PCB manufacturer, how important is temperature?

Using Arduino General Electronics
#1

Hi,

I make my own PCBs. I design them on Eagle cad, print them on transparent sheets (on a laserjet printer), expose the board to UV, develop the boards in Sodium Metasilicate then etch them in ferric chloride. I always use the same chemicals and same boards from the same supplier, I expose all the boards to UV for the same amount of time.

My problem is that I get inconsistent results.

About the only thing that I don't measure/control is the temperature of the chemicals, in the past I've figured this is ok because I can check the progress of these processes simply by visually inspecting the board every few seconds. (I do heat the chemicals, but only control it by feeling the sides of the tanks)

It's a bit hard to say exactly what's going wrong with some of the boards, but basically the board comes out looking 'dirty', I think that maybe the etch-resist is being attacked by the ferric chloride, but if i remove the board any earlier then the bits that do need etching are still joined together!

Advice and suggestions please?

#2

do you stir it when etching ? If you do not, you get this 'dirty' result .
the best way to do that is to lift one side then the other one as regularly as you can during the whole etching process.

#3

do you stir it when etching ?

yes. I use vertical develop/etch tanks and jiggle the boards up and down throughout the process.

#4

do you sensitize your own boards, or buy pre-sensitized boards. if you do your own do you keep the tempreture low during sensitizing? And store the sensitized boards where they will stay cool?

same temp issues when developing the exposed boards. Start with CLEAN boards.

as for etching, yes warm chemicals work better/faster and agitation is essential. you might try adding a small amount of hydrogen per Oxide to the etchant (especially used mix), but watch out, is speeds up the etch process.

Fresher etchant will give sharper (faster) etches. exposure time, developing time, and photo-mask quality (pattern density) will have major effects upon results. it is better to use a UVC source (mercury vapor/unfiltered fluorescent) than a high wattage incandescent.

That is about all I can come up with.

#5

do you sensitize your own boards, or buy pre-sensitized boards.

I buy pre-sensitized boards from mega electronics.

Normally I only make one version of a board before moving on to a new version, this always requires a new set of artwork to be printed, so in the past I've put down the discrepancy in output quality to variations in the thickness in the toner of the negative (which in turn effects the UV exposure of the board) however, for my latest project I've made two identical version of the same board. The only difference (that I can think of) is that for the 2nd version the developer had been in the tank for a few minutes more and had more time to warm up - so my current theory is that if the developer gets too warm it will remove too much photo-resist.

#6

Everything depends on temperature.

#7

What are you using for the transparencies? I have had issues with using some cheap ones from Computer Grafix. There is actually a "clean side" and a side with a coating (which is the side you are supposed to print on) that I got backwards once without noticing. I tried several times to get a workable board from that print and each time I had similar issues. Ground planes seemed "mottled", and traces looked like a mouse chewed on the edges. Some of the bare fiberglass areas had what looked like copper spatters, perfectly round copper drips.

Then I looked at the transparency under light and magnification and saw the reason for the mottled look, being the toner was also that way. I also noticed some scratching in the opposite side and figured out I had my print on the clear plastic side. I used some 91% alcohol and the coating came right off. I reused the print and used less exposure (by 3 min) and the boards came out perfect. Whether it was just luck of the draw or that the coating was interfering with the exposure, I don't know.

I was looking to see if this was a better way to go with doing this
http://www.pcbfx.com/main_site/pages/products/toner_foils.html
It means getting toner transfer paper and the foil, which is worth the cost if it means not throwing bad boards and time in the trash. I just need to find out if it handles .01 traces or not.

#8

How are you UV exposing

#9

What are you using for the transparencies?

I use laserstar sheet from mega electronics.

How are you UV exposing

I have access to a professional PCB/UV exposure unit.

When I make multiple copies of the same PCB I use the same transparencies and expose using the same UV light box for the same period. The only real difference from PCB to another is the number of times I’ve used the fluid before and the temperature it might be at.

#10

After my comment I realized that it might be better to raise the temperature as a function of how "used" the etchant is, to try and keep the etch rate constant.

#11

Fulliautomatix:
I use laserstar sheet from mega electronics.
I have access to a professional PCB/UV exposure unit.

The only real difference from PCB to another is the number of times I've used the fluid before and the temperature it might be at.

Do not reuse sodium metasilicate developer.

At 5g per litre dilution it is only good for 1 use.
Higher concentration may give more uses but inconsistent development.

Using mega boards and the mega exposure unit i would expect an exposure time of around 2 1/2 minutes.

#12

Using mega boards and the mega exposure unit i would expect an exposure time of around 2 1/2 minutes.

Correct - exactly the timings I use.

Do not reuse sodium metasilicate developer.

At 5g per litre dilution it is only good for 1 use.

Really? why?

I use 2g per 400ml (similar to the concentration you suggest) but I'll use it several times (I only ever make quite small boards - limited by Eagle Cad

#13

Fulliautomatix:
I use 2g per 400ml (similar to the concentration you suggest) but I'll use it several times (I only ever make quite small boards - limited by Eagle Cad

The chemistry in principle is simple but the details are complex.
Sodium metasillicate is cheap as chips and it is not worth reusing it.

There are many possible problems and i would suggest simplifying things to get rid of the variables as much as possible.

Made up solutions can have short shelf lives.
Your tanks may be suitable for production use but not occasional use Are they bubble ?

Both sodium metasillicate (developer solution) and sodium persulphate (etchant) have a short shelf life in solution form due to atmospheric effects. Powder lasts forever.

Ferric chloride etchant lasts forever but infrequent use can lead to sediment clogging the bubble bar.

Fulliautomatix:
Correct - exactly the timings I use.
Really? why?

WIKI on and alkaline and alkinility.

They have different meanings
Buffering capacity is also a good search term.

#14

Should probably have done this first.
Attached are two photos of my boards. Both are made using a piece of the same PCB, using the same artwork on the same UV box for the same time (2min 30sec). Each was developed/etched in the same solution but a few minutes apart. The first one is quite clean and shinny, the second one has a dirty appearance.
IMG_0165.JPG
IMG_0166.JPG

Your tanks may be suitable for production use but not occasional use Are they bubble ?

My tanks are homemade, the etch one is a bubble tank, the develop one isn’t. The tanks do drip a little if I leave them full for more than an hour or two so after use I drain the solutions into sealed containers to prevent spills.

#15

Having seen your boards i am almost certain that your problem is exposure / contrast.

Toner or inkjet printing only gives decent blacks for small areas.
Printing just tracks works better than printing large areas.

You have large amounts of copper there.

Try just printing the tracks.

Ground and power planes are virtually impossible to do using standard printers, you need photo quality black.

As i said earlier use fresh developer each time.
Your first example is actually quite good considering its a print rather than a photo.

Fulliautomatix:
]
My tanks are homemade, the etch one is a bubble tank, the develop one isn't. The tanks do drip a little if I leave them full for more than an hour or two so after use I drain the solutions into sealed containers to prevent spills.

Dip tank is fine for etching.

For development for occasional use i would suggest a tray with fresh chemiclas each time so you can eyeball it. I takes away the temperature critical bit that way, time is generally short , less than 30 sec.
Chemical requirement is small also , all you need is about 1/8 in of solution.

#16

Despite my earlier post , i may have missed some bits.

Fresh chemicals every time for consistency (does not apply to etchant though).

Exposing the negative can damage it if not thoroughly dry, the ink can run onto the workpiece.
This can make repeat exposures poor.

With inkjet transparencies i would suggest an oven dry first at 40 deg c for 1/2 hour or so.

#17

Having seen your boards i am almost certain that your problem is exposure / contrast.

Just to clarify - I think what you're saying is that the printer (I use a laser jet) can't deposit really thick and impervious layers of black ink on the transparency sheet. So, when the board is exposed to UV a little bit of the UV light will make it through the black stuff (tracks and ground planes) so even these areas will be partially exposed and then partially developed?

As i said earlier use fresh developer each time.

Understood.

You have large amounts of copper there.

Try just printing the tracks.

Ground and power planes are virtually impossible to do using standard printers, you need photo quality black.

I used to print tracks only, but added ground-planes to speed up the etching times (i.e. to reduce the amount of copper that needs to be dissolved). Also, I read somewhere that a ground-plane was a good idea for isolating adjacent signals from one another.

Exposing the negative can damage it if not thoroughly dry, the ink can run onto the workpiece.
This can make repeat exposures poor.

Presumably this is only an problem with inkjet artwork, not laserjets?

Your first example is actually quite good considering its a print rather than a photo.

Thanks, I have been slowly improving my board quality over the years. Most of the manufacturing process is standardised and repeatable, the only exception is dunking in the developer and etchant where I judge the temperature of the fluids by the feel of the tank, and I judge the timing simply by removing the board and doing a visual inspection.

I think I will look at installing thermostats on the tanks to remove another variable.

In the past I have pondered building a automated etching machine - something that will set the temperature of the fluids automatically and then insert/remove the PCBs for precisely timed intervals

#18

Fulliautomatix:
Just to clarify - I think what you're saying is that the printer (I use a laser jet) can't deposit really thick and impervious layers of black ink on the transparency sheet. So, when the board is exposed to UV a little bit of the UV light will make it through the black stuff (tracks and ground planes) so even these areas will be partially exposed and then partially developed?

Yes, i find inkjet better than laser but doubling up can help considerably

I used to print tracks only, but added ground-planes to speed up the etching times (i.e. to reduce the amount of copper that needs to be dissolved). Also, I read somewhere that a ground-plane was a good idea for isolating adjacent signals from one another.

Etch time should not change but solution will last longer.
Ground planes require large areas of Black me these to not print well

In the past I have pondered building a automated etching machine - something that will set the temperature of the fluids automatically and then insert/remove the PCBs for precisely timed intervals

Heated bubble tank is good for etching.

As i mentioned i tray develop with warm water and no temp control.
Important bit is to eyeball it so you can see when it is done.

Another advantage of tray developing is that not much is used each time.

Tanks hold much more. unless you replace each time it will get difficult to judge the time.

#19

Fulliautomatix:
, I have been slowly improving my board quality over the years. Most of the manufacturing process is standardised

That is how most of us did it.

My biggest ‘score’ though was trying out all my mates printers and some in stores until i found one that gave the best results

Generally i found inkjet better than laser.

For film i found Rymans own coated brand to be the best , slightly better than NOBO

Mega laserstar and jetstar was not so good.

#20

My biggest 'score' though was trying out all my mates printers and some in stores until i found one that gave the best results

I did something similar, except that the conclusion I came up with was that the Samsung ML-2525W laserjet was the best (among the ones I tested anyway).
I also came to the conclusion that although a cheap toner cartridge may be okay for printing off a few emails, when printing PCB transparencies it really has to be an OEM cartridge.

For film i found Rymans own coated brand to be the best , slightly better than NOBO

Mega laserstar and jetstar was not so good.

Ok, I'll try rymans - must admit last time I did any testing I found the stuff from Mega electronics to be the best.

Another advantage of tray developing is that not much is used each time.

I understand what you mean, but I've made my own custom vertical tanks that are built to work with boards that match the size limit of the free version of Eagle CAD - this means I only use 225ml of developer and etchant for each board I make. The other advantage of using a very small amount of liquid is that I can warm the etchant up in just a few minutes with an internal heater (a 20 watt bulb)