In a custom board we created, it is required to use LVDS pins in two banks. I know that the VCCIO of the corresponding bank should be 2.5V. Unfortunately the VCCIO was connected to 3.3V instead of 2.5V by mistake.Can LVDS still work? The bank has LVDS input data pin and input CLK and output data.
It will 'work' but won't conform to the timings specified in the datasheet. If you only have LVDS input signals then, perhaps, it will simply work as you hope. Output LVDS signals would be a different issue.Either way - whether it appears to work or not - I recommend you correct the design error. Cheers, Alex
This FPGA (Cyclone 5 btw) will talk to another FPGA (Cyclone 4) through LVDS.Will output LVDS signals be an issue? Fixing the design will require few weeks for new board manufacturing. I was hoping to test the whole system first because there are other components, .....
Is there no way of modifying the board such that you feed that bank with the right voltage?LVDS output signals will simply be a little misshaped. Being a current based signalling standard, the IO circuitry - driven at 3.3V - will be able to drive the rising/falling edges faster. They're unlikely to damage the Cyclone IV and it will most likely just work. However, you will be operating your devices out of spec. So, Altera won't support you should you suddenly start burning out parts - something I think unlikely to happen but it's still a potential issue. Cheers, Alex
The main problem is that the 3.3V is supplied to few other components and banks and I cannot just modify it. If I fed it with 2.5V, other FPGA banks may not be able to communicate with other chips.I will try to reduce the voltage to 3V and try to power and test the system. I hope it works even for just testing. New boards should be designed later. Thanks.
The input differential clock is seen successfully inside the FPGA. I have other problems with the chip and opened another thread for them.