Date: Thu, 24 Aug 1995 13:26:37 -0700 (PDT) From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com> To: jgreco@brasil.moneng.mei.com (Joe Greco) Cc: fenner@parc.xerox.com, terryl@cs.stanford.edu, freebsd-hackers@freebsd.org Subject: Re: ISDN Anyone? Message-ID: <199508242026.NAA08474@gndrsh.aac.dev.com> In-Reply-To: <199508241950.OAA20237@brasil.moneng.mei.com> from "Joe Greco" at Aug 24, 95 02:50:26 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > Hi Rod, > > > Carefull here, many Uarts are only rated for on f(max) of 5Mhz, others > > are good to 8Mhz. I think all 16550AFN's are rated for 8MHz, but not > > sure if they still had the -5/-8 speed option that late in the game. > > This is something to consider. However I am talking out of a NS data book > and the only references I have seen are: > > NS16450/16C450/ 3.1 MHz max, 56k "top" baud rate suggested > INS8250A/ > INS82C50A > INS8250/ 3.1 MHz max, 56k "top" baud rate suggested > INS8250-B > NS16550AF 8.0 MHz max, 256k "top" baud rate suggested > NS16C451 24.0 MHz max, no suggested top :-) > NS16C551 24.0 MHz max, 1.5M "top" baud rate suggested I am not so worried about the National parts, it is the other manufactures, as you point out later, that have different upper clock rates. I seem to recall the TMS version topping at 5MHz, but maybe confusing that with the Zylog SIO chip. > > Now if I am reading this right the INS8250-B is a slower speed part, judging > from the electrical characteristics. I do not see any lower-speed offerings > on the other parts. Although it IS funny that the 16C451 can have a 24 MHz > clock and they didn't bother to list a top rate, I can imagine that you > could get pretty busy trying to shove a million bits per second through a > device without any real FIFO. I can tell you why they upped fmax to 24Mhz, that is the same clock needed by most floppy controller chips, so now you can run both chips off the same crystal on a multipurpose card. This is quite common in the Winbond and SMC chips, there is an additional divider inside the chip that cuts this down for the uarts, but uses it raw for the floppy controller. Look on any winbond based atio card and you will see 1 24Mhz crystal. The 451/551 integrate some stuff, but I can't recall what ``stuff''. > > Watch the capacitance on that shielded cable, the higher the pf/foot the > > shorter it needs to be. If you use high quality, low loss, low capacitance > > individually shielded twisted pair data cable grounded one side of each > > pair for the TX and RX (yea, okay, so you need a few more wires :-) you can > > run 230KB quite a distance. A scope comes in pretty handy to see your > > signal quality and the receiver as well. > > Or if you are only going 2 feet like I usually do. :-) :-), you can use just about anything for 2 feet < 1Mhz :-). > Anyways the moral of the story is: don't try to hot rod a non-FIFO chip or > you may torch the sucker. (heck, you may even torch a FIFO chip, there are > no guarantees... if you can't afford to lose the UART don't play with fire). Thats basically what I was trying to point out. Not all Uarts are made by National, not all 16XXX clones meet the 8Mhz clock spec of nationals, etc, etc... > And of course if you're using non-NS parts (i.e. Startech, etc) you might > want to check the specs. Inferior chips made for the PC market and all... > I've run Startech at 2x without problems. And I have run the National parts in excess of 10Mhz, but they aren't very stable at that speed, and that was on a bench setup just to see how far I could push one of them :-) Just be aware, YMMV!!! > Joe Greco - Systems Administrator jgreco@ns.sol.net > Solaria Public Access UNIX - Milwaukee, WI 414/342-4847 -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Reliable computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199508242026.NAA08474>