What is wrong with this division?

[tjkelly]

I am compiling this with the long version of PBP and using an 18F2420.

Why do I get the result of the division as shown below?
I expected it to be 0xC455A


P var Long
current_voltage var Word


current_voltage = 49800
P = 78610 * current_voltage ; results is 0xE956C590
P = P / 4868 ; result is 0xFFFECEED in MPLAB watch

[mackrackit]

Looks like you need Extra_LONGS for that. :)
The multiplication over flows the LONG VAR.

Try


current_voltage = 49800
P = 7861 * current_voltage
P = P / 4868 * 10

[ScaleRobotics]

Nevermind.

[tenaja]

Looks like you need Extra_LONGS for that. :)
The multiplication over flows the LONG VAR.

Try


current_voltage = 49800
P = 7861 * current_voltage
P = P / 4868 * 10


Dave, I thought Longs were 4 bytes... ?
49800 x 7861 = 391477800 = $1755 7A28 ...which is 4 bytes...

[Acetronics2]

Hi, tenaja ...

I think you've been cheating ...


Dave is the best ...

LONGS are SIGNED ... ( Manual p.24 )

Alain

[tenaja]

Yeah, but it's still only 29 bytes. Does the "sign" take up 4 bytes?

[mackrackit]

The OP had
49800 * 78610 = 3914778000

[tenaja]

Yes, but he also said it did not match the calculator output. My point was that the calculator proved it fits within the constraints of a Long.

[mackrackit]

A calculator gives
3914778000
LONGS can be
2147483647

[Acetronics2]

Yeah, but it's still only 29 bytes. Does the "sign" take up 4 bytes?

in fact, the sign is bit 31 ...

so, ... better let's say +/- 3 3/4 bytes ... for maximum value

Alain

[tjkelly]

Yeah, what a dumb mistake on my part. I had forgotten about the sign bit being lost resulting in only 31 bits. Thanks.

[tenaja]

A calculator gives
3914778000
LONGS can be
2147483647
My bad. I missed the last zero when entering it in the calculator. Apparently I took the numbers from your code and mistook it as a quote.