Wireless Launch Control

[flight4]

One challenge I have been puzzling over is how best to attach the top panel the field box proper. In the past I have used a few 1/2" pieces of aluminium angle stock to secure the top with some screws. Simple and it works but it's pretty obvious.

I'm not coming up with anything better (and simple) so that's probably what I will do here.

[alexzogh]

I've had a few panels made at frontpanelsexpress.com. On one of the panels I used their 'no screws' idea from their tips & tricks section. Basically you glue, solder, or weld a post onto the back of the panel, and then thread a screw from the backside of the enclosure to tighten down the front panel.

Ok, now when are you going to sell this to me :>

[flight4]

I hear you and that's a great idea. I don't mind the screws, but unlike with a console, here there is no access from below the panel to attach it from the bottom.

Maybe what I'll try is some wider pieces of light aluminium angle stock. Easier to position and attach than skinny brackets. And two screws holding them to the field box will keep them more stable. Then I can drill down into them blind from the top of the panel without having to be too precise, and attach the panel with a few sheet metal screws.

That might work. I have a defective tool box I can try it on.

[flight4]

Well this little break was longer than I planned. I had to stop to build a new computer. My 7 year old workhorse needed to be replaced so I ordered some parts and got it done. The technology out there today in terms of motherboards, CPUs, memory, and peripherals is just amazing. Building a computer from components used to be pretty challenging. Today, not as much.

[flight4]

Spent the last few weeks working on other projects, and also spent a week away on vacation. But now I can finally show some progress on finishing these boxes.

The problem I was working out was how to attach the top panel without brackets on top. The pieces of aluminium angle below the panel worked out well. These are attached with sheet metal screws and will not be visible below the panel.

I had to modify the top right piece of angle to make clearance for the Sonalert which is a big old thing.

[flight4]

A few screws on the outside don't take away from the appearance too much.

[flight4]

And this is the finished product. Well nearly so. I still need to do some labeling. What I'm thinking of is some custom vinyl lettering for Power, HB, pad numbers, Antenna, Beacon, etc. It should be pretty obvious but it will dress it up a lot.

That's an awfully big battery for most applications. A 6AH SLA type battery works really well too and is a lot lighter.

When the day is done the antenna just folds down inside the box, the battery wires tuck in, and it closes up neatly. The tool boxes stack too which is handy.

[flight4]

A few more pix. The second field box is finished (again, except for the labeling) and this system is working and ready to go.

The second picture shows everything ready to store or transport. Except for such extras as ignitor cables, batteries and launch pads. But you get the idea.

[flight4]

So what about antennas? Well the "right" answer is there is only the FCC approved 2.1dBi 4" articulated dipole available from Digi for $5. That's the short little one in the front.

I've tested this one consistently to over 500' across relatively even ground. And it even worked once to more than 1500' when I was on a small hill. That's with the 2.1dBi on both ends.

The middle antenna is a 5.5dBi Omni that I obtained domestically. Theoretically it has around twice the gain as the 2.1dBi. The same company sells a 9dBi Omni antenna that would be a marginal improvement again. I particularly like the 5.5dBi Omni because when it folds over it still fits inside the field box. The 9dBi would be too long.

The monster at the back is a 2.4gHz Yagi I ordered from Hong Kong. What I liked about it was that it already has a RP-SMA connector on the cable so it is plug-and-play with my system. This one is advertised as 16dBi. It is highly directional of course.

I have found in my experiments that getting the antenna up off the ground is much more important than the antenna itself. And this makes sense. A 4" full-wave dipole 6" off the ground will be greatly challenged compared to a similar antenna 5' off the ground.

Out to 100' the small antenna works fine. Out to 500' the small antenna works but it needs to be elevated. In these situations a higher gain antenna at the field box end improves reliability.

The Yagi is the wild card. Line of sight it will work to a mile and more. It's easy to put the Yagi about 5' up on a tripod next to the field box and point it at the LCO. At the LCO we just need the 2.1dBi or one of the longer dipoles. If the antennas can see each other they will work. So this is the solution out to 2000' and beyond.

So what about antennas? It depends (my favorite answer).

[flight4]

This is what the Yagi looks like on a tripod. Not much to it. Press a 1/4-20 socket into the end of some 1" square stock. Any tripod will do. Vertical polarization to match the dipole at the other end.

[flight4]

I really don't have much more to post on this topic. The project is finished and put away. For those following the thread I hope you got something from it. I enjoyed doing it.

Feel free to wave if you liked the thread.

I have enough small parts to build about 8 more of these. In various configurations. Except for the expensive parts I would still need to order like custom circuit boards and engraved top panels and such.

I plan to try building different 8, 16 and 32 pad configurations and see if I can sell any of them.

I know some on this thread have expressed interest. Send me a PM with your email address and we can chat.

I'll occasionally post updates here if I have anything interesting.

Cheers

[Dan_G]

Hi David, the system looks really great. Thanks for posting the pictures and writing up all the details of your design. Very nice work indeed !

Dan

[flight4]

Thanks. I enjoyed doing it.

[flight4]

Up to 16 wireless pads my software is ok, but I know going over that will require some new strategy. The latest version of the LCO master controller bumped the CPU clock from 20mhz to 40mhz. So this will give me a little more elbow room. But I don't have anything close to a 32 pad test platform. I should build a few more field boxes and start to get creative.

[flight4]

Field box-less Launching

A year or two ago a club member and I were talking about building a wireless sawhorse system that didn't need a field box or long launch cables. More suited for low or mid power sure.

So then it's how does one take what's in a field box (relays, radios, CPU, etc) and put it all on a 2x4 piece of sawhorse?

Make it all smaller I guess. That Might be my next adaptation.

I don't know if wireless for low power launching even makes sense. The premium over a simple 30' ethernet cable and a field box will be steep. But it will be cool. Wireless is always cool.

Which would you prefer?

[UhClem]

Up to 16 wireless pads my software is ok, but I know going over that will require some new strategy. The latest version of the LCO master controller bumped the CPU clock from 20mhz to 40mhz. So this will give me a little more elbow room. But I don't have anything close to a 32 pad test platform. I should build a few more field boxes and start to get creative.

For some reason I keep thinking about this but I still don't understand why a 5 to 10 MIPS processor couldn't handle any number of pads. I am planning on a controller for up to 253 pads using a MSP430 running at 4MHz and I can't think of any reason why that isn't sufficient. I actually plan for the processor to be asleep most of the time.

So what is the bottleneck in your design?

[flight4]

For some reason I keep thinking about this but I still don't understand why a 5 to 10 MIPS processor couldn't handle any number of pads. I am planning on a controller for up to 253 pads using a MSP430 running at 4MHz and I can't think of any reason why that isn't sufficient. I actually plan for the processor to be asleep most of the time.

So what is the bottleneck in your design?

It's not really a bottleneck, but rather a difficulty orchestrating a large number of concurrent activities. Everything from blinking LEDs at certain rates, beeping the sonalert at certain rates, keeping track of local and remote state, safety audits, communicating with the transceiver, i2c communication, debugging, all being done by the processor within various modules.

At the LCO master controller this started to get unwieldy and the timing became slightly unreliable. The controller uses an operating system that employs task switching, and I have about a half dozen tasks (processes) and a small number of semaphores. With a 20mhz processor the problem is not cycles, but task logic more likely. Some tasks that were supposed to operate near synchronous were not always in sync. In reality it isn't a big deal unless you are trying to have a multi field-box drag race with millisecond accuracy on the launch. Which is what I was doing.

But anyway I just want to clean that up. I also increased the clock rate at the LCO master controller to 40mhz for better headroom. With that I might even be able to get all the debug logging to work right.

Without wide synchronization drag racing from a single field box will still work great. But more importantly overal timing stability will be much better because the LCO master isn't stepping on its own toes now and then while juggling too many balls in the air.

[flight4]

I have enough parts on hand to build a few more of these systems. Various configurations. Even some of the costly parts like relays and switches.

I'll probably start during the summer sometime.

[flight4]

Here are the top and rear panels for the new project. Just arrived. 16 pad wireless. Range 15 to 2000 feet. 4 wireless field boxes to build (groan). Most of it exactly as before.

[flight4]

A good while back I said I would someday post photos of the earliest prototype MCU controller. Basically showing the internal wiring and components. These boxes haven't been opened in at least 3 years. This is the LCO controller for 8 pads. This is a wired system using cat 5 cable to the pads. Not wireless.

Lots of wires and connections. Lots of patchwork too. But it still works great.

[flight4]

And this is the original 8 pad field box. Again, wired not wireless. At first I thought 8 pads was a good idea. Especially for low and mid power. But the successor to this box was a collection of 4 pad field boxes. That turned out to be ideal from a flexibility perspective. Basically set up your low, mid and high power clusters of 4 in any groupings you like, or not. My wireless systems are the same. I can still get 8 into a box. But 4 is the sweet spot.

Oh, and if you compare the 3rd photo to my more recent field box PCB photos, the layout of 3 boards has basically been combined into the one board. There are some circuit changes of course, but the main thing is a whole lot of discrete wiring eliminated.

[cwbullet]

Very nice work. I have seen many systems and yours are a work of art.

[flight4]

This custom panel is for a future project. Most expensive part on the whole list. But well worth it for the quality and not to have to drill over 110 carefully aligned holes on the drill press.

There's a shorter lower panel too. But I made a mistake on it so it will have to be remade. Ouch!

[cwbullet]

Is it for a wired or wireless?

Chuck

[flight4]

LOL. Wireless, of course. The wired version of this particular panel layout has been in the field for about 3 years now. With wired range extenders we have had success out to around 250'. The theoretical range (with extenders) is greater, but we haven't had to try it. Greater range at lower cost was indeed a driver for developing wireless.

[flight4]

I've kept various hand drawn cheat sheets for where to put all the tiny little surface mount components on the circuit boards. And I lose them or the board changed. And I draw them again. This is over several years. Anyway I have a stable board design so time to start documenting some things. Tedious but very reusable.

This took a few days. Basically first print the board layouts. Then scan them back as images. Import to Visio. Then add all the annotations. The pdf is a little too big to upload to the forum but here's a link. Hopefully it works.

Launch Controller Component Layouts v2.1 Rev 0 20110804.pdf

Some boards have components front and back, some (thankfully) just on the front.

[flight4]

Starting my new project. Sixteen pad wireless. Start with four each x four pad panel modules.

It will make more sense later. No progress pictures this time. Just before and after.

[flight4]

There are actually two versions of this same board. The piece that moves around is the RS232 interface. Not used at all for wireless. But important for the wired version. With the top board the section is integrated. And it works ok for the 8 pad controller even though the section is not used. For higher density such as the 32 pad controller, the rs232 interface needs to be stacked on the pad module to save vertical and horizontal space. So on the lower board the section gets cut off and then stacked on the pad module. But only if wired is needed. The tab pointing down is for the master arm switch in a prototype configuration. Just a place to attach it. Have never used it though and I have just cut it off for every application.

[Blackfly]

Fantastic project, and a great build thread. If you ever start producing these, perhaps in kit form, there's a market, not just for the dozens of HPR clubs, but for the HUNDREDS of pyrotechnic techs in North America. Please introduce yourself and this project over at PyroUniverse.com, the boys(and girls) would love to hear from you!

Cheers,

Patrick T

[cwbullet]

David,

Any more progress on the build?