Hello Adrian and all,
I did indeed implement a four-phase TDC in a Spartan 3 some years ago- it
used a 300MHz clock to achieve 833ps resolution.  I can supply details if
you're interested.
Rick

On Sat, Dec 26, 2015 at 1:44 PM, Peter McMahon <[email protected]> wrote:

> Hi Adrian,
>
>
>
> There’s an open-source TtD converter design available at OHWR that might
> be of interest to you. The paper describing it is
> http://arxiv.org/pdf/1303.6840v1.pdf and the repository is at
> http://www.ohwr.org/projects/tdc-core/repository.
>
>
>
> Cheers,
>
> Peter
>
>
>
> *From:* [email protected] [mailto:
> [email protected]] *On Behalf Of *Dan Werthimer
> *Sent:* Saturday, December 26, 2015 1:00 PM
> *To:* Adrian Sinclair <[email protected]>
> *Cc:* CASPER Mailing List <[email protected]>
> *Subject:* Re: [casper] Asynchronous inputs?
>
>
>
>
>
> hi adrian,
>
>
>
> yes, an fpga inputs can be used as an asynchronous input:
>
> an fpga input block has several D flip flops, but they are optional.
>
>
>
> several people have implemented time to digital converters with FPGA's.
>
> and there a bunch of papers on this.
>
> you can try googling "time to digital converter FPGA".
>
>
>
> these guys get 30 ps resolution ! :
>
> http://cds.cern.ch/record/1158663/files/p383.pdf
>
>
>
> i think casperite rick raffanti has done something like this,
>
> so i'm ccing rick on this.
>
>
>
> if you only need 0.5 ns resolution, a simple
>
> technique is to use the fpga's clock generators to generate several clock
> phases:
>
> eg:  0, 45, 90, 135, 180, 225, 270 degrees,
>
> and then send the input to D flip flips clocked at these different
> phases....
>
> if you clock at 250 MHz, then you would get 0.5 ns resolution.
>
>
>
> if you only need 1 ns resolution, you can use the input block's flip flop
>
> and shift registers - i think the input block is already set up to do QDR -
>
> quad data rate sampling.
>
>
>
> best wishes,
>
>
>
> dan
>
>
>
> On Sat, Dec 26, 2015 at 12:27 PM, Adrian Sinclair <[email protected]>
> wrote:
>
> Hello,
>
> I'm wondering if anyone knows of a way to input an asynchronous signal to
> the Virtex 6 on the ROACH2. My goal is to design an all internal
> time-to-digital converter with sub-clock period resolution, and would like
> the ability to send a signal through a string of buffers with registers
> tapped after each one, and upon the arrival of the clock pulse sample the
> entire buffer lines registers. More specifically my question is, Does the
> signal get latched when using gpio's, even when not packing a register in
> the pad?
>
> Thanks in advance,
>
> Adrian Sinclair
>
>
>

Reply via email to