glowell@hpcupt1.HP.COM (Gary Lowell) (01/14/88)
From glowell Wed Jan 13 14:23:18 1988
Relay-Version: version Notes 2.8 87/11/6; site hpcupt1.HP.COM
From: glowell@hpcupt1.HP.COM (Gary Lowell)
Date: Wed, 13 Jan 1988 22:23:18 GMT
Date-Received: Wed, 13 Jan 1988 22:23:18 GMT
Subject: Wire-wrap Question
Message-ID: <7110003@hpcupt1.HP.COM>
Organization: Hewlett Packard, Cupertino
Path: hpcupt1!glowell
Newsgroups: sci.electronics
Greetings -
Does anybody know if wire-wrap can be used to build a reliable proto-type
at 20 Mhz clock rates and what techniques can be employed minimize
problems ? I am trying to determine the feasibility building a homebrew
computer using 20 Mhz 80386 chipset.
Thanks for the help.
--
Gary Lowell
(408) 447-7997
{hplabs,uunet}!hpda!glowellcoltoff@PRC.Unisys.COM (Joel Coltoff) (01/14/88)
In article <7110004@hpcupt1.HP.COM> glowell@hpcupt1.HP.COM (Gary Lowell) writes: >Does anybody know if wire-wrap can be used to build a reliable proto-type >at 20 Mhz clock rates and what techniques can be employed minimize >problems ? I am trying to determine the feasibility building a homebrew >computer using 20 Mhz 80386 chipset. > 3M make a prototyping system that is much better than wire wrap. I don't know what it is called and the one I use is in another build 15 miles away so I can't check. It works like this. Each socket has pins that look like |\ /| | \/ | (maginified mucho times) | | and you run a wire along and mash it into the pins. You can squeeze upto three wires in a hole. You should be able to run an entire net with 1 wire since there is no reason to cut the wire at each tie point. It avoids alot of the problems that wire-wrap has such as pins nicking nets when you make a right angle bend. (What do you mean you don't have any right angles? You wire point to point and then mash the wires down. I've never done that. Ha Ha) - Joel
gary@grc.UUCP (01/15/88)
In article <7110004@hpcupt1.HP.COM> glowell@hpcupt1.HP.COM (Gary Lowell) writes: > >Does anybody know if wire-wrap can be used to build a reliable proto-type >at 20 Mhz clock rates and what techniques can be employed minimize >problems ? I am trying to determine the feasibility building a homebrew >computer using 20 Mhz 80386 chipset. > >Thanks for the help. > >-- >Gary Lowell I have built a few prototypes using 20 MHz clocks. I have found it can cause lots of little glitches that are hard to track down. Most of these seem to disappear when a printed circuit version is built. We have stopped building wirewrap prototypes of high speed circuits because of these flakey problems, and start with a PC prototype. However, if you still want to try, here is what I suggest: 1) Start with a wirewrap board with a good ground plane. One side should be solid ground (except for small areas for pins), and the other should be power or alternating power and ground. 2) Use *lots* of bypass capacitors. 3) Keep ICs as close together as possible to keep wire lengths to a minimum. 4) Use a twisted pair for long wires and for wires that carry high speed signals (such as clocks), or where a large number of wires are next to each other for a long distance. Connect the second wire to ground at each end. 5) Memory arrays are a special problem because signal reflections can mess up the close timing required to run dynamic RAM. Consider using a memory card from another computer and interface that to your wirewrap CPU. Considering the pain it is to wire up a large array by hand, you would probably be better off buying a memory card anyway. Good luck! - Gary. -- Gary Sutcliffe W9XT (414) 644-8700 GENROCO, Inc. Slinger, Wis. {ames, rutgers, harvard} uwvax!uwmcsd1!grc!gary Disclaimer (data compressed): "f+)Je\ ^_RMv9h|O |j8gELc=tAYw$pfs!1"
ajg@whuts.UUCP (GAETA) (01/15/88)
Joel; The method of board building that you are referring to is called Quick-Connect. Tony Gaeta
jja@tut.fi (Ahola Jari) (01/16/88)
In article <7110004@hpcupt1.HP.COM>, glowell@hpcupt1.HP.COM (Gary Lowell) writes: > Greetings - > > Does anybody know if wire-wrap can be used to build a reliable proto-type > at 20 Mhz clock rates and what techniques can be employed minimize > problems ? I am trying to determine the feasibility building a homebrew > computer using 20 Mhz 80386 chipset. > > Thanks for the help. > > -- Hi there! I use Vero- Bicc's Speedwire wiring system as my prototyping system, and I'm very satisfied with it. Currently I have an Intel 80186 micro wrapped on a plain non-metallized epoxy board and it runs happily at 10 MHz. For the display card I've used a metallized card because of the high (24 MHz) operating frequency. These board are made in various sizes from E1 to extended E3, S-100, STD- bus, PC-AT and multibus compatible. Boards are available as plain epoxy cards or metallized and silk screened. Metallized boards are equipped with power distribution planes and plated through holes. The wrapping pins are made of beryllium copper and the cup is coldplated. Standard AWG30 ww wire is used as wrapping wire. Each of the pins can take two wires passing through the pin, so there is four wires maximum to one pin, and no need to cut the wire when joining two or more (busses) pins. Normally the boards are drilled for DIL- devices, but some boards exist for PGA devices (I've got the only one in Finland :-) ). This system is nearly optimal for high speed ww projects because the wire lengths and impedances can be controlled easily. If you need more info please contact directly or via news. P.S. I'm also building a 80386 box with Austek's cache controller and 82786 display processor, Happy Hacking! jja Tampere University of Technology Finland
johng@trwind.TRW.COM (John Greene) (01/16/88)
> >Does anybody know if wire-wrap can be used to build a reliable proto-type >at 20 Mhz clock rates and what techniques can be employed minimize >problems ? I am trying to determine the feasibility building a homebrew >computer using 20 Mhz 80386 chipset. > We have wire-wrapped several boards with clock rates of 20 and 30 MHz with little or no trouble. The main concern is to keep the wires as short as possible and bypass, bypass, bypass. -- John E. Greene "People are just like frankfurters....You have to decide if you're going to be a hot dog or just another wiener" DLR TRW Information Networks Division 23800 Hawthorne Blvd, Torrance CA 90505 ARPA: johng@trwind.TRW.COM USENET: ..trwrb!trwind!johng
johne@astroatc.UUCP (Jonathan Eckrich) (01/16/88)
In article <7110004@hpcupt1.HP.COM> glowell@hpcupt1.HP.COM (Gary Lowell) writes: > >Does anybody know if wire-wrap can be used to build a reliable proto-type >at 20 Mhz clock rates and what techniques can be employed minimize >problems ? I am trying to determine the feasibility building a homebrew >computer using 20 Mhz 80386 chipset. > Much to my surprise, I learned from some fellow engineers here that wire wrap is actually *more* reliable than say copper clad-type pc boards. We have wired wrapped a few prototype boards that used 22.5 Mhz clocks. To improve reliablity, make sure you use at least one bypass capacitor per every two chips. Place them as close to the chips as possible. For the proto board, use somthing that has copper or tin on one or both sides. You then use a cyldrical-like scrapping tool to remove just enough of the metal around each hole that you plan on having a component's pin inserted. The main thing to remember is that higher frequency circuits generate more noise, and are more suseptable to noise, so use those bypass caps, and employ GOOD grounding and shielding techniques. Try to keep the lengths of the wires short, and keep the design as "clean" as possible. -- ------------------------------------------------------------------------- Jonathan Eckrich | (rutgers, ames)!uwvax!astroatc!johne Astronautics Technology Center | ihnp4!nicmad!astroatc!johne Madison, WI | (608) 221-9001
dale@lamont.Columbia.edu (dale chayes) (01/17/88)
In article <5256@burdvax.PRC.Unisys.COM>, coltoff@PRC.Unisys.COM (Joel Coltoff) writes: > 3M make a prototyping system that is much better than wire wrap. > I don't know what it is called and the one I use is in another Its called Scotchflex. We use it extensively. As long as you are careful with the boards (its not too hard to pull a wire out or press an extra one in) its great. It allows much thinner proto-type boards, and its real quick to move a wire or add one in the debugging process. -- Dale Chayes Lamont-Doherty Geological Observatory of Columbia University usmail: Route 9W, Palisades, N.Y. 10964 voice: (914) 359-2900 extension 434 fax: (914) 359-6817 usnet: ...philabs!lamont!dale
johne@astroatc.UUCP (Jonathan Eckrich) (01/19/88)
In article <2378@korppi.tut.fi> jja@tut.fi (Ahola Jari) writes: > >I use Vero- Bicc's Speedwire wiring system as my prototyping system, and I have been looking for a place that will sell me this but distributers are the only ones who I can find, and they are not interested in the small (<1000) quantities that I need. I understand that 3M has a similiar product. Could you tell me where I could buy some for my own use. Thank you. -- Jon Eckrich | (rutgers, ames)!uwvax!astroatc!johne Astronautics Technology Center | ihnp4!nicmad!astroatc!johne Madison, WI | (608) 221-9001 N1000M, 1948 Stinson 108-3, Flying Station Wagon
mbutts@mntgfx.mentor.com (Mike Butts) (01/19/88)
In article <7110004@hpcupt1.HP.COM>, glowell@hpcupt1.HP.COM (Gary Lowell) writes: > > Does anybody know if wire-wrap can be used to build a reliable proto-type > at 20 Mhz clock rates and what techniques can be employed minimize > problems ? I am trying to determine the feasibility building a homebrew > computer using 20 Mhz 80386 chipset. > In my experience, wire-wrap can work well for small high-frequency runs, IF YOU'RE CAREFUL. Your main enemies are inductance and stray coupling. There are several considerations which will help: 1) I believe the part of your system which actually runs at 20MHz will be small. Place the parts so as to minimize the 20MHz run lengths. Keep in mind, though, that edge rates will be fast for many signals, regardless of their actual frequency, so be mindful of the following. 2) Use twisted pair for the 20MHz runs. That is, use a twisted pair of 30ga. wires, with the ground wire wrapped to ground posts as close to the source and destination pins as possible. Needless to say, keep your runs as short as possible (within reason). 3) Run fast signals directly point-to-point, not in neat up and down channels, to avoid crosstalk. The board may not look pretty, but the signals will look better. 4) Use a board with a ground plane, and ideally with power and ground planes. (I know Augat makes such boards, they or others may even have app. notes on this subject.) 5) Use high-frequency 0.1 mfd bypass caps very liberally - one on each chip and more on the LSI. Place them close to the Vcc and Gnd pins. You might even consider soldering them to the posts. Caps are cheap, and bad bypassing will give you the most hateful kind of flakey bugs. I've never worked with 20 MHz microprocessors, but have seen successful experience with multi-board FAST-TTL systems running 8MHz globally and 16 MHz in small parts. Good luck and have fun! Here's to the home-brewer! -- Mike Butts, Research Engineer 503-626-1302 Mentor Graphics Corp., 8500 SW Creekside Place, Beaverton OR 97005 ...!{sequent,tessi,apollo}!mntgfx!mbutts OR mbutts@pdx.MENTOR.COM These are my opinions, & not necessarily those of Mentor Graphics.
dgt@myrias.UUCP (David Tang) (01/20/88)
Another manufacturer of a similar system is Robinson Nugent (RN). A hint: When distributing a clock signal, you may wish to double up the wiring. We've found this makes for a cleaner looking signal. (Probably has something to do with wire inductance).
jja@tut.fi (Ahola Jari) (01/28/88)
In article <740@astroatc.UUCP>, johne@astroatc.UUCP (Jonathan Eckrich) writes: > Robinson Nugent (RN), 3M, and others make this low profile 'quick' wiring > system that we have been talking about on the net for the past two weeks. > I have been trying to get a hold of some for my own personal use, but I have > found it to be very hard due to the fact that distributers are not willing > to deal with the kind of low volumes that I need. Does ANYONE know of a > mail order house, etc., that sells these little things? > Followup, or E-mail is OK. > I'm working on this, but I haven't got a reply to my telefax from Vero- Bicc yet. Be patient. jja
jja@tut.fi (Ahola Jari) (01/29/88)
In article <740@astroatc.UUCP>, johne@astroatc.UUCP (Jonathan Eckrich) writes: > Robinson Nugent (RN), 3M, and others make this low profile 'quick' wiring > system that we have been talking about on the net for the past two weeks. > I have been trying to get a hold of some for my own personal use, but I have > found it to be very hard due to the fact that distributers are not willing > to deal with the kind of low volumes that I need. Does ANYONE know of a > mail order house, etc., that sells these little things? > Followup, or E-mail is OK. > Hi John! I got a reply from Vero- Bicc UK and they told me the following address in the USA: Bicc- Vero Electronics Inc 1000 Sherman Avenue Hamden Connecticut 06514 tlx no. 510227889 fax no. 2870062 Contact name: Chris Lockwood If you still can't get what you need, let me know so I'll inform the factory as they ask me to do so. And if you are lucky let me know the prices you are paying for them (they are pretty expensive here in Finland). jja