Swordfish

Actually, this raises an interesting question, namely what is the best way to perform the arithmetic required to scale an integer value? My current SF project involves a recursive filter loop in which the variable Output is repeatedly scaled by a factor K and then incremented by a new DataValue , i....

Thanks for the advice Jerry.

I guess that by using Floats I am being lazy, at least during the development stage of my program where parameters are being optimised. In the final version of my program I will replace Float constants with Integer quotients to maintain the precision you describe.

I have run a test of comparisons between pairs of random Float values and another test between a random Float and a fixed Integer (which is where I thought the problem might have arisen). It seems that there is indeed no bug in the latest compiler and so I must apologise for needlessly raising the a...

I am using Version 2.2.0.4 - ICC 1.1.5.5

I will look into this again and report back. I do hope I have not been too hasty!

I have just updated my Swordfish to the latest version and find that the bug when comparing negative Floats still seems to occurs. For example, in my program I want to limit the Float variable Test to be within +/- 20000. A check for Test>20000 works OK but a similar test for less than -20000 does n...

:D Cheers Doj. Thanks for your helpful and detailed clarification. As it happens, the 2.5MHz pulses which I am counting peak at around 4.5V so I am safely avoiding those naughty electrons. Actually, I was'nt aware that fundamental electrons had personalities. Presumably, protons chum up with electro...

Thanks Francis and rmteo. It seems that the safest solution is for me to ensure that the pulses are constrained to no more than VDD+0.3V (=5.3V). This will make for a more reliable product.

:? I am using an 18F252 in a project which counts pulses into RA4 (T0CK1) and RC0 (T1CK1). On my board VDD = 5V. What is the maximum voltage of the pulses that I can apply to these pins? Table 22.2 on the datasheet suggests that it is only VDD (5V) but I am not sure I fully understand the specificat...

Steven,

Sorry to be dumb, but I have written an article for the Wiki (in Word) and have 2 pictures to include (as JPG) but cannot email it without your email address nor do I understand the instructions by J Winters!

Thanks for details of the thread. How do I insert images into my post?

David,

I have a PIC based product which uses Swordfish for the firmware. How can I send the details to be included in the "Featured Products" section of your website?

:? Here are two things which have got me puzzled: In the first, I am simply trying to constrain a Float variable B to be in the range +/- 32767. The following code works OK: If B>32767.0 Then B=32767.0 Endif If B+32767.0<0.0 Then B=-32767.0 Endif However, if I change the second test to: If B<-32767....

:? I have a serial connection from a MAX232 on my PIC board to an RS232 port on my PC. The data are used in a PC graphics and PIC-control application, using both the TX and RX lines. Is there any reason why I cannot split the RX line at the PC into 2 so that I can feed a second RS232 port on the com...

:lol: I too coded for a while in Pic Basic Pro until the lack of signed integers, float, procedures, etc, started to feel like swimming in treacle. Swordfish is a brilliant package at a sensible price. However, it needs more vigorous marketing for the developer to reap his just rewards - a search on...

I completely forgot to do a search! Yes there are some good ideas, thanks.