#robotics Logs

May 08 2019

12:18 AM zhanx_ is now known as zhanx
03:15 AM Jak_o_Shadows: yaqwsx, I am looking, will get back to you with rough thoughts before 1.5 hours has passed
03:21 AM yaqwsx: Jak_o_Shadows: Thanks a lot!
03:59 AM Jak_o_Shadows: Thoughts as I have them:
04:02 AM Jak_o_Shadows: * The way we are doing it at work is prettier - I don't much like how far away from the registers & reference manual the lib is
04:02 AM Jak_o_Shadows: * setupRx seems ok - but I would just make it full duplex.
04:02 AM Jak_o_Shadows: * setupTx probably needs the rx stuff setup as well - all SPI comms can be duplex, so that's the way the hardware probably works
04:02 AM Jak_o_Shadows: * I don't see where you're turning SPI on...
04:02 AM Jak_o_Shadows: * You can't rely on the rx or tx DMA transfer complete signals. Because of the buffer in the RX/TX, you can miss bits. You have to do it with the SPI
04:03 AM Jak_o_Shadows: The way I did it, you disable SPI, you enable DMA, you chip select SPI, you enable SPI, DMA finishes, SPI finishes, you disable SPI, you release CS
04:05 AM Jak_o_Shadows: mumptai (who is not here) that leg design is nice.
04:06 AM Jak_o_Shadows: yaqwsx, hmm, I wonder what tx/rx FIFO buffer stuff you are using (if any)
04:15 AM Jak_o_Shadows: also, I gotta ask, why the G series?
10:03 AM yaqwsx: Jak_o_Shadows: I have solved the problem. The problem was in unaligned memory for the DMA. Now I can do both - receive and transmit independently on each other.
07:58 PM rue_shop1: its working!??? yayyy!!
09:00 PM Tom_L: what is working?