New Hubsan Upgraded Version on the way

PhracturedBlue wrote: The MCU is running at 72MHz, so 20 nops should be plenty to guarantee a single read occurs, but maybe something screwy is going on here.

for(i = 0; i < 20; i++)  //Wait > 1 SPI clock (but less than 8).  clock is 4.5MHz
    asm volatile ("nop");

SeByDocKy wrote: And what about the RSSI ?

goebish wrote: looks like there are not only telemetry frames received but other packets in which packet[13] is not the voltage value.

goebish wrote: I confirm I found (and fixed) what was fishy in A7105_ReadReg()

PhracturedBlue wrote: The MCU is running at 72MHz, so 20 nops should be plenty to guarantee a single read occurs, but maybe something screwy is going on here.

Actually this is plenty, this is even too much at times , it lasts more than 20 cycles because the nop instruction is in a for loop (unless you tell the compiler to unroll loops, but I think that's not the case).

for(i = 0; i < 20; i++)  //Wait > 1 SPI clock (but less than 8).  clock is 4.5MHz
    asm volatile ("nop");

Change 20 to 10 and the problem is gone...
This also explains why the Hubsan4 protocol wouldn't bind sometimes.

PhracturedBlue wrote: The problem is that I have no control of the compiler, so it could vary from compiler to compiler. the only safe thing to do is to put the entire sequence into assembly to guarantee consistency on all compiler versions.

vlad_vy wrote: It works with Devo8 and H107C, but background always red (Telemetry monitor), on Main page the telemetry box also is always red and value not changed. Something wrong with telemetry update function.

With H107 it also works, but value frequently alternate between 0.2V and 3.8V.

goebish wrote: No idea why it alternates between 0.2V and 3.8V on the H107, I don't have this issue with the H107L, will have to wait for more reports.