DESY recommendation class D1.4 for RTM connector

elhep / AFCZ

AMC FMC Carrier board based on ZYNQ Ultra Scale + SoC

5 stars 7 forks source link

DESY recommendation class D1.4 for RTM connector #29

Open kaolpr opened 4 years ago

kaolpr commented 4 years ago

[x] Rename at least nets to be compatible with D1.4 recommendation (preferably also ports of the RTM_Connector sheet)
[ ] Change NOT USED to ASSEMBLY VARIANT, precisely highlight signals that are available in assembly variant
[x] J31 / 8-10 / ab do not follow the recommendation, evaluate the possibility of keeping current RTM_FPGA_CFG as an assembly option while routing proper signals to these ports
[x] Conider routing RTM_CLK and AMC_CLK to clock cross-bar to enable distribution of clocks from RTM to FMC (e.g. sourcing from uRFB).

filipswit commented 4 years ago

I can rout differentially these signals J31 / 8-10 / ab, but from recomendation they should be GTP or CC. 2020-06-09_11h34_40

The RTM_CLK (which from recomendation is AMC_CLK ) actually goes to cross-bar 2020-06-09_11h56_27

Another one is is "RTM_ADC_DAC_SYNC.SYNCOUT11" and it is possible to route it to cross-bar, but please pont which input I can use to make it useful

gkasprow commented 4 years ago

The plan for RTM in AFCZ was to eventually replace the Sayma AMC. But for the moment it's not important

kaolpr commented 4 years ago

Up to my understanding, GPTs are just transceiver clocks. In Ultrascale architecture, it is possible to also output signal from the REFCLK pin pair. See #37.

Regarding compatibility - I think generally it would be better to stick to the standard. Even if it's not that widespread. Especially if it does not cost us that much (like f.e. in case of AFCv4).