stagenation

[Raj]

What would you do, if you had a project, you knew would work but you also know that you can't complete it?
I am in such a situation. I designed the hardware but i'm sure I can't finish the software. The project is basically a test & measurement device to test capacitors, leds, resistors ,diodes and inductors.
I don't mind opensourcing it.
Also a object hardness tester.

[SparkyFX]

If making it so far taught you something, then it was a good way of spending the time. Was it intended to be a marketable product in the first place?
What´s particulary hard on the software side? What would need to be changed to change that?

[Raj]

I wanted to open source it anyways. But only after completing it. Now it seems I can't complete it. There is a lot of maths and timing problems that I can't do. Only the led and diode and resistor testing part works but that too needs user to calculate stuff on a sheet of paper.

Now i'm like. Here's the schematic, here's the idea. complete it and it'll be yours to reap benefit from.
I don't even know if someone has already done it.

So here's the concept-
it's like a curve tracer.
a,b,c are the pins of a microcontroller,like attiny85




pin a,b alternate between 5v,0v and 0v,5v for all components (except for polorized capacitor) respectively at certain speed

the arrows are probes.

the characteristics of the device are determined by making multiple readings at pin c, thus essentially turning the resistor at pin b, into a current shunt.

a little math is used to see what kind and value of component will consume current in manner the device is consuming.

the result is sent to another device like an oled, using the remaining pins of microcontroller.



The hardness tester concept is as follows-
We have a tube,transparent to IR.
to it,we affix 2 ir source and receiver that will send signal when ball passes in-front of them,
We place it on an object who's hardness we want to test.
We drop a bearing ball into it.
We record, at what speed it was, just before it hit the surface.
after it hits, and bounces back up, we record the speed it went upwards.

Some brainwrecking physics later, we get to know the hardness of the object.

[Rick Law]

What would you do, if you had a project, you knew would work but you also know that you can't complete it?
I am in such a situation. I designed the hardware but i'm sure I can't finish the software. The project is basically a test & measurement device to test capacitors, leds, resistors ,diodes and inductors.
I don't mind opensourcing it.
Also a object hardness tester.

How about skeleton code?  You don't need polished software (like a great UI or something like that).  You need at least skeleton code to show it is capable of working.

If you don't do the software along side, how do you know your system really works?  Yeah, you can think it works in theory, but until you do make it work, it stays as "work in theory but I really don't know."  The only way to know is to have your software functioning as well.

Then, even if it is a product, you can sell it as end-users programmable and your skeleton code as working example or for basic operations.

[coppercone2]

instead of skeleton code you need to make a clear block diagram
then you can figure out if your CPU is fast enough by identifying bottlenecks and estimating time it takes to complete loops or complex code executions like sorting in filters and such

then you can work on skeleton code.

IMO that is more useful. I declined projects just because there was no comments or block diagram attributed to them.

also you can do things like add pictures to your block diagram using ellipsis and such } { to identify what that code loop is doing in regards to the process being observed or controlled. It will make it much more pleasant to work with then a run of the mill block diagram filled with boring feedback loops.

If you wanna hand it off to someone I would suggest a nice fun looking PDF with some pictures of the process and stuff. I would be much more helpful if people presented their projects in that way, or at least were throughly descriptive ,rather then me having to reinvent the wheel a whole bunch of times because someone was too lazy to make a tight document.

Also some analog blocks and expected signals coming out of them would help allow the programmer to easily see what kind of interrupts and stuff are required and the expected complexity compared to other stuff they worked on.

[Raj]

ok. ill get on it then

[bloguetronica]

Well, I can say I've been in that situation after designing my first function generator. It is a USB controlled function generator, hence the essential software part. I hit a roadblock when I had to thoroughly study the libusb API plus the CP2130 application notes. That, plus programming, was a lot of work for a single person, but it was totally worth it.

In your case, you've designed the hardware and you are in better position to design the software for it. My advice is to continue with your work. The reward feeling will totally worth it in the end.

Kind regards, Samuel Lourenço

[Raj]

Ok.
Also I think I have a solution now (got the idea from you).Instead of making it all microcontroller based, how about ,making a computer process the data and the microcontroller is there only for sensing stuff.

[coppercone2]

I feel like this thread is similar to trying to make a dog take medication. It went from a adulterated cupcake to trying to shove it up his ass. 

now you need a driver and computer link, its even worse IMO

[bloguetronica]

Ok.
Also I think I have a solution now (got the idea from you).Instead of making it all microcontroller based, how about ,making a computer process the data and the microcontroller is there only for sensing stuff.

It is better for you to stick to the original plan. Keep it simple, unless you want the extra challenge (that is, choosing a USB bridge, building drivers, compiling software). And, as you said, you'll still need a micro-controller. So, why not solely based on a micro-controller? A project like this can work stand alone, without computer.

In my case I did it all precisely to avoid a micro-controller, that I wouldn't need anyways. And mind that my objective was to build an instrument without the smarts, no LCD panel, no nothing, except a USB cable and a computer. Your case is the very opposite.

Kind regards, Samuel Lourenço

[Raj]

maybe I should  start with an single board computer gpio.That way i'll get higher speeds in both processing and detection.
But that still won't solve the fact that I don't know how BBs react when made to bounce of a surface

[taydin]

Never give up because you hit a roadblock. Just keep trying different things. Even if you continue trying, failing, trying, failing, trying ... all of these will teach you something new. And maybe after a while, armed with new knowledge, you will finally realize what the problem is. Maybe it was a wrong assumption you had at the beginning. Now with that assumption gone, you go back, redo and now the problem is resolved.