In case your existing plan doesn't work: So what frequency are you running out of the rubidium?
I've admittedly never had to deal with a GPS-aligned T1 reference clock. I'm assuming you both have to control the exact frequency and also the timing of the superframe, and probably even the edge alignment? In my mind, the trick with using an arduino (or other small micro) is to not bit-bang it but instead to clock one of the internal USART's with an appropriate clock. Assuming one disciplines an oscillator to some multiple of the T1 frequency--possibly even exactly the frequency--one should be able to divide this down and feed it into the external clock input on the processor, using that to clock out the bitstream which is generated by the processor. The only difficult trick will be to adjust the superframe bits to match the GPS time. However with the disciplined clock this should become rather trivial as you should be able to fix this within a superframe of the T1, and then clock it every 1544 bits as needed. The clock disciplining could/should also be done with the microcontroller controlling an OCXO. If you're clocking late, just adjust the OCXO to run a bit faster. Early, run it a bit slower. On Thu, Feb 22, 2018 at 6:34 PM, Chuck McCown <ch...@wbmfg.com> wrote: > No, I saw it, but I already had the Pi so I ignored it and hoped for the > best. > I tried it first with arduino. Just not enough speed. And it had the > jitter problem too. > > I have a method using three TTL/CMOS chips that is going to work... I > think... > > The only thing separating the gps disciplined signal and the T1 will be a > D flip flop and a few AND/NAND gates. So that ought to get me super low > jitter. > > Trying for stratum I with rubidium hold-over isochronous performance at > the end of the day. For cheap. > > *From:* Forrest Christian (List Account) > *Sent:* Thursday, February 22, 2018 6:27 PM > *To:* af > *Subject:* Re: [AFMUG] OT Raspberry PI > > Ok, I think you missed this portion of my email last time: > > > *"I'm skeptical that you'll be able to generate a bitstream with enough > accuracy under Linux, without extreme programming measures.I'd suggest a > digilent chipkit wifire and the arduino ide for this. You should be able > to bitbang at least a T1 with this processor (500mhz)"* > > Generally the raspberry pi is great for 'tiny server stuff', or 'user > interface' stuff, and the arduino and/or microcontrollers will work better > for what you're looking at, since there isn't an operating system in the > way. All arduino really is is a c++ ide with some simplified libraries. > > The wifire product I mentioned is really a PIC32MZ dev board, optimized > for arduino. If you've had enough of the arduino ide, you can download the > microchip ide and program it with a full development kit. > > Honestly for what you are talking about a EUSART in even a low end PIC > might be able to handle this. If you program the EUSART into synchronous > mode you'll just have to stuff a byte into it every 8 bit times and it will > clock it out for you. Not sure if the clock rate is adjustable enough for > you, but if you get a PIC with the NCO peripheral you might be able to > dynamically adjust the frequency enough to make it work. > > NCO app sheet: http://ww1.microchip.com/downloads/en/AppNotes/ > 90003131A.pdf > > > > > On Thu, Feb 22, 2018 at 3:48 PM, <ch...@wbmfg.com> wrote: > >> Anyone know how to get my program to run on bare metal? >> >> Or at the very least tell Linux that my program is the most important >> thing in the world and service it above all other things. >> >> I am trying to create a timing signal with the Pi. It is doing it but >> the jitter is pretty bad. >> >> I have researched trying to use an interrupt but there is a pretty low >> limit on how many times per second you can fire a hardware interrupt. >> Too low for my application. >> > > > > -- > *Forrest Christian* *CEO**, PacketFlux Technologies, Inc.* > Tel: 406-449-3345 | Address: 3577 Countryside Road, Helena, MT 59602 > forre...@imach.com | http://www.packetflux.com > <http://www.linkedin.com/in/fwchristian> <http://facebook.com/packetflux> > <http://twitter.com/@packetflux> > > -- *Forrest Christian* *CEO**, PacketFlux Technologies, Inc.* Tel: 406-449-3345 | Address: 3577 Countryside Road, Helena, MT 59602 forre...@imach.com | http://www.packetflux.com <http://www.linkedin.com/in/fwchristian> <http://facebook.com/packetflux> <http://twitter.com/@packetflux>
