Smallest pitch for home made PCB that is doable for experienced guys.

[Rick Law]

As someone with limited experience, I have been kicking around the idea of trying out etching my own PCB just for the fun of it.

I am wondering for home-made PCBs, with laser printed (HP5P) thermo-transfer, what would be the precision limit one can expect from a experienced person.  I want to set a goal that while I may not be able to do immediately, but is doable as I improve.  Printer will be a laserjet 5P (not buying a new printer just for doing PCB's).

I want to (eventually) do a PCB for which amount other things will have an ADS1115 (instead of using the Adafruit breakout).  It is a 0.5mm pitch VSSOP chip.  I wonder if that is even achievable (given home-made limitations) even for experienced guys - each solder pad is merely 0.3mm wide.

If that is too ambitious, what pitch do you thing an experience guy should be able to do with home made PCB?

Thanks for your input,
Rick

[lapm]

With good equipment its doable.. And with good equipment i mean stereo microscope/mantis, proper tip/hotair soldering, patience...

[redshift]

The smallest pin-pitch I've bothered to etch is 0.025" (~0.65mm).

My method: I transfer toner using a laser printer, glossy paper, and a hot clothing iron. Then use ferric chloride to etch away the copper. Lastly, acetone to clean off the toner.

I don't have a hot-air station or a microscope. Just my hakko fx-888 and a magnifying glass to look at things afterward. So I'm willing to bet 0.5mm is possible(especially if you have these things).

Another suggestion: If you don't think you can reliably make a footprint with small trace widths, you can always solder your VSSOP to a breakout board and then solder that into your pcb.

[Cloud]

For very thin lines and pads its better if you use PCB with photo sensitive coating and foil.

[Rick Law]

With good equipment its doable.. And with good equipment i mean stereo microscope/mantis, proper tip/hotair soldering, patience...

Since you talked about proper tip/hotair, I take it you mean the soldering phase of the project.

I am confident with the soldering part.  I blew out the ADS1115 by over-volt and I had to replace it on the Adafruit breakout.  I have merely a 4x loupe and no hotair station.  That ADS1115 replacement job was successful the first try but it did required a great deal of patience.  Difficult for sure, but I know I can do it.

It was the difficult with soldering that ADS1115 replacement that leads me to think: I can't "magic marker" fix the thermo transfer, there is no chance I can draw anything 0.3mm width.  So fix up any imperfection in the thermo transer on the copper plate will be darn near impossible and I would be relying on pretty much a perfect transfer...  Which is why I wonder if it is a realistic "home project" to make PCB for something like that.

Thanks for the input!

That is certainly possible, but it isn't trivial.
One thing that is helpful in the case of a small breakout board is just the small size of the PCB itself.
...
...

Thanks, evb149.  Lots of info for me to digest.  I know 100% success rate is unreasonable.  My bar is: more good ones than bad ones.  I have to digest the details of your post.  It is too interesting to merely gloss over reading it just once.

Thanks.

The smallest pin-pitch I've bothered to etch is 0.025" (~0.65mm).

My method: I transfer toner using a laser printer, glossy paper, and a hot clothing iron. Then use ferric chloride to etch away the copper. Lastly, acetone to clean off the toner.

I don't have a hot-air station or a microscope. Just my hakko fx-888 and a magnifying glass to look at things afterward. So I'm willing to bet 0.5mm is possible(especially if you have these things).

Another suggestion: If you don't think you can reliably make a footprint with small trace widths, you can always solder your VSSOP to a breakout board and then solder that into your pcb.

That is encouraging that .65 pitch is possible.  How long did it take you to get that far?  Or may be a better question is, how many PCB have you made before you get to the point that you try and then succeed in making .65 pitch?  By success, I suppose I mean better than 50/50 chance of having a useful board coming out at the end.

I will forgo breakout board.  I want to get the sense of accomplishment of "doing it myself".  Using the breakout board defeats that.

I only have a 4x loupe for magnification (and my reading glasses) and no hot air station either.

I think if I can do .65mm, I can be happy.  I don't have a .65mm pitch project in mind yet however.

Thanks

For very thin lines and pads its better if you use PCB with photo sensitive coating and foil.

Interesting.  I had thought toner transfer is the better method.  I think I used photo method before.

Over 20 years ago when I last was interested in doing EE, I etched my first PCB taking the layout from a magazine.  I only recalled it worked but I have no recollection how I etched that.  (Funny how the brain works.  I recall the drive to Radio Shack to get the stuff, and the tray I used to etch, but I actually do not recall how I did the etching and not even remember what circuit it was!  But I can picture that white tray in my mind just now, and picture my squarting down at the store inspecting the bare copper side of the PCB to select the one to purchase at the store...  Strange)

I have to keep that in mind.  May be if I do decide to give etching a go, I have to research present day photo sensitive PCB etching.

Thanks

[Dago]

Why do you want to make your own PCBs? Practically nobody makes their own PCBs anymore because you can get professionally made 2-layer PCBs (10 pieces, 50x50mm) for 9,99$ from China. 4-layer boards are like 25$.

Photosensitive board is definitely the better method. You can easily make boards for 0.5mm pitch. It is pretty useless though because there is no good method for making vias, other than making "huge" holes and soldering pieces of wires there. I never got the toner transfer thing to work well or reliably.

[lapm]

Becouse there are times you dont want to wait couple weeks to get prototype going. Its art form thats is fading how ever...

[tautech]

15 mil tracks and 10 mil clearances is as fine as I'm happy to go for reliable results.

Do give it go, for us that mucked with Dalo pens years ago, CAD and TT is very rewarding.
Plenty of threads if you hunt for them, a recent one:
https://www.eevblog.com/forum/projects/pcb-at-home-guide/

[tggzzz]

You should carefully distinguish between designing and manufacturing the PCB and assembling the components on the PCB.

Manufacturing the PCB at home is, of course, possible. The major limitation in my limited experience is the quality of the etch resist - imperfect resists means the tracks will be pitted. Using the laser toner on grotty paper method, it is just about possible to get a single track between the pins of a 0.1" pitch through-hole IC. I will continue to use toner transfer if I need a medium complexity board fast; if I need a tiny board I will use a dremel/cnc, and for most boards I will wait 10 days for a 10 cheap boards from China (and visually inspect them carefully!)

Assembling the components onto a PCB is much easier; SMD 0.5mm/0603 is easy with cheap equipment, as I have described here: https://entertaininghacks.wordpress.com/tag/smd/
(note the split into designing, manufacturing, assembling and reference material)

[tautech]

Using the laser toner on grotty paper method, it is just about possible to get a single track between the pins of a 0.1" pitch through-hole IC.

What, haven't you heard of modding pads for better clearances or to open other routing posibilities?

[tggzzz]

Using the laser toner on grotty paper method, it is just about possible to get a single track between the pins of a 0.1" pitch through-hole IC.

What, haven't you heard of modding pads for better clearances or to open other routing posibilities?

Of course! It is more a comment on my ok-but-not-wonderful etch resist.

[mikeselectricstuff]

The limiting factor is often via size and lack of plating rather than line width - you can't go too small with vias as it's hard to get good alignment, and laser printers can have small dimensional errors.

[ion]

I've not had much luck with thin traces using toner transfer so I've been meaning to move to photo sensitive boards.

So far I have only done a couple of exposure tests, but looking at the board right now the lines only start breaking up below 0.4mm.

There's also a QFP-32 footprint on the test board and >75% of the 0.2mm wide pads came out fine.  I suspect tweaking the exposure time would fix that.

[Psi]

I've done 0.5mm TQFP with press-n-peel blue toner transfer.  It was pushing things about as far as you'd want to go.
I had to tin the small pcb tracks with solder to bridge some tiny scratches.

If you attempt it i'd advise you to make the tracks wider than the gaps because they will side etch. The smaller the track is the worse this gets.   For example, a 0.2mm track will end up more like 0.1mm after etch.
So, if you want 0.2mm track and 0.2mm gap, make it 0.3mm track and 0.1mm gap

[redshift]

That is encouraging that .65 pitch is possible.  How long did it take you to get that far?  Or may be a better question is, how many PCB have you made before you get to the point that you try and then succeed in making .65 pitch?  By success, I suppose I mean better than 50/50 chance of having a useful board coming out at the end.

I probably had less than 50% success when I tried this. I remember wasting at least a few pcb's but I don't etch boards very often...

[PeterFW]

If that is too ambitious, what pitch do you thing an experience guy should be able to do with home made PCB?

Even with my horrible setup and the worst process control imaginable i etched a 0.3mm pitch board just recently on a 4 year old photo PCB that was never properly stored.

0.3mm should be no problem at all if you have a decent printer.
Your printer and paper is the most important piece.

I use normal copy paper, no fancy stuff, brush the oposite side of the paper with acetone to set the toner a bit after printing.

With decent paper and a good printer 0.3mm is absolutely no problem at all.

Edit:
Do not bother with toner transfer, that technique is inferior and only for people who are afraid of harmless chemicals
(Harmless as long as you wear eye protection)

[tggzzz]

I use normal copy paper, no fancy stuff, brush the oposite side of the paper with acetone to set the toner a bit after printing.

Interesting.
How does acetone improve the end result? Are there any other tips/techniques associated with that?

By "copy paper", do you mean filled, unfilled, inkjet, laser printer paper, etc?

Do not bother with toner transfer, that technique is inferior and only for people who are afraid of harmless chemicals
(Harmless as long as you wear eye protection)

What chemicals do you use?

[fubar.gr]

I've done 0.65" pitch using the toner transfer method with amazing results.

The problem with this method is that you need to find the "magic" combination of printer model, printer settings, toner type and paper.

If something of the above is changed, the final result might suffer significantly.

[PeterFW]

How does acetone improve the end result? Are there any other tips/techniques associated with that?

It does not have to be acetone, you can use a number of solvents.
The best i came across is d-limonene (orange terpene, if that is the correct english translation), often sold as label remover.

The acetone will increase the opacity of the toner so the layout will have a higher contrast. In turn this gives you a wider exposure time sweet spot.
The fuser in the printer will not bake the toner properly, the acetone dissolves the toner slightly and when it evaporates toner settles more dense.

I use a brush from the back side because otherwise i will smear the toner because the acetone will dissolve every aerosol pump i have
If you get the label remover you can just lightly spray it from the top side.

Edit:
Here is a comparison picture on transparent film, but it has the same effect on laser printer paper.
http://www.pixelklecks.de/pictures/solvent50.jpg

Transparency film may be allergic to acetone, in that case you have to use d-limonene.

By "copy paper", do you mean filled, unfilled, inkjet, laser printer paper, etc?

Normal laser printer paper, the cheapest you can get, because the high end may have UV blocking stuff in it and is thicker.
But it helps if it is smooth paper but it does not have to be.

The draw back is that the exposure time is very, very long.
With proper transparent film the exposure time can be 30 senconds, with laser printer paper its about 10 to 20 minutes.

If you can, you should use the transparent film but im to cheap for that... 

What chemicals do you use?

Just NaOH (sodium hydroxide) to develop and Na2S2O8 (sodium persulfate) to etch, both are rather harmless unless you drink it but the same can be said about every cleaning agent in your house.
The NaOH is caustic so eye protection is a must, if you get that in your eye, thats bad news.

The sodium persulfate needs to be warmed to 40° to etch properly but in turn it is translucent and you can observe the etching process.
You can etch with ferric oxide but that is a brown goop, i do not like brown goop.

I would prefer another method of etching but that one can only be done outdoors or in a fume hood. Too much trouble...

Greetings,
Peter

[Rick Law]

I've done 0.5mm TQFP with press-n-peel blue toner transfer.  It was pushing things about as far as you'd want to go.
...
...

Psi, that is impressive!  Beautiful PCB you've made.  Fair to say, that is the kind of DYI board that I am hoping I can do some day.

I have to rethink what I should do for "before some day".  This discussion alerted me I am overlooking something.

The boards I have in mind will be mix of thought hole and SMD.  The idea of drilling dozens of precision holes could be very difficult with just a drimmel.

I had in my mind imagine my DYI board like the production boards I've worked on - in fact, the finished DYI board would not be like that!
- I come to rely on the solder mask when soldering the SMDs; it gives me some confidence on my ability easy to reduce or to remove solder bridges.  Not having soldered SMD on home made boards, I am not sure exactly what the bare-copper impact would be on my ability to solder them nicely.
- Often, I have tiny flicks of solder on the board that I later remove with my sharp tip needle.  Without the solder mask, these tiny flicks would be sticking where they fall.
- I have in mind I can do two sided, worst cast is doing to boards and glue them together on the non-copper side.  How about vias.  How about alignment.  The idea of just gluing the non-copper side is not as simple as I pictured in my mind.

Another thing for sure is that my DYI board will look nothing like say the Adafruit breakout or other production-boards I repaired.  So techniques I have practiced on those production boards may no work well on a DYI board, and the end product may not look so good.

So, my ideas of how to do it was way too simplistic to begin with.  The knowledge you folks brought to the table is waking me up a bit.  I just saw the expected road block (small pitch) and lost sight of the other difficulties.

Still I am itching to try.  I have to think it some more.  I now know for sure the ADS1115-using project should wait - it is not a case of print/toner-transfer/etch/solder and done.  I had hope toner-transfer would be easy.  From what I learned here just now: It still may be easy but would be more limiting and likely not for VSSOP until I get very very good with etching it and drilling it and...

With the project that needs ADS1115 put aside, now I've to first find myself a project that uses larger than .65 pitch for the first go.  I also need to do more learning before I hit this.  So, for now it is on hold - till I found something else interesting that is > .65 pitch.

Thanks guys, for the very educational responses you have given me.  Thanks again!

Rick

[tautech]

I've never tried a double sided board.
Just the odd wire link were a trace would be. 

It's a great challenge IMO to route a SS PCB, I'll spend considerable time        untill routing options are exhausted, some overcome with TH components to provide routing pathways. Don't be too precious and do mix SMD and TH to make a PCB "work".
A good SS PCB is the holy grail, shown mine to a Dr of EE and have been highly commended.

From there the etching is easy by comparison, only the TT artwork reproduction is the tricky bit.
No room for error, if it doesn't look perfect, scrub it off the PCB and try again.
Your 5MP shoud be OK, I used a 6MP (600 dpi) untill I wasn't happy with the "blackness" of polygon pours.
Some porosity on pours can be tolerated, but for traces 15 mil or less NO.

Finding the "right" combination of printer, paper and transfer method is critical to good results.
I thought I "had" it untill I changed to my OKI 1200 dpi printer and had to rework my methods again.

Home laminator for TT, in the card carrier, 2 passes.

I did use persulphate but changed to HCl and peroxide for re-useability and lesser cost.
Now I can etch cold and always in a home made air bubble etching bath.
Aquarium air pump pushing into a needle perforated tube horizontal in the bottom of the bath.
PCB suspended in "gutter guard" plastic mesh for easy removal and checking of progress.
Recent PCB with a new etch batch was 2 minutes flat, 1 oz copper SS.

Again artwork reproduction is the key.

[Psi]

You're far better off learning how to use a PCB CAD package and paying the $10 to get the PCBs made from Elecrow or similar company.

You're also right about home made multilayer pcbs being hard to align and then drill out. It really is quite annoying and time consuming.

When it cost like the same as a few coffee's to get PCBs made you really have to ask yourself, "do i really need to make these myself"

[tggzzz]

You're also right about home made multilayer pcbs being hard to align and then drill out. It really is quite annoying and time consuming.

A double-sided laser toner board is relatively easy to align. Put ~four fiducial "standard" size holes in easily recognisable places. Use sticky tape to hold one side's paper in position, and drill through those fiducial holes. Then position the other side's paper over those holes, tape, and put through a laminator or iron in the normal way.

Without trying hard, I achieved positioning accuracy of <1mm across a 75mm board - i.e. the holes were within the pad on both sides of the board, albeit not centred.

[redshift]

You said you're concerned about turn-around time. So I thought of another method you may be interested in.

PCB Plotters/Routers are a quick and cheap way to make pcbs. They use a subtractive process where a moving drill cuts around both sides of the traces into the substrate. Having traces widths of even 0.01" are very possible. Double-sided boards are ok too they just take a bit of aligning. You still don't end up with solder-mask or plated vias, but you don't get those with etching either.

These machines are expensive(for individuals), but perhaps if you live near a university or hackerspace they might let you use one.

[PeterFW]

You can etch double sidet in one go, just buy photo board with copper on both sides.
About the soldering... the lack of soldermask is no problem.
Soldering SMD components is as easy as on normal boards, even a bit better because too much solder will just flow away instead of piling up on the pad.

Here is a crappy picture of the first board i ever made years ago and my first attempt at smd soldering, but this one is not etched at home because it is double sidet and i gave up after a few tries



But as you have pointed out, you should think about why you want to etch your own board.
I only work with SMD parts and only do single sidet boards.
No drilling, no alignment issues and no jumper wire mess.

Most times you can get away with single sidet boards when you make it big enough, if you use 1206 passives you can easily get a wire between the pads but i dont do that anny more.
There are maybe a few jumper wires from time to time but thats it, just a couple of mounting or connector holes to drill.

You can make "semi double sidet" boards, you route just one side and use the other one as one big ground/power plane.
Every via gets a little hole drilled, wire and soldered on both sides.
Of course you need a FR4 board with copper on both sides.

And i only make those boards because it is faster, way faster, i dont like soldering on proto board with smd parts, it is just a mess.
I only make small boards to test out parts or circuits.

That way when i order a proper board i have nerly all the problems solved allready.