From: John Edgecombe
Subject: [Paparazzi-devel] Hello
Date: Mon, 4 Aug 2003 17:44:49 -0700
Greetings:
My name is John Edgecombe. I don't have a nickname, so I just use John. I am
a retired Land Surveyor, a one time private pilot (real airplanes), and an
Amateur Radio Operator. About 15 years ago, I did a bit of programming of
trig functions in assembly language and Forth on a 6502 (Apple 2) and a
Z-180 (circuit cellar SB-180). Alas, my old machines were too crippled to be
effective. I did learn a lot before I had to give up. So, while my matrix
algebra is over 30 years old, I think it will come back to me, at least
enough to understand what is going on. I also have an interest in navigation
and geodesy.
The world has changed a lot. I notice that the AtMega has a mul instruction
and is another eight bit processor. In that other world out there, the
Pentium is king. Somehow, I never got used to Intel's way of doing things.
My skills are very rusty, and my pocket book is not flush.
I have been interested in helicopters since the mid 1980's when I bought a
kit. I did have a problem with my wife over that, and a very big concern
about the main rotor blades. This kit had wooden blades with a nylon blade
holder which was supposed to be glued together and this assembly was to
pivot on a 4mm bolt. My analysis at the time told me that without the nylon
blade holder to spread the load, the bolt would shear out of the wooden
blade. This meant that my personal safety depended on the glue joint between
the nylon blade holder and the wooden blade. I just didn't have enough
confidence in glueing nylon to wood at the time to finish the machine. Now,
there are commercially made blades with integral blade holders that
shouldn't have this problem. However, there are still a lot of unknowns
about the radio system. The range of the radio and the ability to determine
orientation visually are just two of the factors that bite the uninitiated.
Somehow, airplanes are a little more forgiving than helicopters, so I plan
on having both. My current project is a Sig LT-25 to be built as an electric
motor drive.
An autopilot, whether fixed wing or rotary wing, needs a place to be
installed. For the moment, my progress has stopped until I can build enough
of the autopilot's components to determine size, and balance in the
airframe. While it is nice to sketch out a little space and label it for
autopilot use, I have learned the hard way that sizes and weights are
tricky. While it is inefficient, I find it is necessary to rotate through
the various facets of the work to be done in order not to make commitments
which are incompatable with the rest of the project. Decisions have a way of
propagating consequences that require changes almost everywhere.
Am I fully committed to autopilot or paparazzi? Not yet. That is why I
joined this list. As I become more familiar with each project, I will make
that decision. Both projects now use essentially the same inertial units,
and paparazzi has ported autopilot software. The big difference seems to be
in the control boards, with paparazzi using an ATmega 128 and an ATmega 8,
while the autopilot control board uses an ATmega 163 which Atmel has phased
out. The version 3 board is still in the future. I have done just enough
hardware to be uncomfortable with it. There is a real art to picking
compatable parts and getting them to work together without glitches. Board
design is another area where prototyping costs can eat you alive if things
don't work as planned. I can solder most anything except SMD and flat packs.
I have never tried reflow, and I don't want to invest in those tools.
While I have installed the freeware version of Eagle, I cannot look at any
of the paparazzi boards except the main board schematic, and even that one
produces an illegal data error in Eagle. I suspect that the boards have gone
beyond 2 layers and are no longer viewable in the freeware version.
Annonymous said, "The code will expand to fill the available space." So, at
the moment, I favor the bigger is better approach of paparazzi, but building
one is not as simple as the autopilot kit.
John
From: antoine . drouin
Subject: Re: [Paparazzi-devel] Hello
Date: Tue, 5 Aug 2003 14:17:01 +0200
User-agent: Internet Messaging Program (IMP) 3.2.1
Hi folks
Quoting John Edgecombe :
> Both projects now use essentially the same inertial units,
> and paparazzi has ported autopilot software. The big difference seems to be
> in the control boards, with paparazzi using an ATmega 128 and an ATmega 8,
> while the autopilot control board uses an ATmega 163 which Atmel has phased
> out.
I think ATmega 16 is pin for pin compatible and can be used as replacement.
> The version 3 board is still in the future. I have done just enough
> hardware to be uncomfortable with it. There is a real art to picking
> compatable parts and getting them to work together without glitches. Board
> design is another area where prototyping costs can eat you alive if things
> don't work as planned. I can solder most anything except SMD and flat packs.
> I have never tried reflow, and I don't want to invest in those tools.
The paparazzi board is all SMDs. I soldered it with a normal iron using the
'put_a_lot_of_solder/remove_excess_with_braid" method. The PCBs are
"kitchenmade" (no metalized hole). I'll put a photo on the paparazzi page.
> While I have installed the freeware version of Eagle, I cannot look at any
> of the paparazzi boards except the main board schematic, and even that one
> produces an illegal data error in Eagle. I suspect that the boards have gone
> beyond 2 layers and are no longer viewable in the freeware version.
The paparazzi board has 2 layers. I designed it with the freeware version of
eagle. It is strange that you cannot open it: which version of eagle are you
using ? I am using 4.09-linux
The board has a few more componnents: a switching 5V regulator, a linear 3.3v
regulator, a flat cable connector for the ublox SAM GPS, a cmx469
modem. It doesn't have the frequency to voltage converter for the tacho.
That said, there is only one prototype of the paparazzi board and it only has
completed two flights. There are also numerous "details" that i want to change
in future revision (use fewer crystal- the current board has 3 - integrate IMU
in controller board - he current connector is as big as a gyro..- move the
connectors, etc...). I want to make a new revision by october.
If you want to build a Parazzi board, go on, I'll be pleased to get feedback,
but be warned that it is very new and barely tested.
If you don't need the extra components, or the extra cpu, i think it's way safer
to go with an autopilot board.
> Annonymous said, "The code will expand to fill the available space." So, at
> the moment, I favor the bigger is better approach of paparazzi, but building
> one is not as simple as the autopilot kit.
We have alot of space and processing power available in the 16MHz-mega128.
We want to use it to try to implement basic navigation.
Regards
Antoine
Copyright 2003 http://lists.nongnu.org/mailman/listinfo/paparazzi-devel