FPGA sample clock match data

[Adriana-DMIntegrations]

Hi,

I have a question regarding FPGA sample clock. If the data that must be read match the sample clock rising edge, what would FPGA I/O node read? I attach an image for clarification.

Thanks,

Adriana.

[GerdW]

Hi Adriana,

---

@Adriana-DMIntegrations wrote:

If the data that must be read match the sample clock rising edge, what would FPGA I/O node read?

---

That is rather undetermined.

Are you sure the internal FPGA clock and the external signal are the very same down to the picosecond/femtosecond range (in the time axis)?

[Adriana-DMIntegrations]

It's a theoretical question. If they would perfectly match, what would the FPGA I/O Node read. I guess that 1, if the clock triggers reading data on rising edge but I am not sure.

[GerdW · Accepted solution]

Hi Adriana,

---

@Adriana-DMIntegrations wrote:

It's a theoretical question.

---

What's the practical use of this theoretical question?

---

@Adriana-DMIntegrations wrote:

If they would perfectly match, what would the FPGA I/O Node read. I guess that 1, if the clock triggers reading data on rising edge but I am not sure.

---

"Perfectly match" on the outside of your DAQ device or in the ADC internals? Current might flow fast, but is also limited to the speed of light: every millimeter counts!

(Theoretical answer: I guess you would read a 1 when the sampling starts with the trigger edge…)

[Adriana-DMIntegrations]

In the ADC internals. But it's sufficient to know that it would be 1.

Thanks.

[rolfk]

In theory it’s 1, in practice almost certainly 0. Every digital circuitry has something called minimum setup time. It’s the time that a signal needs to be present at the input before the active edge of the according clock/sample control that causes the input to be read.

That can sometimes be 0ns but it is usually a few ns or more. If that setup time is not observed the value that is read is very likely the previous value but the minimum setup time is a guaranteed value so the circuit may be faster (depending on power supply level, temperature and such).

Also note that the voltage level on which the clock switches does not necessarily have to be the same as the data input. Digital signals are in theory just 0 and 1, in reality (assuming TTL level for this exercise) 0 is a voltage between 0 and 0.8V and 1 is a voltage higher than 2.4V. That range between 0.8 and 2.4 is undetermined and one input may switch at 1.2 V while the other may switch at 2V

For 3.3V logic pretty much the same applies although depending if it is TTL compatible or not with slightly different ranges.