Making square Manhattan pads

[rhb]

I built a sled this morning to cut Manhattan pads quickly and easily. It works a treat!

There are some burrs in the copper, but fine sandpaper will take care of those.

I just spaced these by eye. Later I'll put in a set of holes for a stop so they are perfectly square.

Adjusting the blade height is a little fiddly. I was holding these down with my fingers on the PCB and then with a piece of plywood. If you get the depth of cut just right you can snap of rows of pads of any desired length.

BE VERY MINDFUL OF WHERE YOUR FINGERS ARE IN RELATION TO THE BLADE IF YOU DO THIS!

I'm going to add some spring clips to hold the plywood down on top of the PCB so I can keep my hands at the corners of the sled. However, I now have over 100 pads, so I won't need to make more for quite while.

In hindsight a piece  aluminum plate would make a better sled. I have some variation in the depth of cut from the plywood not being quite the same thickness. I'll need to sand that down.  Though I may simply make a new one using a piece of aluminum.

The idea is to *not* have pieces separate.  The small loose pieces are the result of my depth of cut being too deep, so when I made the cross pass, pads broke off.   I lowered the blade to prevent that.

Edit:

In trying to snap pads apart I found that the grooves were not deep enough.  By using a thicker piece of plywood to press the PCB down I got much more consistent results when I rescored the sheet.


Have Fun!
Reg

[rhb]

Watch this:



It is *really* a much better method.

Reg

[tggzzz]

Alternative: instead of using a scratch tool, use a dremel with an appropriate bit.

Reportedly spherical dental burs are effective

[MadTux]

I use single point carbide cutters for this, like 1/8" or 2-3mm, with lots of RPM, the more the better and smoother it cuts, my Dremel runs at about 30krpm.
And you can grind them from broken endmills and resharpen them about indefinitely yourself on diamond wheel.

You have to hold the dremel at about 45° angle to PCB and cut the slot with Dremel at about 90° angle in respect to cutting direction, then you get very narrow slot. Needs a bit of practice, but works wonderfully.
Big free spaces are also possible, if you hold Dremel at about 80-85° towards PCB, then the cutter cuts free space about the width of cutter diameter.

Fluke 5440B replacement PCB was created that way, because I was too lazy to design PCB on computer:

[langwadt]

thin pcb and a hole punch to make small islands, use super glue to put on pcb where needed, free ground plane too

[m3vuv]

is there no artwork for tonner transfer to make boards like this?,ive looked but had no luck.

[Gyro]

It ought to be easy enough to do. Just use a graphics editor to produce an image with a solid black background and a cross hatch pattern of white lines.

[jpanhalt]

Like the TS, I use a table saw, except it is a miniature saw (4" blade).  No tray is needed -- just a fence.  As for the Dremel, a small round burr (about 1/16" diameter) works for me and doesn't grab or walk on the surface.   That is, it is very easy to control.  I use that mainly on boards when I need to cut a trace.

[rhb]

I'm planning to get a 4" blade table saw and set it up just for PCB cutting with a sliding table.

I want to get to where I can design, fabricate  and package a functional block for a radio  in a couple of hours.  I've got some ~18" x 12" x 3" chassis which will let me mount various modules and connect them with short SMA jumpers so I can easily compare the performance of design changes without a lot of effort and make absolute apples to apples comparisons.  Only one section of the system has changed.

It's easy enough to etch a grid of squares, but that doesn't give you the ground plane that Manhattan construction provides.  Even at HF that is essential to getting good results.  However,if you want a grid of squares and are working at audio frequencies, a laminate scriber will make a grid as fast as you can draw what you want.

At higher frequencies, copper foil tape microstrip works well and is highly recommended by Thomas H. Lee in "Planar Microwave Engineering".  I'm getting ready to try that out next.

Have Fun!
Reg

[RoGeorge]

Nice rig!   

Watch this:



It is *really* a much better method.

Reg

When we were kids we used to use exactly the copper clad scratch method shown at 2:30 in the video, except we didn't have Exacto knives or Dremels back then, so we were just using a hand file and a piece of blade from a former dented metal sawing blade, or a broken knife, or a broken drill.  Almost any metal was good enough, because copper is very soft, and we were not shy to reach the file and reshape again the scratching pick when necessary.

We also used to use concentrated nitric acid HNO₃ as a PCB etcher, for circuits too complicated to build on a scratched PCB.  I still recall the brown fumes and the specific smell of nitric acid.  We were about 4th grade, I think, and we were handling a bottel so big that would probably put any adult in prison nowadays, when the chemicals are so tight controlled.  It was a huge glass bottle with thick walls and a matte glass stopper, the kind used for chemistry, 5 liters I guess, and it was about half with acid.  The other half was always filled with some brown fumes. 

It was hard to pour the acid without spilling, and every drop landed on the pavement was fizzling and bubbling a lot, and if a drop happen to land on the skin, the blister would have took many days to heal.  However, our greatest fear was the smoke.  We were told the brown smoke from the nitric acid is poisonous and to avoid it (which is correct), but it was outside.  With acid in an open tray, the brown smoke was moving with the air currents, many times heading straight to us.  We were many kids around that fuming tray, without any adults around, and we were running away like little squiggling piglets when the smoke was heading toward us  , then we were flocking again around the acid tray, and so on, back and forth for about a minute or so, until the etching was complete.

Being around the tray was a very high stake game, because the acid was etching very fast, with fizzling bubbles like it was sparkling water.  With as little as 10 or 20 seconds too much etching, we would have ended with completely ruined PCBs.  It was not only about loosing all the work and drawing effort (the PCB traces were drawn by hand, with paint or with tar, printers or computers were not yet a thing).  It was mostly about wasting the copper clad that was so scarce.

However, the biggest fear of all, for us, was to avoid any drops of acid falling on our cloths.  Or else, when we got back home our moms would have slapped us and punished us badly for ruining our cloth.   

It's funny how back then no mom was thinking to sue the electronic club or the teacher "for gross negligence with the kids", and instead just slap and guilt trip us for not being careful enough with our cloth. 

Today I would probably be arrested if I'll be caught outside, puoring 2 kilos of fuming HNO₃ in a tray!     I bet HNO₃ is illegal now, as a bomb precursor or a drug precursor, or something.

Yet, with all that acid and all that behavior back then (outrageously irresponsible for today norms), nobody "dieded" or got permanently "damaged".   

Sorry for the short offtopic novel.

w2aew has a few videos about circuit prototyping methods:

[rhb]

I bought nitric and sulphuric acid at the local pharmacy and used it to make flash paper by nitrating 100% cotton bond paper when I was about 12.

I'm sure Dad knew I was doing it, but I doubt there was any discussion other than his mentioning that when he was in college they all wore corduroy pants in chem lab which they did not wash all semester lest they fall apart.

It is rather sad that all the things I did as  kid are no longer permitted.  And the local library has no technical books at all.  I had a great raft of books on electronics, fireworks and explosives on almost perpetual loan.  I was told by the librarian currently in  charge of non-fiction, "You can't learn electronics from a book.  You have to go to college for that."

Reg

[cdev]

Arrggh.. 

  I was told by the librarian currently in  charge of non-fiction, "You can't learn electronics from a book.  You have to go to college for that."

Reg

[cdev]

What do you mean by "chassis" here?

I want to get to where I can design, fabricate  and package a functional block for a radio  in a couple of hours.  I've got some ~18" x 12" x 3" chassis which will let me mount various modules and connect them with short SMA jumpers so I can easily compare the performance of design changes without a lot of effort and make absolute apples to apples comparisons.  Only one section of the system has changed.
.....
At higher frequencies, copper foil tape microstrip works well and is highly recommended by Thomas H. Lee in "Planar Microwave Engineering".  I'm getting ready to try that out next.

I've used copper tape in antennas (typically with cardboard as the substrate) quite a bit. It works quite well.

[rhb]

Standard dictionary definition in an electronics context, an enclosure.  These were some thick  wire ethernet repeaters which I bought for this purpose which were killed by a lightening strike.

What appealed to me about them was there was lots of front panel space horizontally so all the modules could be on the same plane and sufficient depth to make it easy to place a dozen or more small  modules without struggling.  With a few controls permanently mounted to the front panel and a few connectors on the back panel I can swap out the internals very easily.  It's intended to be a universal experimental radio framework which looks like a finished unit with the cover on.

This is to let me  evaluate different mixers, filters etc and not have the variation due to part tolerances in the rest of the circuit inherent in building a new board with one section changed.  Or have to build an entire radio to change one section.  With all the modules in metal enclosures it also provides very good shielding.


Have Fun!
Reg

[cdev]

Magnets could be used underneath PCBs perhaps with some conductive material to provide an RF ground while letting them still function as magnets, that would make them easy to move.

[RoGeorge]

Ferrite properties go south near permanent magnets (and have a tendency to remain there , until demagnetized).

[janoc]

It is *really* a much better method.

Reg

It is not, especially if you are prototyping RF stuff. The Manhattan/"ugly" style construction is a staple there because the continuous copper to which you stick those islands acts as a ground plane, helping with stability and EMI problems.

[cdev]

Somewhere I have a coil with a magnet attached, dont remember where I got it.

Ferrite properties go south near permanent magnets (and have a tendency to remain there , until demagnetized).

What's it for?

[rhb]

Two points:

If you use Leo's technique to build a subsection and then glue that on a ground plane you can use SMD parts for Manhattan construction.

Remnant magnetization causes the core to saturate sooner at one polarity than the other.  That is *seriously* not good.

Reg

[JohnG]

Watch this:



It is *really* a much better method.

Reg

If you need a ground plane, just carve up one side of a double-side blank PCB. The other side is your ground plane.

I will be making one of those tools. I couldn't make a straight line freehand with a Dremel to save my life (or with a pen or pencil, for that matter), and my curved ones don't go where I want.

The wire wrap wire and the teflon tubing I have used almost my whole career. However, I will not ever do any actual wire wrap. I will learn to make straight lines with my Dremel longer before that.

John

[janoc]

If you need a ground plane, just carve up one side of a double-side blank PCB. The other side is your ground plane.

Double sided copperclad is quite a bit more expensive than single sided. Not to mention that if you use the islands, you can easily remove them and reuse/modify the board. No chance if you carve the surface up. Sticking the islands on (whether using superglue or doublesided tape) is also much faster than carving the board using a knife, file or some rotary tool.

Keep in mind that the Manhattan style of building is used mostly for RF prototyping, replacing breadboards which don't work at those frequencies anymore.  Frequent (and possibly extensive) modifications of the circuit are completely normal and expected during the development. Carved/engraved boards are not very amenable to that, those are more suited for finished circuits.

[Howardlong]

I've become a massive fan of Busboards, in particular SP3T-50x50-G and SP3-100x100-G, with 0.05" and 0.1" square pads respectively.

These have a solid ground plane on the back with non PTH holes: you choose which ones to through wire.

You can solder SOICs and SOTs directly onto the SP3T-50x50-G, and it's more useful for RF stuff up to low UHF, with through ground connections available every 0.2", and a 0.8mm thick.

On the SP3-100x100-G, you're never more than 0.4" away from a possible ground connection, but typically when building you manipulate the layout around the ground through connections. I then use breakouts as necessary, clipping the edges to make castellations to solder directly on to the Busboard.

The key here is that because they have a ground plane, the number of flying hook up wires is quite limited.

I recommend keeping a good stock of breakouts. These, as well as Busboards, good way to make up the minimum order for free shipping.

I realise Busboards aren't exactly cheap. However they are the best solution I've found for RF and mixed signal. One of the boards below is a precision current meter, it measures down to 100pA, although it has to sit inside a foil tray to get those kinds of measurements!

The other board is a work in progress, a portable dual voltage source and current measure unit, 0 to 30V +/-50mA source/sink on each channel with 500uV resolution, and current measurement down to 100nA. Not quite an SMU, but does most of what I use an SMU for, i.e. sitting in source V, measure I mode.

[rhb]

That looks to me like an expensive way to do what Leo showed in his video.  Obviously nice, but as you note not cheap.  Though a thin slitting saw and sliding table would make similar boards to order as needed at fairly low cost.

In the past I've used dead bug and just glued the parts to the copper upside down and put the connections in the air.  That works for leaded components up through HF, but becomes increasingly problematic above that and completely impossible with SMD which you *really must have* at higher frequencies or where you need well controlled impedances because of transition times.  Multiregion Manhattan pads as Leo showed makes that quick and easy.

I'll be going to Harbor Freight next week so I'm going to buy one of their 4" table saws and tool it up for cutting PCB material with a nice sliding table running on ground bar stock rails, carbide blade, etc.

I had a 15 year hiatus from electronics which was mostly driven by the mechanical fabrication issues imposed by my electronic projects.  Having resolved those it's fun to get back to electronics.

Have Fun!
Reg

[tggzzz]

I've become a massive fan of Busboards, in particular SP3T-50x50-G and SP3-100x100-G, with 0.05" and 0.1" square pads respectively.

Seconded, for the reasons you give. The only negatives are the cost and having to have them in stock.

[langwadt]

I've become a massive fan of Busboards, in particular SP3T-50x50-G and SP3-100x100-G, with 0.05" and 0.1" square pads respectively.

Seconded, for the reasons you give. The only negatives are the cost and having to have them in stock.

wouldn't take many minutes to draw up in kicad and order from jlcpcb or what ever your favorite is