Prototyping PCB patterns

53 views
Skip to first unread message

Douglas Miller

unread,
Nov 8, 2022, 9:23:56 PM11/8/22
to SEBHC
A posting on another forum recently got me thinking about how we did
prototyping back in the old days at Magnolia Microsystems. Brad Gjerding
had a pattern he used for wire-wrap boards, which he used for his NASA
contracts. He also made some general-purpose PCBs with that pattern that
we used for prototyping. Since I've never seen a pattern quite like it,
and found it very useful, I have created a kicad PCB layout as an
example. Attached are some 3D views plus a photo of some old pieces of
one of those boards.

The pattern is intended for wire-wrap, but could probably be modified to
suit other types of prototyping. It can easily accommodate 0.3" and 0.6"
DIP sockets (0.6" only in certain columns) and has a lot of power/ground
foil and is setup for bypass and filter caps. The DIP power/ground
traces default to DIP-16, but those can be easily cut as needed and
power rails are close by for easy connection to any DIP package.

The benefits of the pattern may not be obvious until you've used it, but
I throw it out here as something to think about. I've stored the kicad
project at https://github.com/durgadas311/MmsCpm3/tree/master/pcb/bkg-77000.
bkg-77000-b.png
bkg-77000-f.png
bkg-77000-f-pop.png
bkg-77000-sample.jpg

Joseph Travis

unread,
Nov 16, 2022, 2:03:46 PM11/16/22
to se...@googlegroups.com
It's a very cool / useful pattern.  Are there any plans for an H8 and/or H89 version?

--
You received this message because you are subscribed to the Google Groups "SEBHC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sebhc+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/2c7edb5a-f900-ade0-5777-31dbf3ea896e%40gmail.com.

glenn.f...@gmail.com

unread,
Nov 16, 2022, 2:17:47 PM11/16/22
to se...@googlegroups.com

Ed Aumiller, the guy from Virginia from whom I collected some hardware a few years ago, used to use H8-7 prototyping boards as the foundation for mounting protoboards (in his case he used wire wrap boards).  He would order the blank H8-7 PCBs from Heath, cut out a rectangular opening and mount the wire wrap board, then use the bus interface logic on the board to interface to the H8.  A little crude but effective. He built a number of prototypes this way…

 

Pix of one of his boards from my rescue:

https://photos.app.goo.gl/njX7d4nQfuAKwog7A

https://photos.app.goo.gl/vg6ATLaSiqV9rC9WA

Douglas Miller

unread,
Nov 16, 2022, 2:28:14 PM11/16/22
to se...@googlegroups.com

I was presenting the pattern as the potential for H8/H89 proto boards, or for certain boards that have an area set aside for prototyping (like the WizNet board) or for boards that don't fill the entire full-sized H8 board. The one board I showed was just what Brad had created for general experimental use, but the pattern could be used to fill almost any size board or area. I first encountered the pattern on full-sized tack-mount boards for a multi-CPU 6800 system, something like 18"x18".

norberto.collado koyado.com

unread,
Nov 16, 2022, 3:42:30 PM11/16/22
to se...@googlegroups.com

Terry S. and I already brainstorm this idea of using a removable proto board on the H8 board as Glenn pictures described. I think this is the way to go to re-use such board as needed. This is in the To-Do-List.

 

Also, we need to consider that the pattern should support PLCC44 DIP sockets.

 

Thanks,

Norberto

Douglas Miller

unread,
Nov 16, 2022, 6:05:39 PM11/16/22
to se...@googlegroups.com

Yes, adding space for a PLCC44 through-hole socket would be good, although then one also wonders about PLCC32 as well, and maybe PLCC68. Perhaps the answer to PLCC packages would be the PLCC-to-DIP adapters, which then would fit into the pattern as-is. Depends on how much PLCC work we think we need.

The PCB I showed was really just a "demo", to visualize the concept. We can make a PCB of any form-factor using that pattern (or whatever pattern is needed). Since there are no boards in existence, we might as well design the board to be what we want.

norberto.collado koyado.com

unread,
Nov 16, 2022, 6:31:23 PM11/16/22
to se...@googlegroups.com

We were considering the attached proto-card to be installed on the H8 wire-wrap board or something similar (picture attached).

 

I’m more on having the same pattern on the whole board with the +5V and the GND on the outside pattern. Something similar to this:

 

Vector_Eurocard_Layout.pdf

Douglas Miller

unread,
Nov 16, 2022, 6:59:02 PM11/16/22
to se...@googlegroups.com

Understood. As I stated earlier, without actually using the pattern I suggested it is difficult to fully appreciate its benefits. I personally prefer a pattern that supplies a strong and stable VCC and GND (which includes easy/clean addition of bypass caps). I'm not sure how much of that is placebo and how much can actually be measured. Part of that depends on whether one is throwing something together "quick and dirty" or whether they actually plan on running it indefinitely. I would argue that if one is building a true prototype - with the intent of transferring to PCB "finished product" - that you don't want to be chasing problems you'd never have on the final PCB. I've been forced to use a lot of "off the shelve" patterns over the years and they always frustrate me. But, I'm not likely to use either design of prototype board, so the decision needs to be left to others. I just present it as a pattern that I prefer over all others that I've seen (so far).

Terry Smedley

unread,
Nov 16, 2022, 7:26:56 PM11/16/22
to SEBHC
I thought it was important to choose a commercially-available board size, like a 3U Eurocard, for the insert.  That way, you could roll your own with something like the pattern that Douglas suggested, or you could use one of the dozens of commercial 3U designs if one of them fit your needs.  I suspect most prototypers will have different preferences, some of which might require a custom pattern.  But I didn't want my pattern choice to be limited to a custom design.

I agree with Douglas about the strong Vcc and GND paths.  The commercial patterns I have favored (like the Vector 4614) have parallel ground and power buses running underneath each column of ICs.  

At least in my world, the pattern I use is a function of the type of circuit I'm trying to prototype.  For analog-heavy circuits I'd probably choose a different pattern (with 3 or 5-hole pads) than, say, a bussed memory design (with zig-zag  traces like a BusBoard).  

None of this is an argument for or against any particular pattern.  What I had discussed with Norberto was a "pattern neutral" platform, where builders could  easily and economically choose whatever pattern they preferred.  All the grunt-work bus interface infrastructure would be handled on the carrier board, allowing builders to concentrate on their specific project using whatever pattern they chose.  This is the same idea that Glenn described from Ed Aumiller, made a little "prettier" by a purpose-built carrier board.

tas

norberto.collado koyado.com

unread,
Nov 16, 2022, 7:28:09 PM11/16/22
to se...@googlegroups.com

The holes pattern described here are for using normal IC sockets for point to point soldering. For Wire-wrapping the holes needs to be bigger.

 

Here is an example of a wire-wrap IC sockets as holes are bigger. I used this board to develop/test the H37/H67 controllers before production.

 

Douglas Miller

unread,
Nov 16, 2022, 8:56:00 PM11/16/22
to se...@googlegroups.com

A modular design is interesting, but I worry about the strength of the base board - I imagine, with such a large cut-out, there needs to be some sort of strengthening bars horizontally?

I guess one could design this such that the mounting screws actually doubled as making the power and ground connections, thereby providing the best possible conductivity. Possibly, mounting the "secondary" prototype board might also provide the mechanical strength needed. Depends on whether the secondary PCB mounts overlapping the base board, or fits flush inside the cut-out (harder to do mechanically).

Would there be a set of pads for the H8 bus signals? Or something more formal like a connector (i.e. the base board is sort of like a bus extender)?

What's the cost of a large PCB with a very large cut-out? Seems like a lot of board material is wasted. I wonder if it is more cost-effective to just have several layouts available - all with the same H8-bus-interface circuitry. Sort of "drop-in" patterns but at the gerber level rather than physical PCBs.

Terry Smedley

unread,
Nov 16, 2022, 9:30:36 PM11/16/22
to SEBHC
My thought was simple (perhaps simple-minded).

No cutouts in the board, just a blank area (perhaps with power traces heading out to the regulators) over which a smaller board would be mounted on short standoffs, 3/16" or 1/4".  Acknowledged that this makes the board "thick" and would not fit in all backplane spaces.

KK headers for the bus signals.  Ribbon to interconnect with protoboard.  Was trying to avoid a specific edge connector for the attached protoboard since often times I need only a handful of signals and don't want to make too many assumptions about the mechanical layout of the protoboard.   A little like the front panel interconnection to the bus.

Ground connection available through metallic standoffs if desired (as well as through the KK headers).  Ground bump has been an issue for me.

Address decoding done by GAL.

In other words, take an H8-7 board and update the bus interface circuitry to use modern buffers, GAL logic, and switching regulators.  Use the read/write latches that have appeared on recent Collado boards (that don't glitch with wait states).  Throw away (please!) the solderless patch panels.  Drill some holes for standoffs to match the protoboard being used.

I didn't do an economic analysis, but if the regular "busy boards" are selling for $25 each, this board shouldn't be any more expensive than that.  Agreed that the blank space in the middle is in some sense wasted space.

tas

Douglas Miller

unread,
Nov 16, 2022, 10:00:40 PM11/16/22
to se...@googlegroups.com

OK, I was thinking that wire-wrap was still desired, for which you'd need a cut-out or something. Or else you have to remove the protoboard from the main board every time you work on it. I guess the same is true for point-to-point soldering as well, although one might be expecting to make changes with WW more than soldered wires.

The only issue I'd see with using a ribbon to connect signals is just further signal degradation. Even the MMS H89 boards that did that made me uncomfortable, and I guess Heath did it, too, with the H37. I guess this ribbon would be ultra-short, but it's also the connectors themselves that attenuate - perhaps more than the cable at that length.

Terry Smedley

unread,
Nov 18, 2022, 1:52:31 PM11/18/22
to SEBHC
Douglas has some great observations about prototyping with wire-wrap, an issue that wasn't on my radar.

My thru-hole solder prototyping needs are already handled pretty well by pirating signals from  the 8255 socket pads on Norberto's PPIO card.  Everything I need except the bus clock is there (and the Z80 IORQ if I choose to use that).  Norberto leaves a spare set of solder pads next to the bus pins where I can pickup anything not on the 8255 pads.  I just use 3M mounting tape to hold whatever commercial board I'm using for prototyping to a PPIO board that is bare except for the bus interface and power regulation.   I've pirated other boards for similar purpose - thanks, Douglas, for this host to the prototype "Blinkenlights" controller.  :-)

It seems like it might be difficult to allocate fixed space for the larger PLCC packages without compromising the space available for more frequently-used components.  The DIP breakouts suggested by Douglas make sense to me.

For SMT components, I've also used off-the-shelf mini-breakout boards to turn them into thru-hole.  That's been fine for prototyping the simpler circuits that I've dealt with.  

tas
PXL_20221118_181150076.jpg
Reply all
Reply to author
Forward
0 new messages