wire wrap advice
smail...@yahoo.com
memory, latches, ... I think there will be about 10 ICs. I am new to
wire wrapping and need to purchase the tools, wire, and prototyping
board.
Can someone give me advice on what tools, prototyping boards to use?
Also best practices advice would be nice. I realize the importance of
decoupling caps, but what about things like:
choosing boards with ground planes, power buses, and connecting power
and ground pins.
what wire gages to use. Same for power, ground, and signal?
Any tricks for connecting data and address buses?
Specific manufacturers and part #s for tools and boards would be great.
Thanks in advance,
Scott
Tim Wescott
By the time you spend enough to be up and running efficiently doing wire
wrap you will have spent enough for your first two quick-turn 2-layer
prototype board designs -- and soldered boards are much better than
wire-wrap.
-------------------------------------------
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
Luhan Monat
Wire wrapping is a good choice if you are not very experienced.
Microcontrollers, unlike microprocessors, are very forgiving on layout
and such since rom and ram are internal.
Wire up all the power and ground lines first. Run redundant connections
in a 'matrix' manner. Put some 0.1 or so ceramic capacitors in several
places spaced around the board and tie into the power and grounds.
I use #30 wire for everything. Pad-per-hole boards allow you to solder
two diagonally opposite pins on DIP sockets to hold them in place.
Quarter watt resistors can be trimmed, bent, and inserted into DIP
sockets. The same for small capacitors.
Most all of the projects shown on the site below are done in wirewrap.
--
Luhan Monat: luhanis(at)yahoo(dot)com
http://members.cox.net/berniekm
"Any sufficiently advanced magick is
indistinguishable from technology."
cbarn...@aol.com
for wire wrap. The whole job would take only 2 hours or so to do once
the circuit has been designed. Reliability is also good, wrap joints
are good for at least 10 years allthough I have 20 yearold boards that
still work fine.
dave garnett
I'd advise using a protoboard with a collander ground, and soldering
ground pins directly to it. You will save yourself untold trouble !
Similarly you can solder your decoupling caps directly between ground
and the Vcc pin.
Lots of colours will help you follow wiring more easily. You used to be
able to get little labels which pushed onto dips and gave you the pin
numbers - saves hours of counting.
Another useful tip is that there are two sorts of wire-wrap wire -
tefzel and kynar insulated. Kynar is tough, resists soldering irons and
so on - but a pig to strip. Tefzel is easy to strip and not very robust.
If you buy a 'cut strip and wrap' bit they only work with Tefzel.
Go for 'modified wrap' tools (adds a couple of turns of insulated wire
at the start of a wrap, gives greater strength to a joint).
Your bus wiring strategy needs to take account of the fact that you get
a limited number of joints on a single pin - three typically, unless you
go for short pins.
Posted Via Nuthinbutnews.Com Premium Usenet Newsgroup Services
----------------------------------------------------------
** SPEED ** RETENTION ** COMPLETION ** ANONYMITY **
----------------------------------------------------------
http://www.nuthinbutnews.com
Paul Burke
>
>> I want to do make some prototype microcontroller boards with an 8051,
>> memory, latches, ... I think there will be about 10 ICs. I am new to
>> wire wrapping and need to purchase the tools, wire, and prototyping
>> board.
>>
>> Can someone give me advice on what tools, prototyping boards to use?
>>
As one who has done a lot of wirewrapping in my time, my advice would be
not to do it. Prototype PCBs are now much more cheaply and quickly
available than in the past, and you are unlikely to get bad wraps or
signal problems. Wirewrapping gear is expensive, the less expensive the
slower it is. Use a simple free PCB package, go to PCB Pool or PCB Train
(in the UK- local alternatives otherwise), and you are also free to use
SM components in the design without clumsy and terrifically expensive
adapters.
I still use it occasionally if I need a one off very simple (slow,
coarse analog or single- chip microcontroller) test circuit in a blazing
hurry, but then I've got all the gear lying idle already.
Paul Burke
Luhan Monat
Since I've been doing it so long, I forgot to mention. Use precut and
stripped #30 Kynar wire in a variety of colors. I buy them with a
different color for each length. Using a standard 'modified wrap' tool,
you can actually get about 5 wires on a '3 level' wirewrap socket. If
you need more, cut off up to half of the 1 inch bare end, 5 turns of
wrap is plenty good most of the time - thats a total of 20 contacts.
The big advantage of wirewrap over circuit boards, is that is both good
as a prototype (allowing experimentation and changes), and works as
reliably as a soldered board (some claim more so), over long periods of
time. I have never seen one fail - even after 30 years.
Good luck,
Guy Macon
dave garnett wrote:
>I'd advise using a protoboard with a collander ground, and soldering
>ground pins directly to it. You will save yourself untold trouble !
>Similarly you can solder your decoupling caps directly between ground
>and the Vcc pin.
Good asvice. Use a small clip-on heat sink to stop the solder from
wicking up the pin.
>Lots of colours will help you follow wiring more easily. You used to be
>able to get little labels which pushed onto dips and gave you the pin
>numbers - saves hours of counting.
Find them here:
http://www.action-electronics.com/pdww.htm#Id
http://www.wassco.com/socketwrapid.html
http://www.abra-electronics.com/catalog/sockets/8id.html
Another tip: if you are connecting many pins together and go 1 to 2,
2 to 3, 3 to 4, 4 to 5, 5 to 6, you end up ripping the whole thing
out when you discover that 2 is in the wrong place. If instead you
go 1 to 2, 3 to 4, 5 to 6 then 2 to 3, 4 to 5, you only have to redo
three wires to correct an error.
Guy Macon
Paul Burke wrote:
>As one who has done a lot of wirewrapping in my time, my advice would be
>not to do it. Prototype PCBs are now much more cheaply and quickly
>available than in the past, and you are unlikely to get bad wraps or
>signal problems. Wirewrapping gear is expensive, the less expensive the
>slower it is. Use a simple free PCB package, go to PCB Pool or PCB Train
>(in the UK- local alternatives otherwise), and you are also free to use
>SM components in the design without clumsy and terrifically expensive
>adapters.
In my opinion, this depends a lot on your experience. If your skills
are such that the circuit is right the first time, buy a PCB. If
your skills are such that you end up rewiring again and again as you
learn, use wire wrap.
Ken Smith
<smail...@yahoo.com> wrote:
>I want to do make some prototype microcontroller boards with an 8051,
>memory, latches, ... I think there will be about 10 ICs. I am new to
>wire wrapping and need to purchase the tools, wire, and prototyping
>board.
For most things, the soldered prototyping PCB patterns are a better option
than wire wrapping. With the soldered stuff, the board ends up larger in
the horz dimemsions but thinner. All the connections are made on the
component side so there is no constant flipping of the PCB.
Some years back DEC made (had made for them) a wire wrapped socket where
the pins came up on the component side. This made for very easy wire
wrapping.
I prefer a hand wrapping tool to an electric gun. In the hands of a
skilled person, the gun is very fast but it takes more care to use.
There are wire wrap pins you can force or solder into hole of the board.
These make very good tie points to help connecting things up. Power and
ground connections often need extra tie points.
Extra tie points can also be used to modularize the circuit. If the
system breaks naturally into sections, making the connection between
sections via a row of pins can help to keep the changes managable when a
fault is found.
Buy good quality sockets. Cheap sockets are a false economy.
You want to run your power and ground wires as a grid. Each chip should
have a bypass of a 0.01 or 0.1uF. Line driver chips should have two
bypasses, one pointing North and the other pointing South so that they
connect to two different ground runs.
Twisted pairs of wire wrap wire are a good way to carry a sensitive signal
from place to place. It can also be helpful if you have a bus such as a
data bus and command signal running near each other.
--
--
kens...@rahul.net forging knowledge
John Larkin
My advice: don't wire-wrap. If your time is worth anything, you'll
come out way ahead laying out a PC board and having four or so fabbed
by AP Circuits or somebody for less than $100 total.
You said "some boards" not just one. There's nothing more tedious than
wrapping the same thing multiple times.
John
Luhan Monat
Wire-wrap the first one. Make sure it works. Have boards made for the
rest (http://www.expresspcb.com}. Reworking PC boards is also tedeous.
John Larkin
>Wire-wrap the first one. Make sure it works. Have boards made for the
>rest (http://www.expresspcb.com}. Reworking PC boards is also tedeous.
Design it, think about it carefully, and fab boards. It's good
discipline for real life. We never breadboard entire circuits; we go
directly to multilayer PCBs, and most of them are sellable first pass.
Breadboarding teaches a number of bad habits. Careless design is
self-reinforcing.
John
Luhan Monat
Hey, I've done that for purely digital circuits. For analog and mixed
signal stuff, no way!
Terry Given
hear hear!
Its amazing what a good production department can do with rework :)
a contract mfg we used in NH showed us a rework they did for one
customer, who forgot a wire under a 512-pin BGA. So these guys (alas,
cant remember name) whipped the BGA off, re-balled it, soldered a dangly
wire onto the BGA then replaced it on the PCB. 100% success rate, almost
1000 boards. wow.
an OEM ups I got lumbered with had 90 rework mods. It took longer to
rework the pcb than to assemble it in the first place. So we built
17,000 like that :) [long story why we didnt re-spin the pcb]
careful design + careful spice + careful layout = its close enough to
working that we can make it go.
Cheers
Terry
John Larkin
>John Larkin wrote:
>
>> On Wed, 18 May 2005 10:20:30 -0700, Luhan Monat <x@y.z> wrote:
>>
>>
>>
>>>Wire-wrap the first one. Make sure it works. Have boards made for the
>>>rest (http://www.expresspcb.com}. Reworking PC boards is also tedeous.
>>
>>
>>
>>
>> Design it, think about it carefully, and fab boards. It's good
>> discipline for real life. We never breadboard entire circuits; we go
>> directly to multilayer PCBs, and most of them are sellable first pass.
>>
>> Breadboarding teaches a number of bad habits. Careless design is
>> self-reinforcing.
>>
>> John
>>
>>
>>
>
>Hey, I've done that for purely digital circuits. For analog and mixed
>signal stuff, no way!
Oh, come on, take a walk on the wild side.
John
Luhan Monat
Hey John, did you notice what the this thread is called???
Game, set, and match.
Rich Webb
Everyone has mentioned a modified wrap tool but nobody listed a specific
one. Get the OK Industries WSU-30M (these guys have it cheaper than
Digikey): http://www.web-tronics.com/wrto30awgmow.html for 30 gauge
Kynar wire. I've had one for <cough-several-cough> decades and still
reach for it occasionally.
You can use pre-stripped wire but it's easy enough to cut and strip from
a bulk spool. Use the tool above to do the stripping and use your finger
joint as a quick'n'dirty gauge for the length to strip. Goes fast.
--
Rich Webb Norfolk, VA
Guy Macon
There are times when it is impossible to know what a circuit will do
without breadboarding and modification. For example, I once designed
a system that measured hydrocarbons in exhaust gasses by injecting them
into a Hydrogen-Oxygen flame that contained two gold plated electrodes,
applying 1,000 volts across the electrodes, and measured the resulting
microamp current. I didn't know how many stages of filtering I would
need to handle turbulance and still meet the response spec until I did
a handwired prototype and took some measurements.
Richard H.
> Everyone has mentioned a modified wrap tool but nobody listed a specific
> one. Get the OK Industries WSU-30M (these guys have it cheaper than
> Digikey): http://www.web-tronics.com/wrto30awgmow.html for 30 gauge
> Kynar wire. I've had one for <cough-several-cough> decades and still
> reach for it occasionally.
Hmmm... double-ended?? That'd seem awkward to spin one-handed.
I got one of these when Radio Shack still had vacuum tube testers
in-store... it still works great.
http://www.radioshack.com/product.asp?catalog%5Fname=CTLG&product%5Fid=276-1570
The head spins freely, making it pretty low-fatigue to use one-handed;
stripper stores in the handle. It looks like they still stock it
in-store too.
Richard
John Larkin
True. Sometimes we breadboard a small circuit, like an oscillator or a
GaAs fet switch, or we lay out a little board to test a part when the
datasheet isn't clear or entirely believable. For example, we use a
lot of microwave parts in the time domain, where there's little or no
data. But there's no point in breadboarding an entire product, or even
part of one, when the elements are understood and when the datasheets
are clear.
Straightforward analog and digital and uP circuits will just work if
you're careful. Breadboarding is a pernicious habit that encourages
sloppy thinking and can let a lot of bugs and marginal designs sneak
through to production. The Space Shuttle took off under full power,
with a full crew, and flew to orbit on its first powered flight. The
A380 took off and flew at altitude for hours, first time off the
ground.
John
cbarn...@aol.com
handed.
Keith Williams
says...
> John Larkin wrote:
> > On Wed, 18 May 2005 10:20:30 -0700, Luhan Monat <x@y.z> wrote:
> >
> >
> >
> >>Wire-wrap the first one. Make sure it works. Have boards made for the
> >>rest (http://www.expresspcb.com}. Reworking PC boards is also tedeous.
> >
> >
> >
> >
> > Design it, think about it carefully, and fab boards. It's good
> > discipline for real life. We never breadboard entire circuits; we go
> > directly to multilayer PCBs, and most of them are sellable first pass.
> >
> > Breadboarding teaches a number of bad habits. Careless design is
> > self-reinforcing.
> >
> > John
>
> hear hear!
>
> Its amazing what a good production department can do with rework :)
>
> a contract mfg we used in NH showed us a rework they did for one
> customer, who forgot a wire under a 512-pin BGA. So these guys (alas,
> cant remember name) whipped the BGA off, re-balled it, soldered a dangly
> wire onto the BGA then replaced it on the PCB. 100% success rate, almost
> 1000 boards. wow.
Sanmina-SCI in Salem NH?
--
Keith
Rich Grise
It still isn't worth it. Just use regular sockets, and solder wires
to them. It's not quite as quick as wire-wrapping, but it's nowhere
near as kloogey-looking. Can you even still get wire-wrap stuff?
Somebody was asking about soldering to a vectorboard. This isn't chips,
but hopefully shows that it's doable:
http://www.neodruid.org/images/PWM-prototype-top.jpg
http://www.neodruid.org/images/PWM-prototype-bottom.jpg
Ew. In the close-up, it looks like pretty crappy work. But WTF, I got
paid $150.00 for six of the little bastards, and the guy who ordered
them was buying the drinks.
Why the big fat (#18) wire around the perimeter of the board? EMI
protection. It's a motor speed controller for a spool gun for a
battery-powered MIG welder; it mounts right alongside the ~60A weld
current cable. I first did it with a 555, and it died instantly, so the
primary consideration was robustness. The main current switch isn't shown
- it mounts to the heatsink that the little board gets bolted to. Its
leads poke up through the three enlarged holes by the turret terminal at
the bottom. It was also fun cutting notches in the heatsink fins with an
endmill bit in a drill press, to accommodate the mounting screws. He also
showed me how to use a tap in a hand drill.
Yes, it's a Radio Shack perfboard. :-)
Cheers!
Rich
Rich Grise
> I prefer a hand wrapping tool to an electric gun. In the hands of a
> skilled person, the gun is very fast but it takes more care to use.
I once saw a "wire-wrap gun" that was just a hand wire-wrapper clamped
in an electric eraser.
Cheers!
Rich
Rich Grise
> Rich Webb wrote:
>> Everyone has mentioned a modified wrap tool but nobody listed a specific
>> one. Get the OK Industries WSU-30M (these guys have it cheaper than
>> Digikey): http://www.web-tronics.com/wrto30awgmow.html for 30 gauge
>> Kynar wire. I've had one for <cough-several-cough> decades and still
>> reach for it occasionally.
>
> Hmmm... double-ended?? That'd seem awkward to spin one-handed.
The other end is an unwrapper, and actually, is a very convenient
handle for spinning.
Cheers!
Rich
Joerg
>>I hate to disagree with you Tim but a small project like this is ideal
>>for wire wrap. The whole job would take only 2 hours or so to do once
>>the circuit has been designed. Reliability is also good, wrap joints
>>are good for at least 10 years allthough I have 20 yearold boards that
>>still work fine.
>
> It still isn't worth it. ...
Agree. Also, 10 years isn't a whole lot of life. Even 20 isn't for some
stuff. When wire wrap has turned 30 and the posts look black, oh boy...
> ...Just use regular sockets, and solder wires
> to them. It's not quite as quick as wire-wrapping, ...
It can be as quick. When in a rush I use wire where the lacquer
insulation melts while soldering. For very speedy work you have to learn
how to simultaneously hold the spool and wire cutters in one hand,
solder iron in the other. But that's no harder than learning to eat with
chop sticks. Sometimes I clamp a contraption to the lamp above me and
roll off the spool from there.
Regards, Joerg
Tom Woodrow
Tom Woodrow
smail...@yahoo.com wrote:
> I want to do make some prototype microcontroller boards with an 8051,
> memory, latches, ... I think there will be about 10 ICs. I am new to
> wire wrapping and need to purchase the tools, wire, and prototyping
> board.
>
> Can someone give me advice on what tools, prototyping boards to use?
>
> Also best practices advice would be nice. I realize the importance of
> decoupling caps, but what about things like:
>
> choosing boards with ground planes, power buses, and connecting power
> and ground pins.
>
> what wire gages to use. Same for power, ground, and signal?
>
> Any tricks for connecting data and address buses?
>
> Specific manufacturers and part #s for tools and boards would be great.
>
> Thanks in advance,
>
> Scott
>
Rich Grise
I took one of these tools, a WSU-30M:
http://www.okindustries.com/products/4.1.1.10.htm
and ground off the rivet holding the little stripper blade, and clamped
it in an X-acto handle. It was handier than a rat, especially for doing
buses. I'd strip a couple of inches off the end of a spool of wrap
wire, and solder just the tip end to the one point, then go to the
insulation, and cut it at the length I needed to go to the next
point; slide the insulation up to the joint, route it, and solder
the next point; then again for the next one, and so on. So, essentially,
it was one wire with gaps in the insulation. You can do it the other
direction, but if it's a long run, sometimes the insulation doesn't
slide along the whole length of wire very well.
Cheers!
Rich
Joerg
> I took one of these tools, a WSU-30M:
> http://www.okindustries.com/products/4.1.1.10.htm
> and ground off the rivet holding the little stripper blade, and clamped
> it in an X-acto handle. It was handier than a rat, especially for doing
> buses. I'd strip a couple of inches off the end of a spool of wrap
> wire, and solder just the tip end to the one point, then go to the
> insulation, and cut it at the length I needed to go to the next
> point; slide the insulation up to the joint, route it, and solder
> the next point; then again for the next one, and so on. So, essentially,
> it was one wire with gaps in the insulation. You can do it the other
> direction, but if it's a long run, sometimes the insulation doesn't
> slide along the whole length of wire very well.
That's pretty clever.
In the days when I had to debug wrapped boards (not mine, I'll never WW
a board or anything else for that matter) I used mostly hand tools like
these because the techs had the electric wrap guns. Except late at
night, then we could have the wrap guns. Got a lot of blisters from the
repetitive motion.
For daisy chains I also use the melting-insulation wire. It's pretty
easy, no tools.
BTW, the modified wrap they show can cause some grief. Many people don't
realize that they create a small inductor of unknown properties that
way. Usually ok for old TTL but for ECL that's another story.
Regards, Joerg
keith
> Hello Rich,
>
>> I took one of these tools, a WSU-30M:
>> http://www.okindustries.com/products/4.1.1.10.htm
>> and ground off the rivet holding the little stripper blade, and clamped
>> it in an X-acto handle. It was handier than a rat, especially for doing
>> buses. I'd strip a couple of inches off the end of a spool of wrap
>> wire, and solder just the tip end to the one point, then go to the
>> insulation, and cut it at the length I needed to go to the next
>> point; slide the insulation up to the joint, route it, and solder
>> the next point; then again for the next one, and so on. So, essentially,
>> it was one wire with gaps in the insulation. You can do it the other
>> direction, but if it's a long run, sometimes the insulation doesn't
>> slide along the whole length of wire very well.
>
> That's pretty clever.
>
> In the days when I had to debug wrapped boards (not mine, I'll never WW
> a board or anything else for that matter) I used mostly hand tools like
> these because the techs had the electric wrap guns. Except late at
> night, then we could have the wrap guns. Got a lot of blisters from the
> repetitive motion.
Out techs hated it when we used their tools. We had our own and had to
take care of 'em. Touching their 30ga strippers was a crime. Using them
as pliers was a capital offense!
> For daisy chains I also use the melting-insulation wire. It's pretty
> easy, no tools.
Kynar? We were required to use teflon wire. Kynar resulted int too many
"cold" (really cut insulation) connections.
> BTW, the modified wrap they show can cause some grief. Many people don't
> realize that they create a small inductor of unknown properties that
> way. Usually ok for old TTL but for ECL that's another story.
Ok, but the "modified" wrap isn't in any way "modified". It is the
standard wire-wrap connection. Using their "regular" wrap results in
broken wires down the road. Gardner-Denver always made the "modified"
wrap. ...and yes WW, even with the "modified wrap" was used for *years*
with ECL. Mainframes boards used tons of WW, and were all ECL. I used
WW with ECL in the 70s, and with 74Fxx and 74ASxx in the 80s. It
works, but it takes some care. I had several boards with 6K wires (my
techs loved the OT ;-).
--
Keith
Joerg
> Out techs hated it when we used their tools. We had our own and had to
> take care of 'em. Touching their 30ga strippers was a crime. Using them
> as pliers was a capital offense!
Oh yes, I remember how closely the techs guarded their tools.
> Ok, but the "modified" wrap isn't in any way "modified". It is the
> standard wire-wrap connection. Using their "regular" wrap results in
> broken wires down the road. Gardner-Denver always made the "modified"
> wrap. ...and yes WW, even with the "modified wrap" was used for *years*
> with ECL. Mainframes boards used tons of WW, and were all ECL. I used
> WW with ECL in the 70s, and with 74Fxx and 74ASxx in the 80s. It
> works, but it takes some care. I had several boards with 6K wires (my
> techs loved the OT ;-).
I should have said ECL in an ultrasound machine. There half a nanosecond
deviation can mess up the whole beamformer performance. So one day I got
pretty tired of debugging other engineer's WW boards and asked the techs
to solder one. Should have seen their looks.... but then again they were
paid overtime for it so their grudge only lasted until payday. That
fixed all the delay inconsistencies.
BTW, back at the university I soldered the backplanes of some dead
mainframes that had been written off as too expensive to repair. After
that they raised like Phoenix from the ashes.
Regards, Joerg
Terry Given
a *rose* by any other name?
>
> Regards, Joerg
>
Cheers
Terry
Terry Given
The name doesnt ring any bells. But shit, their techs could solder.
Cheers
Terry
Keith Williams
says...
Where about in NH? I've worked with Sanmina in Salem (formerly Hadco)
and a company in Dover NH, though I can't place their name. Both
companies were great to work with.
--
Keith
Joerg
>> BTW, back at the university I soldered the backplanes of some dead
>> mainframes that had been written off as too expensive to repair. After
>> that they raised like Phoenix from the ashes.
>
> a *rose* by any other name?
Um, yes, should have said rose from the ashes. I guess I should allow
the brain to adjust for a minute after switching languages...
Regards, Joerg
keith
> Hello Keith,
>
>> Out techs hated it when we used their tools. We had our own and had to
>> take care of 'em. Touching their 30ga strippers was a crime. Using them
>> as pliers was a capital offense!
>
> Oh yes, I remember how closely the techs guarded their tools.
>
>> Ok, but the "modified" wrap isn't in any way "modified". It is the
>> standard wire-wrap connection. Using their "regular" wrap results in
>> broken wires down the road. Gardner-Denver always made the "modified"
>> wrap. ...and yes WW, even with the "modified wrap" was used for *years*
>> with ECL. Mainframes boards used tons of WW, and were all ECL. I used
>> WW with ECL in the 70s, and with 74Fxx and 74ASxx in the 80s. It
>> works, but it takes some care. I had several boards with 6K wires (my
>> techs loved the OT ;-).
>
> I should have said ECL in an ultrasound machine. There half a nanosecond
> deviation can mess up the whole beamformer performance.
Ok, a half nanosecond translates to 2GHz. No, I wasn't doing 2GHz in the
70s. ...and no, I don't wire-wrap such circuits today. ;-)
It seems your engineers didn't understand the requirements (or physics).
BTW, I don't know how you're maintinging 500pS with SSI anyway, ECL or not.
> So one day I got
> pretty tired of debugging other engineer's WW boards and asked the techs
> to solder one. Should have seen their looks.... but then again they were
> paid overtime for it so their grudge only lasted until payday. That
> fixed all the delay inconsistencies.
I had one case where the technician decided that he knew how to wire power
and ground better than I. He daisy-chained *all* the connections. He
then went on to exacerbate his stupidity by demonstrating that they all
checked out on a continutiy tester. I went rip-shit, but a couple of
weeks of work were already down the drain. He didn't work for me anymore.
Instead, I got a new kid and trained him right (he ended up as an engineer
in the research division).
OTOH, we also had an engineer that argued grounds. He couldn't undestand
why "his" ( he took over responsibility) analog stuff looked so ratty on
the scope, but it still worked, sorta. His scope ground was connected to
the bat handle of a toggle switch on the test fixture. It took an hour to
convince him that his "ground" wasn't *ground*.
> BTW, back at the university I soldered the backplanes of some dead
> mainframes that had been written off as too expensive to repair. After
> that they raised like Phoenix from the ashes.
As long as it's junk, why not try?
--
Keith
Joerg
>>I should have said ECL in an ultrasound machine. There half a nanosecond
>>deviation can mess up the whole beamformer performance.
>
> Ok, a half nanosecond translates to 2GHz. No, I wasn't doing 2GHz in the
> 70s. ...and no, I don't wire-wrap such circuits today. ;-)
>
> It seems your engineers didn't understand the requirements (or physics).
> BTW, I don't know how you're maintinging 500pS with SSI anyway, ECL or not.
What mattered was phase shift between dozens of transmit pulses. 500psec
can start messing things up. The frequency was much lower.
> I had one case where the technician decided that he knew how to wire power
> and ground better than I. He daisy-chained *all* the connections. He
> then went on to exacerbate his stupidity by demonstrating that they all
> checked out on a continutiy tester. I went rip-shit, but a couple of
> weeks of work were already down the drain. He didn't work for me anymore.
> Instead, I got a new kid and trained him right (he ended up as an engineer
> in the research division).
>
> OTOH, we also had an engineer that argued grounds. He couldn't undestand
> why "his" ( he took over responsibility) analog stuff looked so ratty on
> the scope, but it still worked, sorta. His scope ground was connected to
> the bat handle of a toggle switch on the test fixture. It took an hour to
> convince him that his "ground" wasn't *ground*.
Oh boy. I have seen that too, but not quite this bad. Except for the guy
that wired all his stuff up on white board.
>>BTW, back at the university I soldered the backplanes of some dead
>>mainframes that had been written off as too expensive to repair. After
>>that they raised like Phoenix from the ashes.
>
> As long as it's junk, why not try?
They became remarkably reliable.
Regards, Joerg
Keith Williams
notthis...@removethispacbell.net says...
> Hello Keith,
>
> >>I should have said ECL in an ultrasound machine. There half a nanosecond
> >>deviation can mess up the whole beamformer performance.
> >
> > Ok, a half nanosecond translates to 2GHz. No, I wasn't doing 2GHz in the
> > 70s. ...and no, I don't wire-wrap such circuits today. ;-)
> >
> > It seems your engineers didn't understand the requirements (or physics).
> > BTW, I don't know how you're maintinging 500pS with SSI anyway, ECL or not.
>
> What mattered was phase shift between dozens of transmit pulses. 500psec
> can start messing things up. The frequency was much lower.
I understand, but even ECL has a Tpd variation more than 500pS. I
designed an ECL clock generator in the mid '70s that had 64
programmable clocks with 1nS edge resolution over a 25nS cycle. I had
to use tri-lead (twin-lead with an extra ground) to tune the delays in
each channel. It was a real PITA. The control logic was all wire
wrapped, but the clock drivers were on PCBs.
> > I had one case where the technician decided that he knew how to wire power
> > and ground better than I. He daisy-chained *all* the connections. He
> > then went on to exacerbate his stupidity by demonstrating that they all
> > checked out on a continutiy tester. I went rip-shit, but a couple of
> > weeks of work were already down the drain. He didn't work for me anymore.
> > Instead, I got a new kid and trained him right (he ended up as an engineer
> > in the research division).
> >
> > OTOH, we also had an engineer that argued grounds. He couldn't undestand
> > why "his" ( he took over responsibility) analog stuff looked so ratty on
> > the scope, but it still worked, sorta. His scope ground was connected to
> > the bat handle of a toggle switch on the test fixture. It took an hour to
> > convince him that his "ground" wasn't *ground*.
>
> Oh boy. I have seen that too, but not quite this bad. Except for the guy
> that wired all his stuff up on white board.
"white board"?
> >>BTW, back at the university I soldered the backplanes of some dead
> >>mainframes that had been written off as too expensive to repair. After
> >>that they raised like Phoenix from the ashes.
> >
> > As long as it's junk, why not try?
>
> They became remarkably reliable.
Properly done WW is as reliable as solder. I certainly wasn't
qualified to WW customer deliverables, but someone wrapped many
thousands of wires in each. Of course, GD machines did most of it. ;-)
--
Keith
Guy Macon
Keith Williams wrote:
>Joerg says...
>
>> Except for the guy that wired all his stuff up on white board.
>
>"white board"?
Commonly used term for white plastic borads with holes in them,
designed to allow solderless breadboarding by plugging wires
and components into the holes. Pictures here:
[ http://www.globalspecialties.com/solderless.html ].
[ http://www.circuitspecialists.com/level.itml/icOid/8099 ]
Has a rather undeserved bad reputation because of morons who
force thick-lead components into the holes and weaken the
spring contacts. Later another engineer has intermittent
contacts and concludes that whiteboards are all junk.
Some engineers/technicians make whiteboard circuits that are all
neat parallel runs of color-coded wires with 90 degree angles
hugging the board. Others like to make a gentle arc from one point
to another. Some simply make a rat's nest.
Whiteboards are not suitable for high speed or low capacitance
circuits and are clumsy for large designs, but they are great for
prototyping small chunks of a design before adding them to a PWB.
If you want to try something *really* interesting, try a SchmartBOARD.
[ http://www.schmartboard.com/ ]
[ http://www.schmartboard.com/index.asp?page=products ]
Keith Williams
_see.web.page_@_www.guymacon.com_ says...
>
>
> Keith Williams wrote:
>
> >Joerg says...
> >
> >> Except for the guy that wired all his stuff up on white board.
> >
> >"white board"?
>
> Commonly used term for white plastic borads with holes in them,
> designed to allow solderless breadboarding by plugging wires
> and components into the holes. Pictures here:
> [ http://www.globalspecialties.com/solderless.html ].
> [ http://www.circuitspecialists.com/level.itml/icOid/8099 ]
Ah, I've always called these things "EL sockets". IIRC, EL was the
company that first made these things in the early '70s. I hated them
then. After graduating I never had to look at one again.
>
> Has a rather undeserved bad reputation because of morons who
> force thick-lead components into the holes and weaken the
> spring contacts. Later another engineer has intermittent
> contacts and concludes that whiteboards are all junk.
I don't see their purpose. They may work for some really simple
circuits, but it's not difficult to solder up a bread-board of similar
complexity. I *hate* chasing intermittent connections. Time is
important.
> Some engineers/technicians make whiteboard circuits that are all
> neat parallel runs of color-coded wires with 90 degree angles
> hugging the board. Others like to make a gentle arc from one point
> to another. Some simply make a rat's nest.
Whatever I tried it turned into a rat's nest. I remember losing
compensation on 709s with these things. Ouch! Them things get *hot*.
> Whiteboards are not suitable for high speed or low capacitance
> circuits and are clumsy for large designs, but they are great for
> prototyping small chunks of a design before adding them to a PWB.
Not IMO. You can have 'em.
> If you want to try something *really* interesting, try a SchmartBOARD.
> [ http://www.schmartboard.com/ ]
> [ http://www.schmartboard.com/index.asp?page=products ]
Those are more interesting. Much pricey though.
My last few PCB designs I went straight to PCB. There were a few
circuits that I was concerned about. *Those* worked perfectly. I F'd
up the simple stuff. Though where I did screw up, the wires were all
on the outer planes. Dumb, but easily recoverable.
--
Keith
Joerg
>>What mattered was phase shift between dozens of transmit pulses. 500psec
>>can start messing things up. The frequency was much lower.
>
> I understand, but even ECL has a Tpd variation more than 500pS. I
> designed an ECL clock generator in the mid '70s that had 64
> programmable clocks with 1nS edge resolution over a 25nS cycle. I had
> to use tri-lead (twin-lead with an extra ground) to tune the delays in
> each channel. It was a real PITA. The control logic was all wire
> wrapped, but the clock drivers were on PCBs.
I found ECL to be better than that but we soon abandoned it anyway for
discretes and FPGA because it gulped so much power.
>>Oh boy. I have seen that too, but not quite this bad. Except for the guy
>>that wired all his stuff up on white board.
>
> "white board"?
The real name slipped my mind because I never ordered any and never
will. It's the white experimenter's boards where you stick components into.
> Properly done WW is as reliable as solder. ...
Yes, it can be. What I saw in these mainframes was that it all turned
black over time. On prototype it was even worse because people touched
it with their hands a lot. With solder that wasn't a problem.
> I certainly wasn't qualified to WW customer deliverables, ...
I wasn't either but we had techs that received training it WW. They did
a pretty neat job but for boards I never authorized WW.
WW still was much superior to another technique. Forgot what it was
called but you pushed wires into IDC slots. Once in a while you heard a
faint "ping", had no idea where it came from and the board would refuse
to work. That was often a real pain.
Regards, Joerg
Guy Macon
Joerg wrote:
>Yes, it can be. What I saw in these mainframes was that it all turned
>black over time. On prototype it was even worse because people touched
>it with their hands a lot. With solder that wasn't a problem.
Why do you care? Yes, silver oxide is black and tin-lead oxide is
gray, but in neither case is the oxide in the path that the electrons
take - wire-wrap connections are gas-tight. Besides, Silver oxide
conducts electricity just fine.
Mike Monett
Most likely silver sulphide from the sulphur in your fingerprints. It is
conductive, but not quite as good as silver. However, it is soft and
contact pressure may break through the film.
Old WWII transmitters often used silver plate on the tank coils,
especially for VHF and UHF. They used sliding or roller contacts to
adjust the tank. The discoloration had little or no effect on the power
output.
Mike Monett
Guy Macon
In the case of wire-wrap, there isn't even a need for the
film to conduct or for any breaking through of the film.
Every wire-wrap has multiple gas-tight connections in
parallel. Neither oxygen from the air or sulpher from
finger oils can reach the contact point.
Robin
http://freespace.virgin.net/john.hardaker/Musatek/vero.html
The trick is to bend all IC socket-legs "outwards" beforehand (see
photo) otherwise the vero-wire will slip off.
When wiring the MCU sockect,
1) Pre-tin the vero-wire (don't melt the tool tip).
2) wrap a couple of turns around the MCU socket pin and solder it.
3) Then you can run the wire around all other nodes (don't solder
untill board complete).
Wiring up memory is a snip... d0 to d0 to d0 to d0 ad infinitum takes
seconds *and* you don't solder anything until the end.
You can make ten or twenty connections to one node (you can only get
three or four with wire-wrap).
Wire-wrap ends up about an inch thick and looks ugly
Vero-wire only adds 0.1 inch thickness and looks neat.
Cheers
Robin
Guy Macon
Robin wrote:
>
>Vero wire is better:
>
>http://freespace.virgin.net/john.hardaker/Musatek/vero.html
MIL-STD-1130B says that it isn't.
>Wire-wrap ends up about an inch thick and looks ugly
Not when I do it.
>Wiring up memory is a snip... d0 to d0 to d0 to d0 ad
>infinitum takes seconds *and* you don't solder anything
>until the end.
Slit-N-Wrap wire wrapping is faster than vero wire, and
you don't have to solder at all.
Keith Williams
_see.web.page_@_www.guymacon.com_ says...
silver-oxide or sulfate in there. ;-)
--
Keith
Guy Macon
Keith Williams wrote:
>
>Guy Macon <_see.web.page_@_www.guymacon.com_> says...
I believe that if you checked, you would find that the wires were
all silver-plated copper.
Keith Williams
_see.web.page_@_www.guymacon.com_ says...
no-no after many major disasters.
--
Keith
Terry Given
Hadco sounds familiar. This was 1997-2000 when I was in MA.
Cheers
Terry
John Larkin
wrote:
>In article <CEOie.1040$U4.1...@news.xtra.co.nz>, my_...@ieee.org
>says...
>> John Larkin wrote:
>> > On Wed, 18 May 2005 10:20:30 -0700, Luhan Monat <x@y.z> wrote:
>> >
>> >
>> >
>> >>Wire-wrap the first one. Make sure it works. Have boards made for the
>> >>rest (http://www.expresspcb.com}. Reworking PC boards is also tedeous.
>> >
>> >
>> >
>> >
>> > Design it, think about it carefully, and fab boards. It's good
>> > discipline for real life. We never breadboard entire circuits; we go
>> > directly to multilayer PCBs, and most of them are sellable first pass.
>> >
>> > Breadboarding teaches a number of bad habits. Careless design is
>> > self-reinforcing.
>> >
>> > John
>>
>> hear hear!
>>
>> Its amazing what a good production department can do with rework :)
>>
>> a contract mfg we used in NH showed us a rework they did for one
>> customer, who forgot a wire under a 512-pin BGA. So these guys (alas,
>> cant remember name) whipped the BGA off, re-balled it, soldered a dangly
>> wire onto the BGA then replaced it on the PCB. 100% success rate, almost
>> 1000 boards. wow.
>
Just put a via on every ball, used or not.
John
keith
That's what I did. It was easier leaving the via, since it came with the
image.
--
Keith