Stuffed my project up!!

[spec]

Take up Chinese https://en.wikipedia.org/wiki/Written_Chinese

While I am fascinated by all things Chinese: people, country, food, art, I am radically opposed to the pictographic alphabet. While it looks attractive, because of its irrational and non expansive nature, it is very difficult to learn and thus is a hindrance to the Chinese people, and others who would like to learn their language. For example you hear of Chinese typewriters with 4, 000, or more, keys. To a lesser degree Roman numerals are also a hindrance.

Hey, I can write script forward, backward, and mirror image forward and backwards. The weird part, it effortless, if you don't think about it. All I have to do is start correctly. With practice, I make fewer mistakes. I got bored in glass one day. So, I could write "Happy Holidays" on the inside of a window with spray snow and it will turn out correctly when read from the outside.

That is amazing and something I just could not imagine being able to do. It is a natural born talent you have and I would say nothing more

I've been told I have some characteristics of Asberger's Syndrome, but not enough for a dx. Things that are important to me are not important to others: e.g. Shoveling a path across the 12" of snow on the grass for the mailman or shoveling snow in the curb for drainage reasons. Why? I've seen what can happen under the right conditions. The storm drains handle melting snow too. Reasons seem to be wrong to other people. e.g. The boss wants to make an impression by checking off things that are done. I, on the other hand, want to have want to have the next unstarted phase of the project started. The stuff that I know will work without complications, I leave to last. In reality, I should have set the boss take the fall for any problems.

Putting labels on people is what psychologists do, but I am afraid they do it too freely without taking the trouble to find out what is going on. Their job has no feed back either. If a surgeon cuts off a wrong leg or makes a pigs ear of an operation it is woefully apparent. A Psychologist, on the other hand, can spew forth the biggest load of rubbish and no one will check or know. They are also far to free with dishing out drugs, some with horrific side effects.

Getting back to your experience and thoughts about what is important and how to solve problems. There is something very strange in this world. Time and time again the views of someone who knows the score and has enough experience and foresight to mitigate future problems is ignored in favor of someone who does not have a clue. I have never got to the bottom of this, but it is very common.

I feel a story coming on:

We had a contract to design, develop, and build a helicopter radiometer. Being airborne, power, size, and weight were extremity critical. One of the fundamental controlling factors of the performance was the stability of the master clock. Basically, the better the clock the better the resolution was for a given antenna size. So initially, I focused on getting the clock as good as was practical. After some investigations it was clear that oven controlled xtal oscillators had an order better performance than other types.

At the first technical review meeting I described my proposed hardware design, including the oven controlled Xtal. At the first mention of oven there was an uproar from the project manager. 'You are not using an oven' and that was that. I could get no further discussion on the subject. After the meeting I went to see the project manager to find out what was going on. He was most aggressive and just repeated that I was not going to use an oven. About a week later a large project needed a problem sorting so I got transferred to that as it had a higher priority.

I still kept an eye on the goings on at Helicopter Radiometer and knew the engineer who had taken over from me. He had not used an oven controlled Xtal, but instead used a normal type, with a load of compensation circuits to try and stabilize the master clock over the military temperature range. I learned that the program manager was spreading the word that I was an idiot specifying an oven on a job where power and size were a premium. He also said that my digital approach to the design was not right and would never have worked.

The truth of the mater was that the program manager was a physicist and to him an oven was a big power hungry thing with a door on the front that you cooked samples in. He knew nothing about electronics and was a dyed in the wool analogue advocate, because at least he could understand that. The oven controlled Xtal that i proposed consumed IW maximum at low temperatures. It measured about 10 x 10 x 8mm and it cost, in modern value, £500. It was also off the shelf. Another factor controlling the ultimate performance of the radiometer was drift. A digital approach not only allowed drift to be largely cancelled but, by using CMOS logic, power was much less and the output from the radiometer was digital anyway. I never did clear the innuendo in that department, even after the offending program manager left the company.

As to the Helicopter Radiometer: it stumbled from one problem to another and clock stability and drift compromised the performance badly. So instead of having a tool which would have been leading-edge, the customer got a troublesome run-of-the mill performer. We never got another radiometer job after that.

 

[spec]

You still do that here. No biggie.

I know

Dyslexic has to be weird. I was one of those left-hand babies forced to be right-handed. The real term is what is your dominant hand? You should learn early to use both hands equally without favoritism. I use whatever hand is convenient, if a left handed-person is eating next to me, I'll eat left-handed. What's the big deal. just no elbow bumping.

Weird- true. If you compare my wife and I, we are complete opposites. She can spell and writes naturally, although I would say not succinctly. Me, I can't spell to save my life, but I can visulise things, in 3D too, and my spacial appreciation is good. My wife, on the other hand, has none of those skills.

Left handedness is another common theme with engineering types. I too was naturally left-handed and after a year in primary school was treated like a leper until I changed to right handed. It was a big thing at the time and put me in the low achievers category: no spelling, no writing, poor reading, quiet. But if some situation need to be sorted or a bit of logic was required I was the top dog.

As it turned out, I am very glad that they did force me to be right handed and now I am naturally so.

 

[KeepItSimpleStupid]

A Neurologist I had, said he could visualize every word from a text book. It's extremely fascinating that people have entirely different skills. I doubt you can be taught them?
e.g.
Visualizing 101
Visualizing in color 102
Visualizing in 3D 201
Math your head 101
Recording sounds in your head 101
Recording speech in your head 201
Game time 101 - Play chess in your head with no board with an opponent
The spacial evaluation - visual, verbal and Kinesthetic scores. (Test required at kindergarten) and every 2 years thereafter.

 

[spec]

A Neurologist I had, said he could visualize every word from a text book. It's extremely fascinating that people have entirely different skills. I doubt you can be taught them?
e.g.
Visualizing 101
Visualizing in color 102
Visualizing in 3D 201
Math your head 101
Recording sounds in your head 101
Recording speech in your head 201
Game time 101 - Play chess in your head with no board with an opponent
The spacial evaluation - visual, verbal and Kinesthetic scores. (Test required at kindergarten) and every 2 years thereafter.

Keep,
What do the numbers after each aspect signify?

 

[KeepItSimpleStupid]

Course numbers. e.g. We had M241, M242 and M243 for the 3 Calculus courses. CS105 was an intro Computer Science course.
So, I picked course titles and numbers. 101 is a typical intro course # for Freshman in college.

e.g. https://www.google.com/url?sa=t&rct...-1.htm&usg=AFQjCNG4a9C_7Zbs-zHrfsYol463flTkxw

 

[large_ghostman]

Demo went like a charm, power wise it was close! I had around 2% battery left or around 20 seconds full blast no gizmo on.
Points...................

Only got a rough idea as we wont get the scores until Monday, but apparently I am in the "90 something percent range" I thought my machine would be back home today, but apparently they are keeping hold of them for a week (no idea why, probaly to play with them lol).

I will tell you what I came up with later, sounds boring at first but took alot of work and when you see it in action then it becomes clear why it is pretty good .
Was a couple of others not done any pre testing, the switched on and.................NOTHING, that is the sort of thing keeps me up at night!

There was a quad copter that hadnt been given its medication and totally lost the plot, went completely woppy before committing suicide into the wall. Quick but messy death!
Was pretty spectacular to watch but I decided against congratulating the owner on providing the best self destruction of a non controlled quad copter I had ever seen!

Someone made an ant (huge ant) that got lost 2 meters into the room and just stood there until the battery died.

The doorbell girl................. Great bumpy bits but man I am never sure if she even visits our planet! I have no idea if she even read the brief or not, but considering her short skirt and long legs I decided to go talk to her and be friendly . Was a struggle being impressed with a wired door bell when all projects in the group were meant to be wireless. She insisted the sound was wireless and the cable was just for the button and power.............. Did I mention she has great legs and bumpy bits? She likes making stuff but not sure I could handle her in my work shop. Apparently she got her dad to solder the project as she dosnt like the smell of solder! I did tell her to say nothing to the marking team about doing that.

Not sure she got my point about it having to be all her own work, but I did mention I dont mind giving her soldering lessons (her workshop not mine!!).

 

[cowboybob]

So, I could write "Happy Holidays" on the inside of a window with spray snow and it will turn out correctly when read from the outside.

First time at sea I was assigned the "Status" board in CIC (Combat Information Center) on my ship. The guy behind the big clear plastic board marking ship/target positions with a white crayon like marker. In reverse text.

Somehow, from the git-go, I fell right into the task with no mistakes. Didn't think about it but later discovered my dyslexia had, probably, made it possible.

More surprised that the Navy didn't even bother to test me for that "General Quarters" assignment. Just said, "Here. Do this.". Gotta love the NAV...

 

[throbscottle]

spec - you should have told him "thermally stabilised". Same thing (well, arguably), different words...

LG - what, you mean she has great knobbly knees? (I have grand-kids around your age - great fun winding them up about bumpy bits)

I'm left handed myself (and dyslexic), so mirror writing came fairly naturally in my teens, but then I worked out a way of turning my writing around so even in the mirror it was still backwards. Don't know what you call that - I think I might have done it vertically mirrored instead of laterally.

 

[large_ghostman]

To be honest I have no idea about her knees, my eyes didnt get as low as that , I guess they were distracted on the way down

 

[spec]

spec - you should have told him "thermally stabilised". Same thing (well, arguably), different words...

I just quoted the standard term that was used for that class of Xtal oscillator. I may not have remembered the name exactly right, but it certainly had oven in it. The point is though it should not have mattered what I called the oscillator. If the program manager had any doubts about my design approach, and baring in mind that he knew nothing about electronics, he should have discussed the situation and established the facts instead of reacting in a thoroughly un professional and obstructive way. That particular department was staffed by managers who lived off past glories and had not been able to/ could not be bothered to keep up with technology.

I'm left handed myself (and dyslexic), so mirror writing came fairly naturally in my teens, but then I worked out a way of turning my writing around so even in the mirror it was still backwards. Don't know what you call that - I think I might have done it vertically mirrored instead of laterally.

Interesting- another engineer with the left handed dyslexic thing. I have met many the same in real life. One of the kids at our school was a whiz at metal work including all the theory, but he could hardly write and reading, apart from magazines about cars, motorbikes and mechanics, was a real problem.

When he left school at 16 he opened his own garage which was a great success. He was rolling in loot, while the rest of us were living on a pittance while still under training. He then made a few property deals and finally got a high paid job as technical manager of a concrete products company. He retired at about 55 and seems to be having a grand time. He remembers the school headmaster's parting comment that he would never amount to anything. His buddy in crime who got the same comment from the headmaster, has a successful clothing business and lives on a huge dude farm in an upmarket area north of Bristol UK.

 

[throbscottle]

I preferred pics as it seemed you could get more done although there was a steeper learning curve. But pics seem to have stood still, I have used ARM and other micros and all including the arduino have better easier tools and larger libraries, also one of the things I think really holds back the pic today is attitude.

I had the same experience when I started using Slackware Linux. Very much a hard-nosed attitude on the forum (Usenet back then). RTFM (and the rest) is alright if there is a manual, if you can find it, and if there is one it's comprehensible. I was a fairly experienced Linux user by then and was always trying new things, I loved Slackware but the so called "community" put me off (and now I finally settled on Arch. Looks like I won't be changing again).

 

[spec]

I had the same experience when I started using Slackware Linux. Very much a hard-nosed attitude on the forum (Usenet back then). RTFM (and the rest) is alright if there is a manual, if you can find it, and if there is one it's comprehensible. I was a fairly experienced Linux user by then and was always trying new things, I loved Slackware but the so called "community" put me off (and now I finally settled on Arch. Looks like I won't be changing again).

Why are people so awkward/rude. I just don't get it
My experience, for what it is worth, is that people who behave like that are hiding a secret- they don't know that much.

 

[large_ghostman]

B+ for write up, A* for practical 91% for demo, overall score A+. , as always the write up didnt do me any favors, because I am doing next years exam this year I have until May to do the MKII version. Thats really tight time wise so I need to have a think and see if I should do a different project.

Its also been entered into a national competition.

 

[DerStrom8]

B+ for write up, A* for practical 91% for demo, overall score A+. , as always the write up didnt do me any favors, because I am doing next years exam this year I have until May to do the MKII version. Thats really tight time wise so I need to have a think and see if I should do a different project.

Its also been entered into a national competition.

Congratulations LG, great work!

Your dad would be proud

 

[throbscottle]

Yay - congratulations, well done
Now do we get to find out what it is????

 

[cowboybob]

Atta boy, LG!! Knew you could (would) do it. Did anybody get any video of the demo??

 

[large_ghostman]

There is Video!! I dont have it yet as the school took it, but I dont see why I cant get hold of my segment.

Ok I can explain some of it, until I decided if I am going to do a MKII for school you cant offer advice or upgrades! I have upgrades in the pipeline and these were left out on purpose so I had something to do for the next exam without starting from scratch (although I might anyway lol).

So it was a kind of Hexacopter, it has 6 motors but only 3 are 'NORMALLY' on at any one time, it kind of cycles through the motors depending on what its doing, it isnt remote controlled as such but does send data back to the user including video and a multi sonar picture (uses those sounder things you can buy like car bumper ones) the sounders are mounted on servos so it can 'feel' its way around.

The other feature (DO NOT RECOMMEND UPGRADE OR MENTION THE OBVIOUS)is the pi had a camera and facial recognition software, at the moment it looks for a face (any face).
Its designed to self guide itself through a single story building, it looks for doors and windows (MATLAB and the echo things can tell the difference between door and window).
Basically its for flying into a burning building looking for a person if one has been reported as missing, it send back video feed and a kind of floor plan as it maps it.

So it will enter a building and map the doors etc it can 'see' from its current position, it also looks for a face. If it finds a human face it reports back the location using sil labs radio transceiver chips (all back to user comms via this, I will link to these chips in a bit). If it dosnt find the person in the room it guides its way to the nearest door and repeats for that room (I had 4 rooms in total (mocked up rooms in the hall). The main motors use micro steppers (less than 8mm) to alter the main motor position so it can fly at angels for short periods (almost on its side).
Lets say there was a something blocking the door it can go through gaps by itself (self guided), I also hacked a cheap laser temperature gun to take temp readings of the room etc. I found that really hard at first then I discovered a neat trick by accident, I could get a signal I could read off the board, I could trigger the temp gun but couldnt find a way to read the data. It seems like some kind of odd encoding (I tried Manchester etc no luck), then while looking at the signals from the screen back to the board I discovered the LCD is SPI . So to read the temp I trigger the gun and just read the spi line that went back to the LCD .

There were a couple of other things it does and does well but rather than tell you and possiably mess up being able to use it for second exam, I will get the video and you can see from that the other bits I put on.

It flies fairly slowly because it has alot to do and being self guided etc I decided to make sure it didnt fly too fast. The whole thing was run with 2 x Arduino uno's 1 x R PI
1 x sil labs TX/RX chip, 6 X 18f1330's.

I know thats alot of chips etc but because of the time constraints it was easier to build it modular so I could have something that worked even if all the features were not there. It started as a quad copter but 4 motors didnt me the precision and moveabilty I could get with 6. But 6 uses ALOT of juice and I couldnt afford the mega expensive batteries :O, so I settle for it to mostly run with 3 powered at a time, the on board Gyro and Accelerator (still cant get those working spot on) with a MATLAB routine take care of working out which motor to power and when, because it runs mostly with 3 that constantly change you always have at least 1 corner thats dropping, this took a huge amount of the PI's time calculating, the PI then literally just sent simple commands to each the pics saying on or off, if the pic (and hence motor) is told on it asks one the Arduinos for the speed information etc over the I2C bus.

The pics also had some ability to read the sensors so it could calculate PWM etc.

Thats a very very brief outline, there is alot more to it and the upgrade will be alot better. I also used a Sil Labs Giant Gecko chip with 1 meg on chip memory and CORTEX M3 core, the next one will use the Wonder Gecko instead which has less memory but a cortex M4 Core and floating point maths etc on chip. I also want to move away from the Arduinos and pre built libs, I felt in many ways they held it back a bit but did make the development very fast. The PI was essential for the camera and sensors etc and most the hard bits were done with MATLAB, to speed things up like moving room to room the PI can switch from scan/video mode to processing mode, this allowed me to map the room send back the data then switch to move around obsticles and through doors mode.
It works and works well but is 'Clunky' in places, one of the motors has burnt out (it happened right at the end flying back to base). I dont know why yet but there is several little bugs I need to iron out. It dosnt recognize who the person is it just looks for any face. I need to change this to something else, it also has IR sensors on for doorways etc but they didnt do work well, you cant tell by looking at it but it logged a huge amount of data so I can examine exactly what it did and when it did it.

I need to learn XC8 and mplabX, code in general. Many of the things I wanted to do I couldnt and had to use inbuilt libs alot which in places holds it back. Also I would of like more sensors and maybe I will add them once I can afford one or two of the arduino sensor kits (I think is cheapest way), I also want GPS module and maybe bluetooth or wifi for it.
Cost stops me at the moment and I need to think a bit more about it, I am not sure how useful GPS is on the scale we were testing. The wireless chips I used are excellent and huge range, but they were streaming back huge amounts of data, I might use another of the same so I can send more commands back or maybe have a 'Drive' mode where you can fly it.

The idea though was to enter a smoke filled building and act as a scout while the fireman get ready to enter or whatever. was alot of fun and took alot of time to build and design.
You can ask questions but I might not be able to answer all of them, the last thing I want is to be disqualified from using it for next exam if I want. I havnt seen the video yet but it flies better than my description, I would love to add some fire and gas sensors etc.

Loads of ideas but I thought I had another 15 months or so to do it in, as I have been offered the chance to take the next exam a year early I will only have a few months. Money and time will limit what I can do, I wont be buying the mega expensive batteries as I dont see them being worth it. The motors were fantastic and were donated from sil labs power demo boards, they are really powerful and I might take a vid of one the demo boards working. I had 4 of those motors and 2 that were slightly different model, it was a nightmare having 2 motors that were different. I will stick with a micro for each motor, it gives me alot more options and makes code easier, it also meant the PI etc wasnt struggling too much speed wise. If it was a real device then it wouldnt need to stream back all the data I had to for a development model.

I need a major rethink on the H bridges as the mosfets got hot and were poorly driven by the transistor mash up I did. Again cost cutting meant I didnt use drivers, but I would still use my own H bridges any way.

The next model might be more for going into a chemical fire or whatever and sensing the environment to feed back data before fireman enter. When I first started to design it there were no self guiding self deciding quads/hexapod around, then a few weeks before the exam there was one shown ob CLICK the BBC techy program. they have won an award for theres but I dont think its anywhere as good as mine .

Total build cost was around £370 all in, but most of it was done in small bits, I also saved alot by having the motors and batteries used donated. I also made the airframe and this is going to get a major rebuild! When you see the vid you will get a better ide, the motors on mine can alter there angel and the copter arms can move a bit by using the tiny stepper motors, I built the driver boards for these and have burnt some out!

I will need to replace them but they are reasonable cheap on ebay and really tiny. Much of the motor burn outs were due I think to poorly done drivers or bridges, and the arm motors that move the arm were down to code errors! Things like not checking stall current or because I cheated power wise, I used 12V power rail for 5V motors but used PWM to manage it, in a couple of places looking at the code the motors got the wrong pwm signal and therefore over voltage/Current.

Mainly I want to learn much more programming, I havnt programmed pics in a very very long time, and havnt used the new pic tools before, so stuff like that needs looking at so i can do the MKII.

Talking of code!

RANT

MC have dont support the 28f in there SD card lib now! or with the new compiler, plus the lib they do have is a complete nightmare! much better easier libs are around for the 8051 and 32 bit Sil lab chips I used, and the Arduino make using SD cards dead easy!! Why cant MC do libs that dont require a PhD just to write some stuff to a file for logging? in the end I used another UNO BUT didnt get to use the SD shield as I over volted it!! before the demo , I will have to see about replacing it.
I sold my Tektronix analogue scope to pay for most of the build, I might see what else I have that I dont use and can sell to get the MKII moving. I have a signal generator or I might sell my camera as I have use of a different one.

I dont have the time this time to get bits in dribs and drabs, and to be honest I can live without the camera more than I can live without a decent degree . I have some other bits that I dont use and likely wont use again so little point keeping them. as a new seller the pay policy of holding the money for 21 days when you sell something is a nightmare!! Robbing gits.

I have some power rail questions to ask but will post in another thread.
Sorry its a long post but it was a massive project . Oh and why does Magnesium Alloy used in engine cylinder heads STINK really bad when machined? Its an awful smell

 

[KeepItSimpleStupid]

There isn't a Wow! rating, otherwise you would have got it. Magnesium will burn.

 

[large_ghostman]

There isn't a Wow! rating, otherwise you would have got it. Magnesium will burn.

It is what I had to hand and to be fair its covered in the poor man's carbon fiber stuff they use at school, not sure what it is (not real carbon fiber for sure) and its not fiber glass, but its some composite.

It's not all Mag alloy but parts are as it was what was to hand, had it been something I was seriously developing for manufacture or whatever at a later date, then i would have gone the carbon fiber route. The actually copter I built is unlikely to get near anything much hotter than a cup of coffee , but it seems more robust compared to aluminum but a little more brittle.

I havnt explained it very well above, to really get how cool it turned out you need to see a video of it in action, its the first project I have done that has actually impressed me . The anti collision stuff is really good, but would need alot more work to do it seriously. The biggest innovation from my perspective (apart from the 3 motor control system) is the non fixed angle for the main motors, with decent fast gyro's and accelerators, the quad/hex copters are far more stable and alot more nimble, they can also virtually fly on there side.

Dont get me wrong it works, but its a long way from polished. I knew early on due to time and money I would have to use alot more chips than a polished product would, but it did give me speed of development and also meant if I had come up against something I could have had the project working but without a few features.

It seems more complicated than it was, Matlab/simulink for the PI and arduino took most the pain out of the complicated stuff, the fast heavy processing was done with Giant gecko dev kits sil labs gave me a while back when I was reviewing boards for them. I will grab some links of the boards I used at some point. If I had been forced to code it all on a single system or without tools like Matlab I wouldnt have been able to do it, but tools like that make it easy.

Again this is what annoys me about microchip, it took longer to mess with motor controllers and pics than the entire face recognition and avoidance system!! Once I have done the next exam I will release all the code and schematics etc, it might make a good starting point someone. I nearly used Matlab/simulink for the pics, then I found MC was charging nearly £1000 for something alot of other companies give away for free! The micro controllers with built in transceivers are something else, 32 bit processors and a range out in the open or nearly 6Km.

Would be less in a town or in and around buildings, but not much that much. Not going in today as its Tuesday which has become home study day for me , but emailed the teacher and I will be able to publish my video segment once they have separated them all and cleaned them up, I think they are being put on the school blog so I will link to it once they put it up. I have also asked what details I can give etc and waiting for an answer.

My favorite bit by far in the motor control system, that has been all my own work and makes quads/hexapods much better to fly, especially the moving arms and motors.

Now I have it home it isnt flying GRRRrr, looks like the battery took a pounding and mainly the motor H bridges and control circuits I built were not great, I will use spice for the next ones

 

[throbscottle]

I am in awe of your awsomeness, you awsome youth, you! Well done, have pat on the back and a glass of whatever makes my piddling around with things look like child's play!