Date: Tue, 7 Oct 1997 04:34:59 -0400 (EDT) From: Peter Dufault <dufault@hda.com> To: mdean@best.com (mdean) Cc: dufault@hda.com, hackers@freebsd.org Subject: Re: A world of unexplored pain. Message-ID: <199710070835.EAA14531@hda.hda.com> In-Reply-To: <Pine.SGI.3.95.971006232803.27801A-100000@shellx.best.com> from mdean at "Oct 6, 97 11:39:17 pm"
next in thread | previous in thread | raw e-mail | index | archive | help
> > > > on the A-D board to watch for lost ticks by setting it for something > > like (N + fudge) times the tick interval and verifying it wasn't > > pending when I got to the control. > > So are you saying you were running at an interrupt frequency of 10000 Hz > (gee wiz ma I only want 1000hz)? So what did you see in the way of lost > tick statistics? I did say that (100 time 100 Hz), but I lied. A check of the code shows #defines of 4, 8, and 32, and the last enabled one was 8. Sorry. I wasn't DOING much - an ordered delta timer list to when to toggle a bit, so decrement a counter, see if it was zero, and toggle a bit if it was. (...) > Many people would have me believe this can't be done adequately with > freebsd. You wrote labpc.c (I've been using it as a reference, I have one > of those boards I want to connect the stepper motor drivers to.) excellent > piece of code, there are a lot of features of the labpc I've never even > ventured into. I particularily like its ability to post an interrupt on > completion of forming an AtoD value. Unfortunately that doesn't work right > once youv'e connected it to an external multiplexer system, because there > is no good way to tell that the Analog Value has stabilised, you just need > to _WAIT_ a prescribed amount of time. This is precisely why I wan't to put > scans inside a 1000hz interrupt routine. What are you using for an external multiplexer? If you're using the National SCXI stuff that uses a serial bus on the SCXI backplane to setup channel lists, then I have some user land code to set that up. The SCXI bus controller (in the SCXI backplane) has scanning hardware on it to multiplex the external signals back to the system A-D board and provide the trigger for that A-D board to do the conversion. They must provide for a settling time after switching. Have you looked to see if that does what you want? A mux settling time is going to be extremely short compared to a software clock tick. > > Things have grown a lot since then. > > Grown to what? I meant FreeBSD has grown a lot and there are more places to look for problems. I wish I'd meant that the aproach had grown. -- Peter Dufault (dufault@hda.com) Realtime development, Machine control, HD Associates, Inc. Safety critical systems, Agency approval
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199710070835.EAA14531>