Electric Stunt > New electronic technology
What do you want in a timer?
Tim Wescott:
Suddenly I'm working on not one, but two timer projects.
One's a teeny tiny timer for Larry Renger's program to put planes in the hands of kids.
The other is a project to make a timer that -- with the exception of size (it's big) -- can be all things to all people. It's turning into a bit of a monster, but I think I'll get it done. I'm modest, so I'm calling it Tim's Universal Timer.
* What it definitely has:
* Four connections for input or output, pulse or logic
* A freaking insanely powerful processor for a mere timer (STM32F303)
* A USB port (I said it's turning into a monster!)
* The ability to talk to, and be programmed by, a JETIBox
* A 3-axis gyro, 3-axis compass, and a 3-axis accelerometer
* 2 megabytes of flash memory, for data-logging
* A blinking light, because no board with a processor should be without one*
* What it definitely will have:
* The ability to emulate a KR timer or a Hubin timer
* The ability to sequence landing gears along with motors
* The ability to control more than one motor in a Renicle-esque way
* The ability to sequence things other than motors, if that's what you want (consider a FF-style motor cutoff after six minutes)
* The ability to detect premature motor or IC engine stall, and extend the landing gear**
* The ability to detect one motor or IC engine cutting out prematurely, and cut out all the others
* The ability to record RPM and all motion of the aircraft into flash, and upload it to your computer after a flight, for review
* And more. Read on
* What it will probably have:
* The ability to emulate Igor Burger's timer
* The ability to regulate an IC engine speed***
The way that I'm achieving such flexibility, and a good part of the reason that this project has turned into a monster, is because in order to meet my minimum requirements, I needed to give the thing a programming language**** of its own. This fact means that you'll be able to take a variety of modules -- RPM detectors, PID controllers, timeouts, and the like, and combine them in ways that I have not yet imagined. Then you'll be able to download a program to the timer (remember the USB port?) and have it run the next time you turn it on.
So, given that you have an idea of what this critter can do -- what do you want in a timer? Forget any constraints -- just say what you'd like a timer to do, and ignore anyone (even me) that says "oh, that's not possible*****".
* Well, almost none. To quote a new customer of mine who's very power-consumption sensitive, and who will be potting his board in opaque epoxy: "What the hell is this light here for?"
** Or drop the landing gear immediately after a crash. I suggest no retracts if you're still at the crashing stage.
*** Whether there's any advantage to it or not. This kind of comes for free with the KR timer emulation.
**** Any time a software engineer for an embedded product waltzes into a design review and announces that he's working on a programming language for the product, the correct and rational response is for every other engineer in the room to turn to him as one, and say ARE YOU CRAZY!!!. Then, as follow-up, they each need to collar the guy one by one and say "OK, explain to me why you're not crazy."
***** Except for teaching you how to pick up women in bars. If I were born with the technical chops to make a gizmo that would help me pick up women in bars, I would still be single. Or quite happily dead from exhaustion.
John Cralley:
Can you make it the size of a Hubin FM-9 and under $10??? VD~ VD~ VD~
Tim Wescott:
--- Quote from: John Cralley on February 25, 2014, 03:40:29 PM ---Can you make it the size of a Hubin FM-9 and under $10??? VD~ VD~ VD~
--- End quote ---
The one I'm making for Larry has a target BOM cost of $5.00, which we may meet in quantities over 80. The only thing standing in the way of it being the size of a Hubin timer is someone willing to volunteer to solder it up in surface mount, rather than through-hole -- and it'll get cheaper, because a big part of the BOM cost is the bare board.
How are you with 0402 resistors and SOT-23-6 packages?
steven yampolsky:
Few ideas:
1) a mechanism to measure line tension. Imagine an active line tension control system that maintains constant line tension in all quatrants of the sphere.
2) a video camera or a plug for one. maybe several cameras. wireless video transmitter. Imagine being able to view the flight from multiple angles live on multiple monitors.
3) interface for wireless telemetry.Google Glass has made some big waves. Imagine being able to see telemetry in flight. It would be a boon for trimming.
4) bluetooth interface to program. Why use a programming box? I should be able to program it with my iPhone.
5) ability to define a multi-dimensional map for IC engines performance like in cars. the axis' would be fuel, air amount, barometric pressure, humidity.
6) bellcrank position or rate of deflection sensor. Imagine being able to predict engine power requirements as a maneuver begins.
7) relative altitude sensor. it would be great to know how high the model is off the ground. The training value of being able to see(wireless telemetry and google glass) where 5ft is tremendous.
Tim Wescott:
--- Quote from: Steven Yampolsky on February 25, 2014, 06:56:39 PM ---Few ideas:
1) a mechanism to measure line tension. Imagine an active line tension control system that maintains constant line tension in all quatrants of the sphere.
--- End quote ---
Anything that can generate a 1 to 2 millisecond pulse train can be registered, recorded, and maybe acted on. I'm not sure if there's value in active line tension control, but if someone makes the gizmo, it should take little or no systems work to make it affect motor speed.
(This is why I was asking -- stuff out of left field, yea!)
--- Quote from: Steven Yampolsky on February 25, 2014, 06:56:39 PM ---2) a video camera or a plug for one. maybe several cameras. wireless video transmitter. Imagine being able to view the flight from multiple angles live on multiple monitors.
3) interface for wireless telemetry.Google Glass has made some big waves. Imagine being able to see telemetry in flight. It would be a boon for trimming.
--- End quote ---
These can be bought for RC planes. Top-o-the-line RC receivers these days transmit telemetry back. So, buy one and shove in. Ditto, the in-flight video stuff is a solved problem, used by the quad copter guys the world around.
--- Quote from: Steven Yampolsky on February 25, 2014, 06:56:39 PM ---4) bluetooth interface to program. Why use a programming box? I should be able to program it with my iPhone.
--- End quote ---
With an appropriate USB cable you'll be able to talk to it with your iPhone. Or, at least you can talk to an Android phone -- I'm not sure what apps are out there for iSpend phones. Bluetooth is horrendously expensive for small production runs -- you either need to spend on the order of $100,000 on FCC compliance testing, or you buy a certified module that starts at about $45 a pop that takes up half a square inch of board space.
USB, on the other hand, takes up about 1/10th inch of board space on one side (leaving the other side free for other things), and less than $2.00 worth of parts in small quantities.
This board uses an FTDI FT230X USB to serial chip and will use a human-readable serial data stream; search on iPhone terminal apps that talk to that chip, and you've found an app to talk to the board.
--- Quote from: Steven Yampolsky on February 25, 2014, 06:56:39 PM ---5) ability to define a multi-dimensional map for IC engines performance like in cars. the axis' would be fuel, air amount, barometric pressure, humidity.
--- End quote ---
Why? I'm not challenging -- I'm curious.
--- Quote from: Steven Yampolsky on February 25, 2014, 06:56:39 PM ---6) bellcrank position or rate of deflection sensor. Imagine being able to predict engine power requirements as a maneuver begins.
--- End quote ---
* Cheap: gut a servo, leaving the output shaft and pot. Replace the board and motor with a board that generates a 1-2ms pulse train (see my answer to your comment #1, above)
* Spendy: Build a bellcrank sensor circuit ala Kim Doherty (http://stunthanger.com/smf/index.php?topic=32528.0) with that 1-2ms output
You don't have to stop there -- with a gyro on the board, it's conceivable that you could move the CG back and play with eletronic stability enhancements. I have no freaking clue if the plane would fly better as a result, but you could do it.
--- Quote from: Steven Yampolsky on February 25, 2014, 06:56:39 PM ---7) relative altitude sensor. it would be great to know how high the model is off the ground. The training value of being able to see(wireless telemetry and google glass) where 5ft is tremendous.
--- End quote ---
Maybe next year. In theory, with gyro and accelerometer information you could deduce this. In reality (thank you, Howard Rush, for bursting my bubble) the available chip-scale sensors aren't up to the task, and a sensor suite that could do the job would probably weigh a pound.
You may be able to use sonic ranging over pavement, I dunno. I don't think radar altimeters work well at five feet.
Navigation
[0] Message Index
[#] Next page
Go to full version