Another Home-Brew - Z180

[Bill B]

20191227_140543.jpg (908.75 KiB) Viewed 18758 times

Well this is an improvement on my previous attempt

I purchased a 230 x 160 mm board from RS. That gives me two 160 x 100 mm boards, plus a 30mm spare strip to provide exposure test pieces.

On close examination it is not perfect. There are places where pads are connected to the adjacent tracks. But hopefully with some careful scalpel work it is usable.

I need to tweak my design rules, and perhaps use a slightly longer exposure time, for making the second board.

[Bill B]

Including a Propeller Chip and level shifters on the Peripherals board is defenitly tight. I did not have any spare space for mounting holes.

By doing some address decoding for the Propeller in the CPLD, there are some Propeller I/O pins spare, so as well as VGA out it is (from a hardware point of view) possible connect a keyboard input and an interrupt back to the Z180. Yet to work out whether there are enough cogs to support this. Possibly PS/2 in, USB in would require at least three cogs which may be a stretch.

Learning the lessons from etching the CPU board, I have reduced the width of the IC Pads and the Via diameters from 0.06 inch to 0.05 inch. Also the width of the non-power tracks from 0.02 inch to 0.018 inch. This gets closer to when I used to lay out boards using my MTX, for which I wrote a PCB paint program, with pixels at 1/60 inch. These were then smoothed to 1/300 inch when printing.

[Bill B]

PCB_Detail.png (75.57 KiB) Viewed 18727 times

I am going to have to re-think my PCB designs again. With a home made PCB it is necessary to solder small pieces of wire between the upper and lower pads of a via. My soldering is not good enough to do this between pads at 0.05 inch spacing without creating shorts between pads or tracks. It does not help that my UV box is single sided. It is very difficult to keep alignment between the upper and lower layers as I turn the board over. I did quite well with the CPU board, not so well (~1mm out) with the Peripherals board.

Dave is undoubtedly right, it would be more reliable and higher quality to have the boards made professionally. By the time you count wasted materials, broken drill bits and effort taken, probably cheaper and quicker as well. But it seems a shame to have the tools and not use them. Also the "I made that" satisfaction.

One option is to do single sided boards and use wire links instead of tracks on the upper layer. It will still be necessary to move the via pads further apart (0.1 inch spacing) in order to be able to solder the links to the tracks without shorts. Ideally I would like to be able to make the via pads rectangular, but would then need to be able to orient them horizontal or vertical. And with the extra spacing, possibly a double layer board will work.

Back to the drawing board

[Martin A]

If you want the "I made that" satisfaction, there's always the matrix board route. With the components on one side and wires on the other, it looks quite neat too.

I use Kynar wire for hook-up, since it's simple to just melt the insulation when doing a multi point path like the data bus. Because it's wire hook-up and not tracks, the 3rd dimension is available when working with awkward PLCC sockets etc. The 0.1" spacing is pretty easy to solder too.

I use 2 different "eurocard" size matrix boards from RS. RE200HP which is a basic rectangular board, and RE060-HP that has a few less holes, but takes a DIN41612 connector for backplane based designs. Neither board costs more than a few pence more than the equivalent size copper clad board. And there's no messy developing etc. required.

[PMG]

Hi Bill,

2 tricks I have used in the past for double sided etches:

1) Put fiducials on your design (These are markers that are in defined positions on the board - usually 2 or 3 of the corners). Then when you stick down the mask it helps you to apply it to the exact spots.

2) Drill the mounting holes before masking and use them as mechanical fiducials.

It's many years since I did any of this, so my memory may be playing tricks but the boards were a bit simpler than yours too.

Good luck.

Cheers,
Paul

[Bill B]

Did both of those.

The board cut lines are extended well beyond the edges of the board, and can be used to align the top and bottom masks.

The boards do not have any mounting holes, too packed. However, prior to etching I drilled two of the component mounting holes (in opposite corners) through both the masks and the PCB. However it is still difficult to align the mask under the board with the drilled holes.

Next time I will try something different, put a roll of blue-tack (or similar) along each edge of the board. Still have to be careful, it needs to be tight against the board to stop it slipping, but must not lift the mask away from the board at all, otherwise it will blur the tracks.

[1024MAK]

You need some thin wire or thread to go via holes in the board and the mask to help hold everything in the correct place. Best if you can manage this at four corners or equivalent. The hole size for the thin wire or thread should be only big enough for the wire/thread to pass through, otherwise there will be too much slack.

The other trick, but only practical on small boards, is to tape the two masks together as a sandwich, along three edges, leaving the last edge open. Then slip the board into the ‘pocket’ formed by the taped together masks. As I say, only practical if the mask sheet is significantly larger than the size of the board.

Even using these tips, it’s still difficult to get exact alignment. So you design needs to allow for a certain amount of misalignment. This does limit the complexity of the design.

I only ever tried relatively simple double sided designs because of this.

Mark

[Bill B]

The problem with that is if the wires go right through, and are then bent over the bottom mask is that they then lift the bottom mask off of the the glass plate, so that the mask is no longer pressed firmly against the copper.

[Bill B]

Progress at last

20200126_103449.jpg (1.03 MiB) Viewed 18364 times

Lots of steps involved:

• The layout was revised to ensure that all vias were at least 0.1 inch apart.
• I created a number of footprints consisting of ovals of different orientations on the upper and lower layer, such as

  

  Via_Image.png (2.42 KiB) Viewed 18364 times

  and placed them over the vias, to give larger pads.
• My new technique of using Blue-Tack to hold the top and bottom masks in place worked reasonably well. I still didn't get the top and bottom copper perfectly aligned but it is good enough. Next time I will combine techniques, and use a couple of holes through the board to align the templates, and then use Blue-Tack to hold them in place.
• Once the board was etched, the next task was to join the top and bottom layer of each via together. I previously used via pins that I got from Maplin. They came in a long strip,the pins joined end to end. The end pin would be pushed through a hole in the PCB until it jammed. It would then be broken off and soldered in place on both side. The pins were a bit big (they required a 1.0mm hole) but the long strip made them easy to handle and quick to install. I have not found anywhere to buy the equivalent now.
  
  The best alternative I have found are these, from Amazon. Despite their stated size, they will not go through a 0.8mm hole, they require 0.9mm. Each pin is loose and they are tiny and incredibly fiddly to install. Minutes per pin, instead of seconds for the old pins from Maplin. Also, drop one and it will vanish.
• Once all the via pins were in place, fitting most of the remaining components was quick and easy, which is why I prefer making a PCB to point-to-point wiring. The one exception was the PLCC socket for the Z180. I had got it 80% soldered in when I discovered that one pin was not through its hole, so it had to come out again I never have much luck with a solder sucker. I do slightly better with some solder wick. Even then it required extra flux paste to mop up enough solder to get the socket out again. When I finally got the socket out I found that the pin in question was broken off and not there to go through the hole. It is possible that it broke when I was installing it, but I don't think so. It came from China (via Amazon). So put another socket in after inspecting in carefully.

Next step is careful testing for continuity and particularly shorts between adjacent tracks.

[Martin A]

Looking good!

Have you though about just using a piece of stripped back single core wire instead of pins for the through board connections ?