Spike in the ESC

[lang-simon]

Hello !
I have a weird problem with your code.
The compiled executer works like a charm on windows.
But if i compile it with Qt, the esc has what seems like current spikes/ very fast speed changes . The motor thus make a strange noise. I could record it for you if you want.
The frequency of that noise change with the speed. Ie for a throtlle of 1200s it appears every 2 seconds, and faster for faster speed.

Do you have an idea of the differnce between the compiled and the code ?
Many thanks

[iforce2d]

No idea sorry. I do recall that some Arduino boards with CH340/CH341 chips had a problem, not sure if it's solved yet.
martinling/libserialport#29

Perhaps you could make sure that the .dlls being used by each build are the same.
http://www.dependencywalker.com/

Maybe when building, there is one on your computer that is preferred over the ones in my distribution.

[lang-simon]

OK! Thanks a lot! My board doesn't have a ch480 chip. At least if the code is the same for both that doesn't leave a lot of room for such differences! I'll investigate more next week. I guess that if the computer is sending speed change, I'll be able to see them with on oscilloscope and/or log them from your program. Eitherway thanks a lot for your tools, they are very useful for my master thesis! :) Le ven. 5 avr. 2019 à 15:43, iforce2d <notifications@github.com> a écrit :

…

No idea sorry. I do recall that some Arduino boards with CH340/CH341 chips had a problem, not sure if it's solved yet. martinling/libserialport#29 <martinling/libserialport#29> Perhaps you could make sure that the .dlls being used by each build are the same. http://www.dependencywalker.com/ Maybe when building, there is one on your computer that is preferred over the ones in my distribution. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#2 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AuGKYx8u-bYHl4Vg5RG3r82S3MG2RxBRks5vd8OIgaJpZM4cfrg9> .

[martinling]

Hi, libserialport author here.

It looks like this project is not setting all of the configuration parameters for the serial port, so you are getting some random settings that will depend on the OS, the serial port driver or the last program to use the port.

I looked at this version of the code: https://github.com/iforce2d/thrustTester/blob/60a5e78e00620b318d4a3971294f9b8eca87e6c5/src/mainwindow.cpp

I see calls to:

sp_set_baudrate(port, 9600)

and

sp_set_bits(port, 8);
sp_set_parity(port, SP_PARITY_INVALID);
sp_set_xon_xoff(port, SP_XONXOFF_OUT);

Calling sp_set_parity with SP_PARITY_INVALID will not do anything. SP_PARITY_INVALID is a value used to indicate an unset or unrecognised parity setting in a port configuration. If you want the port configured with no parity bits/checking, you need SP_PARITY_NONE.

You need to also call sp_set_stopbits() with the number of stop bits you want (usually 1).

Finally, you're setting SP_XONXOFF_OUT which is fine but doesn't guarantee that RTS/CTS or DTR/DSR hardware flow control isn't also configured on the port by defaults or a previous program.

So my suggestion if you need XON/XOFF flow control for TX only is to do:

/* Start with all flow control disabled to get a known configuration. */
sp_set_flowcontrol(SP_FLOWCONTROL_NONE);
/* Now enable XON/XOFF software flow control for output only. */
sp_set_xon_xoff(port, SP_XONXOFF_OUT);

I realise the documentation wasn't previously clear about how the configuration API is supposed to work and what settings are essential. I have added some more explanation here:

https://www.sigrok.org/api/libserialport/unstable/a00022.html#details