Hello everybody, here are some preliminary news on the protocol decoding stuff (btw, the stored sine waves have been repaired):
- The control message that the FP sends seem to be the FULL configuration, that is, the frequency, amplitude and whatever is send for the both channels and the FPGA reconfigures fully.
- The numerical information (at least for the frequency and amplitude) seem to be organized into 48bit !!! numbers ( 3x16bit), there are at least 32 of this 48bit parameters that are send.
- I've managed to identify the frequency slot, it does increase with this stupid mHz increment, and the number reflects the display.
I'm so curios what is happening if somebody forces a value larger than 60MHz on the slot, especially on the 30MHz units !!!- I've managed to identify the amplitude slot, this somehow varies monotonic with the display, but there on 1 to 1 correspondence, seem to be some quanta, even values from a table, to be investigated.
- I've managed to quickly reach the limits of the Intronix logic analyzer, to get somehow reliable and repeatable measurements at 18MHz clock, I need to sample at 50MHz and the little buffer fills quickly, even with compression active. the software on the other side was so cool, it took me like 5 seconds change the SPI protocol interpreter from 8 bits to 16bits. I really hope that the owner of the LogicPort IP will produce an updated variant or at least open-source it, because I really like it.
- To continue work I need to use the LA advanced features, like repeated condition trigger, to skip groups of 3 SPI transfers, but that means doing single acquisitions instead of looping and comparing values, and this pisses me off.
!!! WANTED !!! Kind Soul to Lease a Logic Analyzer with a RELIABLE sampling mode (3V3 levels) of at least 50MHz (100MHz preferable, not every Chinese firma is Intronix) on at least 4 channels (I need an extra for the extra chip select, to see where this long string of words are going, remember we have 2 chip selects!!!) and yuuuuge buffer, at least 4Msample without compression, there is a large interval in between transfers compared with the actual transfer time of 3 words, and if the LA is of naive design it really needs a large buffer. Hopefully I'll be soon able to get the control message, without a powerful LA will be painful, but it's doable.
Humanitarian appeal: Dear gentle beings that lurks on this thread, I know that some of you have these cool scopes with LA and/or preformant LAs, kindly please try to reproduce my measurements, the connector is gigantic at today's standard and there is no pain in hooking the LA, 5 pins with plastic balls it's all that is needed for the grabbers to have a grabbing point
.
If you don't want, at least some encouraging words will be appreciated
.
I'm posting some pictures with my setup to encourage you to contribute
Cheers,
DC1MC