Uploading protocol

@barrie 

When you're dealing with mixed number types you have to bear in mind their type and limitations as you work.

Generally speaking, integers operating on integers result in integers. floats dealing with anything result in floats.

So, suppose you declare 

int x = 5, y=10;

You'd expect Serial.println(x/y); to report .5 but it would actually give you ZERO. That's because they're both integers the result will be an integer and the closes integer to .5 is zero.

On the other hand

int x=5;

float y=10;

Serial.println(x/y); now reports .5 because the value 5 is converted to a float before the operation, so when the division is done it can produce a floating result (and does so).

You can get really screwed up if you're not careful, so it's best to think about what accuracy you need and maybe do the calculations in floating numbers and then convert to an integer after the calc is complete.

Consider the examples

int x=1, y=3;

float q=50;

now (50*x)/y is 16.66 (as you'd expect) 

but (x/y)*50 is zero ! That's because the brackets force the evaluation of 1/2 (both integers) so that gives zero multiples by 50 which is still zero.

Another tangle you'll find is that each category of numbers has a size limit. For instance, the "byte" type is an 8 bit integer, so it can only hold values with 8 bits (=2 to the power 8 = 256 values). So if you multiply 100 x 100 you get 16 (because you've wrapped it around the clock a few times, just like your odometer).

@inq Thanks. I thought the same. The answer was in an Int. It should now be an easy fix.

@will Thanks for all that good info. I now know what to do. 

The float was from the output of the sonic module I had to convert it to an integer in order to use it in a Switch Case function.  the conversion was:   25(int derived from float) * 14(int) = 75(int). I will try an integer with the module.  the 25 above was converted from float:  int x  = y where y is a float 

Posted by: @barrie

@inq Thanks. I thought the same. The answer was in an Int. It should now be an easy fix.

Do you remember about integer overflow/underflow from when you programmed before?  On most computers, they exception when you go over the variable type's limit.  On these microprocessors, they simply wrap to the other side.

I believe on your Arduino, the int is a 16 bit signed integer.  So it goes from -32768 to 32767.  If you were to increment it when it is 32767, it'll actually wrap and its value will be -32768.  Decrementing will hit the bottom and wrap back up to the top.  Point being... If you had something like 

float f = 1020.3;
int i = 42;

// And did...
int j = i * f;

// Internally, the MPU would have:
// 1) upgraded the i to a float = 42.0
// 2) do the multiplication  42.0 * 1020.3 = 42852.6
// 3) In attempting to convert it back to the int so it can place it in j, it would
//    a) truncate it (not round) to 42852
//    b) wrap it since it is higher than 32767 
// So you'd get j = -22684

... is that the "silly answer" type thing you were getting when you were really expecting 42,853?

VBR,

Inq

yes exactly. one ofthevariables the should have bee 2 digits had about 6  or 7. 

@inq @will There is something wrong. I have tried all integers, long's and floats. Still get silly answers. It may be a full memoryl. How can I check Eprom ?

@barrie 

Post your code where it's going wrong so we can see what you're doing.

its this line in the loop at case 2

 int spd = speedFacter * distance;

@barrie 

I can't make heads or tails of what you're doing, but you are assigning an int value (ranges from -32,768 to 32,767) to the product of a long value speedFactor (ranges from -2,147,483,648 to 2,147,483,647) times another int (ranges from -32,768 to 32,767).

What value range can spd, speedFactor and distance take ?

I kind of get the feeling that you set spd as an integer so you could use a "switch" statement.  But it seems that a better choice would be to use value ranges if speed is more reasonably represented as a float.

That is, something like

if (spd<10) {

    Serial.println("We appear to have stalled");

} else if (spd<50) {

    Serial.println("Is the parking brake on");

} else if spd<150) {

    Serial.println("So nice to have Tony the Tiger in the tank");

} else if (spd>500) {

    Serial.println("Is it just me or do you hear a siren ?");

}

@barrie 

I just realized that you define ..

 const int speedFacter = 3.3;

which is totally bogus. 3.3 is a floating point number, it has a decimal part and so it can never be 3.3. This value will immediately be truncated to 3.

You MUST use float data types for values which will have non-zero decimal parts !

@barrie Please use the REPLY link, and turn on line numbers in your Preferences. While there, turn on other helpful features, like Compiler warnings ALL. 

@zander Didn't see a line number option in Arduino 2.0.0 but they are showing in my sketch. I noticed that I could not select the line numbers with the code when I posted the sketch. I also realised I hit the wrong reply button and almost wrote an apology. here is a belated apology. Sorry. 😇 

@barrie Yes, I just checked 2 and it's not there but the sketch has line #'s. I will do a test insert code here to see what happens. While I am nit-picking, it is a lot easier to read code that is uniformlyh formatted. There is a menu item under Tools to auto format the code. It doesn't do much but at least the spacing will be uniform. It's surprising how much that helps. Yep, NO line numbers. I will ask for that to be added.

```cpp
void setup() {
  // put your setup code here, to run once:

}

void loop() {
  // put your main code here, to run repeatedly:

}

```

@inq It is working but not yet suitable for a video. It is too jerky and I need to get the hang of changing stepper speed on the fly. When it works to spec I will try to remember to post a video. If you have any examples using AccelStepper changing speeds please let me know.