thanks,
Tyler

 

[schip666!]

For a Q&D demo you might be able to use an Arduino hobby micro-controller board with a wifi or bluetooth "shield" (plugin extension board). Just search for arduino and you should find all sorts of interesting, and sometimes accurate and useful, information. I would also go look at http://hackaday.com/ for tidbits on "wireless hacks", and maybe http://www.sparkfun.com/ for arduino and other hobby level solutions.

Once you've got something that works you may be able to design your own circuit board to do everything rather than cobbling a bunch of toys together.

 

[DragonPetter]

How will the data be collected? If you have access to internet, you could do like schip666! suggested, or you could put a wifi transceiver with a microcontroller on the equipment. If they are in places without access to any internet location, you could use a cell modem to call into an external server, although this is more expensive and complicated.

 

[berkeman]

Hi all,
I'm a Mechanical Engineer but I'm going into uncharted territory both for my company and myself so I could use some help. We currently have a line of trash compactors, of various sizes, which are used in commercial applications from restaurants, to airports, to buisnesses and are controlled by a standard PCB. The controller operates the machine; determines if it is full, if there is an error, etc. We are now trying to integrate Wi-Fi capabilities to these machines so they can be accessed from a single location. Ideally, this would allow the company, or the customer, to network the machines to allow for easier maintenance, i.e. which machines are full/ready to be emptied, if there are any error messages, etc. Ideally, this would be something that could be optional for customers, but it could also be a permanent feature. I really have no idea where to start so any feedback would be much appreciated!

Thanks,
Tyler

Welcome to the PF.

Networking devices using RF has a number of challenges. See the discussion in this white paper, for example:

http://www.google.com/url?q=http://...ds-cse&usg=AFQjCNEss18uPWjE7UpJDgcvyfJdhzAdbQ

Also, I'd stay away from the hobby approach, even for demos. You need to find a workable scheme for use in your target markets, so having a simple demo in your lab is not going to tell you much. You need all the noise and multipath and installation issues thought through as part of your investigation, before you pick an approach to test out.

With all of that said, if you can be certain that there are Wi-Fi routers in all of your target installations (or you can offer to install the routers in locations where they are absent), then I'd just use the standard 2.4GHz Wi-Fi band to do the networking.

I'm not real familiar with the best way to network devices using the Wi-Fi band, but here is a simple Google search to get started with. I'll try to spend some time looking more into it in a day or two if others haven't posted about it yet.

http://www.google.com/webhp?sourcei....2,or.r_gc.r_pw.r_qf.,cf.osb&biw=1149&bih=882

.

 

[gomidas2007]

Yes, I would like to stay away from the Hobby approach, although I do appreciate the suggestion, it just doesn't give a broad enough scale to know whether it will work in the field, as Berkeman mentioned. These machines are generally used in large enough scale operations that Wi-Fi routers should be available. Places like airports, Mcdonalds, that sort of thing. And I like the idea of offering to install routers, I think customers would really go for that. There is a company, called Bramidan, which offers Modems on their balers, as DragonPetter mentioned, which is a solution we are investigating as well, but Wi-Fi is the route we would like to go. I'll do some research into the links you all have posted. Thanks for the help! Hope to hear more input! I'll try to keep you all updated as I move along. Like I said, this is completely new to me as well as the company.

 

[Antiphon]

You can drop on an FCC-approved module.

The module will need access to the product's status on the PCB but otherwise will simply drop on.

Try this:
http://search.digikey.com/scripts/DkSearch/dksus.dll?x=3&y=19&lang=en&site=us&KeyWords=831-1009-nd

It's probably not the cheapest solution but it's got logic level inputs. Most of these modules need serial ports to move data around.

Expect to pay a consultant around $20,000 to program it for you.

(The Serpent in the Garden of Eden was the first consultant.)

 

[gomidas2007]

Ok, so I've been looking into a few things and have come across an interesting method called Power Line networking which uses existing electrical wiring to connect multiple devices. Do you guys know anything about this? Would this be a possibility in lieu of wireless networking?

 

[berkeman]

Ok, so I've been looking into a few things and have come across an interesting method called Power Line networking which uses existing electrical wiring to connect multiple devices. Do you guys know anything about this? Would this be a possibility in lieu of wireless networking?

Please check your PMs.

 

[gomidas2007]

Ok, let me throw another question out there; what is the difference, if any, between using a wireless module that drops into the PCB and using a wireless modem?

 

[DragonPetter]

Ok, let me throw another question out there; what is the difference, if any, between using a wireless module that drops into the PCB and using a wireless modem?

Do you mean one that solders/connects to the PCB vs. one that comes in its own housing? Are you talking about a gsm modem?

 

[gomidas2007]

Do you mean one that solders/connects to the PCB vs. one that comes in its own housing? Are you talking about a gsm modem?

Maybe?...One that comes in its own housing would have to connect to the PCB at some point anyways wouldn't it? Pretend I don't know anything about large scale wireless/wired networking. In fact, don't pretend...

 

[DragonPetter]

Maybe?...One that comes in its own housing would have to connect to the PCB at some point anyways wouldn't it? Pretend I don't know anything about large scale wireless/wired networking. In fact, don't pretend...

Well, I think it might depend on what kind of antenna you want to use, because you have to interface to that antenna some how, and keep the correct impedance in that interface.

You can get modem modules with a USART i/o interface on it, and you only need to make these digital connections and supply it with power to work. I did this with a satellite modem, and we used a patch antenna for this, and so it was very easy to just buy one and plug it into the already provided antenna connector on the module. I'm not sure how these look for GSM modems.

I have also designed a board that used a packaged GSM module that must be soldered onto the PCB and had similar form factor to a big microchip, and it only had an analog pin that must be interfaced to an antenna. The module was a BGS3, http://www.cinterion.com/products/m2m-value/lga/bgs3.html . I had to ensure that the trace impedance from that pin was 50 ohms which is easy enough, and we had to pick out the proper antenna connector to come off the PCB (we were using a patch antenna again) and make sure its impedance was as close to spec as possible. Again, it used simple USART interface to the microcontroller. Some of these PCB chips have active and passive antenna modes, and that can add extra complexity for layout since they might use external components for the active network. Also, depending on the form factor of your design, you might want to use a PCB antenna, like a F antenna, specifically a PIFA one, and those are challenging to design I think.

The other consideration you need to take, which I'm not sure if you have, is the amount of costs for certifying your product. If you use any of these devices, you need to go to a certified EMC testing facility and there are specifications specifically for GSM devices. I'm sure there are other ones for wifi too. For GSM, we had to do CISPR qualification, http://en.wikipedia.org/wiki/CISPR , and the facility to do it at was expensive.

I think all of this might be overkill for what you're trying to do. Could you setup some piece of hardware with access to a network, and that communicates to remote units using zigbee or something else like that? I think that would have less barriers than a wireless modem.

 

[gomidas2007]

So, if I want to design a control panel, which contains the PCB and all the other components, which has networking capabilities, I would have to buy what I'll call a "networking module" and then an antenna as well? Surely, there must be something that contains all the necessary parts and can just drop onto the PCB or connect via USB or pin-connector right? I do need this to be a fairly standard design which can be easily replicated as we will be building hundreds, and potentially thousands of these machines a year with any luck. Also, what is form factor?

Well, I think it might depend on what kind of antenna you want to use, because you have to interface to that antenna some how, and keep the correct impedance in that interface.

You can get modem modules with a USART i/o interface on it, and you only need to make these digital connections and supply it with power to work. I did this with a satellite modem, and we used a patch antenna for this, and so it was very easy to just buy one and plug it into the already provided antenna connector on the module. I'm not sure how these look for GSM modems.

I have also designed a board that used a packaged GSM module that must be soldered onto the PCB and had similar form factor to a big microchip, and it only had an analog pin that must be interfaced to an antenna. The module was a BGS3. I had to ensure that the trace impedance from that pin was 50 ohms which is easy enough, and we had to pick out the proper antenna connector to come off the PCB (we were using a patch antenna again) and make sure its impedance was as close to spec as possible. Again, it used simple USART interface to the microcontroller.


Some of these PCB chips have active and passive antenna modes, and that can add extra complexity for layout since they might use external components for the active network. Also, depending on the form factor of your design, you might want to use a PCB antenna, like a F antenna, specifically a PIFA one, and those are challenging to design I think.

The other consideration you need to take, which I'm not sure if you have, is the amount of costs for certifying your product. If you use any of these devices, you need to go to a certified EMC testing facility and there are specifications specifically for GSM devices. I'm sure there are other ones for wifi too. For GSM, we had to do CISPR qualification, and the facility to do it at was expensive.

 

[gomidas2007]

Yes, I think you may be right. I am looking into Zigbee per your suggestion, but the part that I need to come up with is exactly how and what is required for that kind of setup to be possible. It's the details that I'm really concerned with right now. Unfortunately, if I go to my boss with "we should use Zigbee", chances are I won't get this project :( I certainly appreciate the help though!

I think all of this might be overkill for what you're trying to do. Could you setup some piece of hardware with access to a network, and that communicates to remote units using zigbee or something else like that? I think that would have less barriers than a wireless modem.

 

[DragonPetter]

Yes, I think you may be right. I am looking into Zigbee per your suggestion, but the part that I need to come up with is exactly how and what is required for that kind of setup to be possible. It's the details that I'm really concerned with right now. Unfortunately, if I go to my boss with "we should use Zigbee", chances are I won't get this project :( I certainly appreciate the help though!

You probably should make a spreadsheet with all of the technologies you've researched, and rate them based on certain criteria and then rate your application on this same criteria and see which one matches the best. Here's some examples of criteria:

Range
Datarate
Number of units in a network
Reliability
Availability
Cost
Technical challenge

 

[gomidas2007]

You probably should make a spreadsheet with all of the technologies you've researched, and rate them based on certain criteria and then rate your application on this same criteria and see which one matches the best. Here's some examples of criteria:

Range
Datarate
Number of units in a network
Reliability
Availability
Cost
Technical challenge

That's a great idea! I had just been taking notes, but I like the spreadsheet way more. Much more organized. Thanks!

 

[wifiquestion]

Might be a bit late to comment, but worth a try.
A GSM modem looks far and away the best bet.
Use a cheap, off the shelf unit, and send text messages (SMS)
about the state of your compactor.
You don't have a network... just lots of single items.