PCB Layout 2

[treez]

Hello,

I am trying to do a PCB layout for a mount for a current transducer……

The following is the transducer…

http://web4.lem.com/docs/products/htfs_e.pdf

I am trying to do it with PCB Wizard 3 (Standard edition).

I wish to get a dimensionally correct print out from PCB Wizard and then use that as a mask ready for etching away the copper on the copper clad board.

Here is my effort so far…….(it includes holes for caps and for connecting wires)

However, as you can see from the dimensions at the end of the datasheet, I will need very tight resolution since the 4 signal/power connections are not “true” with the mounting holes…..in fact the 4 mounting holes are at the corners of a square of side 33mm…..the four signal/power holes are in a straight line and are 2.54mm apart. –Also, these four signal/power connections are in a line which dissects the 33mm-side-square….however, the first hole starts at 32.5 mm above the “bottom” hole.

In short, I need the PCB software to have a grid with a resolution of 0.01mm to do this.

Does anybody know if PCB Wizard 3 (Standard Ed) can do this?
Also, if any reader knows of a web help group for PCB Wizard then I would be most grateful for it. I've been to the new wave website but sometimes the best help group operates independently to the main manufacturers.

 

[benta]

Given that the dimensions are given with a tolerance of +/- 0.2 mm, I think you are overdoing this a bit.

Benta.

 

[ScottyUK]

With the etching processes available to the amateur user and the drilling accuracy available on a manual drilling machine you will be doing well to hold 0.1mm tolerances, never mind 10[μ]m tolerances. Oversize your mechanical mounting holes slightly.



----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[MacGyverS2000]

That's a res of 0.4 mils!!! Almost all of my work is with SMD components, with pitches often at 20 mils... 1 mil pitch in the design is more than adequate, and you certainly don't need that kind of accuracy for a huge, 4-post component with such slop in its own tolerance.

The pins are on a standard 100 mil pitch. Give them a 10mil annular ring and oversize the hole by 10-15 mils. Use the nominal dimensions for the pin-to-mounting hole distance, and oversize the mounting holes by 30-40 mils

Dan - Owner

http://www.Hi-TecDesigns.com

 

[ScottyUK]

treez,

I sense you are in the UK. When the US members refer to 'mils' they really mean 'thous'. How a thousandth of an inch ever became 'mils' is a long term mystery to me; a prize to any US reader who can explain how us Brits get from 'thousandth' to 'thou'!


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[melone]

I would have to presume that it is a carryover from the US military:

mil [1]
a unit of distance equal to 0.001 inch: a "milli-inch," in other words. Mils are used, primarily in the U.S., to express small distances and tolerances in engineering work. One mil is exactly 25.4 microns, just as one inch is exactly 25.4 millimeters. This unit is also called the thou. With the increasing use of metric units in the U.S., many machinists now avoid the use of "mil" because that term is also a handy slang for the millimeter.
mil [2]
a unit of angle measure, used in the military for artillery settings. At one time the U. S. Army used a mil equal to 1/1000 of a right angle, 0.1 grad, 0.09°, or 5.4 arcminutes (often written 5.4 moa; see "moa" below). Later this was changed to 1/1600 right angle, or 0.05625° (3.375 moa). In target shooting, the mil is often understood to mean 0.001 radian or 1 milliradian, which is about 0.0573° or 3.43775 moa. In Britain, the term angular mil generally refers to the milliradian. 1 milliradian corresponds to a target size of 10 millimeters at a range of 10 meters, or 3.6 inches at 100 yards.

http://www.unc.edu/~rowlett/units/dictM.html

 

[MacGyverS2000]

melone has it on the nose with definition 1. I was actually caught out with the problem he mentions while working with a CAD guy... the PCB package expressed everything in mils, and he designed the package in millimeters (even when it made no sense... must have been a major brain fart day for him, he was excellent otherwise).

That said, most board houses express their specs in mils, followed by millimeters in parentheses.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[ScottyUK]

Thanks melone, I hadn't thought of milli-inches. My little mystery solved.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[treez]

Thanks for these replies.

I have oversized my holes accordingly. I printed the artwork onto a sheet of paper and found that the mounting holes have come out measuring up as 32mm apart instead of 33.02mm, which is what they should have been from the PCB design development environment.

I hope i can get a dimensionally accurate printout somehow, as this will be the basis of forming the "mask" ("artwork track image") on the (resist-coated) copper clad board.

Could i ask for conformation that getting a dimensionally accurate print-out of the copper tracks and holes is actually feasible? I suspect that i need to play with printer settings or such like?

 

[IRstuff]

You need to use a better printer, or you need to calibrate your printer.

Frankly, trying to do submillimeter accuracy on a home printer is stretching it.

TTFN

FAQ731-376

 

[ScottyUK]

A laser printer should manage it, but use decent transparency film. My old HP4000 does very well at 1:1 printing onto paper from AutoCAD, as do most HPs. I haven't tested many other brands but I doubt they would be significantly less accurate.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[itsmoked]

You should print out a 10 inch and a perpendicular 8 inch line on your page with the same system and then measure them.

Most CAD programs have the ability to adjust this.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[IRstuff]

Your other choice, old school, is to make a 5x or 10x image and photographically reduce it to the correct magnification. This would reduce the scaling errors in the printer.

TTFN

FAQ731-376

 

[MacGyverS2000]

Somebody help me understand this... why is a professional design house using paper masks and photosensitive PCBs for design work? Proto houses would charge you $150 and send you a handful of boards within a week, so why bother doing it any other way? I simply cannot imagine any professional outfit spending so much time on such an error-prone process when a minimal cost and speedy solution is available...

Dan - Owner

http://www.Hi-TecDesigns.com

 

[itsmoked]

and you forgot poisonous chemicals..

Keith Cress
kcress -

http://www.flaminsystems.com

 

[treez]

Hello,

It is intended to make up these test boards (they are not for sending to customers) in-house using a UV light box and developing and etching the resist coated PCB.

I agree that this is slightly messy, -but takes only some 30 minutes, and costs no more than than a few pounds for the chemicals used and the PCB.

Comparing this cost with that of a well-known prototype PCB house (which claims to be able to give discounts because of their 'procedures')....

For 10 single-sided (no frills what-so-ever) PCB's of the type in the first post, it would cost £215 (British pounds, -and this doesnt even include VAT)....for 8 working-day delivery (prices increase sharply for less days delivery time)

The entire cost of the home-etching technique would be less than this...the most expensive item would be the UV light box at £89.00.(and once this is purchased , it can obviously be re-used at no further cost)

Considering that further test-board PCB's might be needed in future, the home-brew PCB comes in a clear winner in terms of cost and delivery time.

 

[GonzaloEE]

From my own experience with home-brew PCBs from college days, I never got any track less than 20mil wide. Not to mention whole days I spent printing, etching, drilling, dropping several failed attempts, etc.
I'm afraid I prefer paying for a photoplotter reading vectorial gerbers, than correcting transparencies with permanent pen myself ;-)

Good luck, anyway!

 

[treez]

Thanks, I'm certainly going to look into the photoplotter to save using a printer.

I wouldnt like to try less than 20 mil (0.5mm) but dont know if its possible.

I used to etch PCB's with very fine spiral antennas on them for 30-40 GHz (prototype) use for a Microwave company. It used to work quite amazingly well.....i was amazed at the beautiful clarity of the thin, sinuous spirals we used to get. Using nothing more than plastic bowl of developer, wash, and one of etchant.....you just had to look at it in the developer to check you didnt over/under develop.

.....we didnt even bother with transparencies...just used the white printer paper with the spiral on it and put that on the coated cu clad board in the UV light box.

 

[MacGyverS2000]

You're telling us you were able to etch photo-sensitive boards using normal white paper instead of transparencies?

Dan - Owner

http://www.Hi-TecDesigns.com

 

[richs]

http://www.5bears.com/pcb.htm

Has a nice little tutorial on homemade "do it yourself" PCB fab using toner transfer from a laser printer.

HOWEVER, I find the transfer paper mentioned on this site:

http://www.pulsarprofx.com/

and sold as digikey p/n

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail?name=182-1003-ND

much better.

Also getting the "green film" to "protect" the toner image is also worthwhile.

http://search.digikey.com/scripts/DkSearch/dksus.dll?pname?site=us;lang=en;name=182-1021-ND

I did a little note on it some time ago. I find that I can, if pushed, get a printed circuit board to a customer overnight and really look like a hero! Anyway, along with the toner transfer and green film products mentioned above, and the little notes below, this is my two cents worth to the thread:

A laminator seems to be one of the best tools that you can use for the dry process. The iron on process doesn't seem to work very well and my reliablity factor increased dramatically when I went to the laminator. Found one at ebay for $50.00 although it came broken..... Simple to fix. The web site at the dry transfer paper site has the modifications they do for the laminator, however, I just use thinner boards and have no trouble with the unmodified one.

As mentioned earlier, clean the board carefully! One of the most important steps for a clean etch. DON'T USE STEEL WOOL IF YOU WANT TO TIN PLATE IT! I screwed up on that one and I contaminated the tin plate. Arrrrgh!

Use the heaviest toner setting on your printer/photocopier. The more toner you can lay down, the better. You still might have some pinhole ommissions on the layed down toner. You can touch those up with a Sharpie.

CUT the paper to the size of the PC board plus a little bit.
The paper is expensive! Mount the paper on a backer piece of normal paper. USE POST A NOTEs for the next couple of steps.
The post a note adheisive (cut off the part of the note without the adhesive on the back and use the sticky part) can withstand the heat of the laminating/printing process. This was a neat trick passed from another user.

Use the post a note adhesive to also mount the transfer paper to the blank copper clad. I usually wrap it around the back side of the board unless doing double sided.

I make 3 or 4 passes through the laminator to fully reheat the toner and transfer it to the copper clad.

Soak the board/paper in water to release the paper backing.

I also make 3 or 4 passes through the laminator after applying the green film to "harden" (cover) the toner.

Touch up if required with sharpie any broken resist traces.

I HEAT my ferric choloride up to about 140 degrees or so. I use a Salton Hot Tray, with my horizontal ferric choride tank (Hefty disposable cassarole pan). On top. Usually about 8-16 oz. FeCu (fresh if I'm in a hurry. New stuff etches much faster).

ROCK the tray back and forth. It seems that even the vertical tanks do not pass enough FeCu over the board. ROCK the board both side to side and back and forth. Otherwise, you will find that the etch times vary, both the location on the board and the density of the traces on the board. Use the same rocking motion you do if you have developed photographic paper in the paper trays. About one cycle a second and slosh the liquid to just barely keep the fluid in the tray.

My friend who passed the post a note tip to me uses a "Sponge Method" of wiping on the FeCu (mentioned on the release paper web site). I haven't done it myself.

I keep a couple of bamboo skewers around for big toothpicks to lift the board from the FeCu and see the state of etching.

You can easily see when the etching process is complete and also get a good idea of when you might want to replace your FeCu.

Remove and rinse the board. AGAIN don't use steel wool to remove the green backing if you want to tin plate. If you don't want to tin plate, scrub away with the steel wool.

You can remove the green backing resist from the board (and the toner underneath) with acetone. I tried finger nail polish remover from my wife's stock. It's not pure acetone so it didn't work as well as the stuff you find in the paint store.

If you don't tin plate and you wait a while before soldering, you might want to run the pads over with a bit of steel wool to remove any oxide before soldering.

With this process I've been able to do 12mill trace widths. I might be able to go lower, and I think that I also do 8mil widths.