Revolutionary no-solder wire-wrapping breakthrough!

[cbmeeks]

I read on Jeff Tranter's blog that he's used Krazy Glue on the sockets before. A couple dabs of that on the ends would probably hold the socket in very well and shouldn't be too hard to remove with some solvents. Assuming you don't melt the plastics.

[Dr Jefyll]

In this application Crazy Glue is more of a threat than a benefit, I'd say. Its low viscosity means it can wick its way inside the socket pin cavities, and that's very bad news indeed. Moreover I don't trust it to adhere well, given that the board you're attaching to is not entirely rigid -- there can be a bit of flexing during handling (such as chip extraction). When gluing onto something flexible, you want the glue itself to retain a degree of flexibility. Contact cement is good for this; likewise silicone adhesives such as RTV.

[BigEd]

Sounds like a case for hot glue?

[Bregalad]

Quote:

but I catch them long before I'm done, like noticing that there are already two wraps on a post that should only have one (where did the other one come from?)

And of course when this happens, the bottom connexion is the one that was wrong, leading to the need to remove half-a-dozen correct connexions just to get rid of one wrong connexion.

[cbmeeks]

I'm getting closer and closer to my first wire-wrapped computer. I've wrapped others but never got them far enough along for a real test.

I've taken my time on this one. Nothing is soldered to the board.

However, I doubt I will wire-wrap another computer in the future. I know all of the pros for a "properly" wrapped board but what I'm finding is that the closeness of the components are actually making it difficult for me to wrap. Mainly because of my vision. But also, because of a "blindness" effect from all of the pins and vias. It's kind of like snow blindness. When I look at the pins, I find it's hard to pinpoint the exact pin I need. Plus, pins that are closer than an inch apart are difficult for me.

I know this take practice.

Anyway, I should have a really neat computer finished pretty soon. ;-D

[EugeneNine]

Well back then we all couldn't afford a multi thousand dollar CAD program nor the high end workstation to run it on.

Now I can run kiCad on a 5 year old netbook on Monday and upload and have a PCB to me by Friday for less than the cost of a couple wire wrap socket. times have change a little bit. Part of me wants to WW another project but part of me says its just not cost effective anymore.

[BigDumbDinosaur]

EugeneNine wrote:

Part of me wants to WW another project but part of me says its just not cost effective anymore.

That's like back in the days when I used to make my own spaghetti sauce from scratch. I could still do it, but the fine folks at Barilla have done all the work for me. Now I open a jar of their sauce, dump it in with my meatballs and other ingredients, do some mixing and cooking, and in 15 minutes or less I'm ready to serve spaghetti. Barilla's sauce is as good as what I could make, so why expend an hour or more making it from scratch, not to mention getting several pots dirty in the process?

[GARTHWILSON]

Bregalad wrote:

Quote:

but I catch them long before I'm done, like noticing that there are already two wraps on a post that should only have one (where did the other one come from?)

And of course when this happens, the bottom connexion is the one that was wrong, leading to the need to remove half-a-dozen correct connexions just to get rid of one wrong connexion.

Well, as I said in the same post, when I find a wire went to the wrong post, it's always still on top, so I don't have to unwrap something above it to get to it. I would have seen that the second wrap should have been the first, so the errant first one would be caught before anything was added over it. Really, I don't think I've ever had the problem you say.

EugeneNine wrote:

Well back then we all couldn't afford a multi thousand dollar CAD program nor the high end workstation to run it on.

I got lucky. The tiny company I work(ed) for bought my CAD in 1993, and nobody but me had any interest or knowledge in using it. It could run even on an 8086, although I started with a '286. I quit taking updates when they were no longer helpful. It was only $375 at the time, but was way better than the OrCAD 386 I used at my previous job, which cost six times as much and needed at least a '386 (but not a workstation) to run it.

Quote:

Now I can run kiCad on a 5 year old netbook on Monday and upload and have a PCB to me by Friday for less than the cost of a couple wire wrap socket.

That is a nice thing we have today, that even multilayer boards are affordable for the hobbyist; but WW may still hold an attraction in that you can modify or add to an already built-up circuit much more easily than you can on PCBs. Sometimes I'm torn between the two.

[Dr Jefyll]

GARTHWILSON wrote:

Sometimes I'm torn between the two.

In some cases a hybrid may be attractive. I can imagine creating an SBC, say, with WW sockets soldered in place on a custom PCB.

Connections of a tentative or experimental nature would be made using WW, whereas connections unlikely to become subject to revision would be implemented via the etched traces. Typically these would include the Vcc and Gnd connections (and their bypass caps) and probably also the data and address buses.

And the divisions needn't be cast in stone. You might choose to use WW for a few of the non-experimental connections simply to increase density (as if you had more PCB layers than actually exist). On the flip side of the coin, if you initially used an etched trace for a given connection but later decided it needed to be modified, just cut the etched trace and install an alternative connection using WW.

[EugeneNine]

Or build a backplane based with card slots spaced wide enough you can use WW boards and then do soldered boards in some slots as well. If you make it modular enough.

I'm getting ahead of myself though, I'm at the point where I need to be able to burn roms.

[BigDumbDinosaur]

EugeneNine wrote:

Or build a backplane based with card slots spaced wide enough you can use WW boards and then do soldered boards in some slots as well. If you make it modular enough.

Running the 65xx buses onto a backplane often significantly reduces the maximum speed at which the system can run and remain stable. Transmission line effects caused by all that wiring on the backplane can sabotage timing in a variety of ways and are not easily solved by the average hobbyist. We generally don't recommend such an approach.

A "mezzanine" configuration, used by several of us to attach boards, tends to reduce the problem of transmission line effects and usually demonstrates better stability at higher clock rates. However, mezzanine construction may be less convenient for experimentation.

[EugeneNine]

Yea, there is that issue too

Most anything I build tends to be smaller and portable anyway as I don't have any kind of bench space.