Best printer for toner transfer

[SilverSolder]

I've been thinking about that too, but I don't have the vinyl cutter. What I'm not particularly crazy about with that idea is picking all the tiny bits of cut outs that need to go.
The results would need to be drastically better to justify all that work, and I'm not sure just how much better can it get and how much better than 10 or 20 mil I need. After all, for me, the biggest issue with diy PCB production are vias, not traces.

Agree.  Maybe some kind of hybrid method where it is only used for large black areas...

I'll be playing a bit with PCBs in the weekend, after getting inspired by this thread LOL! 

I'm going to try to find out the max size of black area that my printer can do reliably...

[KL27x]

...Vinyl sticks to the PCB but only lightly (comes off without tearing or stretching), and after transfer it does appear as if it's pressed to accommodate the thickens of toner, so it might be sealing...

Thanks. That's interesting.

When I tried pre-etch, I got splotchy copper surface that toner wouldn't stick to reliably, so a lot of traces wouldn't transfer.

Barring some difference between acid peroxide and cupric, it's possible you didn't get the board hot enough, yet. It definitely takes a bit higher temp to get the toner to transfer to my pre-etched boards. I suspect I would melt vinyl they way I do my transfers. Anyone with a 1200W heat gun knows how hot it gets. Blowdryer x 1000.

Anyway, this is what I can achieve without any special precautions. I don't have any raw boards around at the moment, so this assembled on will do

This is better than anything I ever etched with HCl and peroxide!

Oh, as standard practice, I tin all my traces as soon as I strip the toner. Black smudges are toner residue, and I have no idea how to get that off reliably. Acetone soaked rag sometimes leaves them and they never go away.

I found that acetone will draw some of the toner right into the FR-4. Depending on the color and thickness of your FR-4, it can be very noticeable, this darkening/discoloration. (On a high voltage board, it might even have practical implications? There's carbon particles in the toner to make it black!) Since my boards have the icky pre-etch surface, I just scrub the toner away same way I prepped the board to begin with. I wet scrub with stainless steel wool and a drop of dish soap. Pat the board dry, and then brush a thin layer of rosin flux over the board.

SilverSolder: Good luck. I think with some persistence and attention, you might find the printer isn't the limiting factor.

[SilverSolder]

By overprinting a thin 8" x 11" frame several times on the same piece of paper, I noticed that you can't rely on the printer (Laserjet 2300) being more accurate than about plus minus 0.8mm across the page,  in other words, the printout is going to be slightly distorted depending on roller slippage as the sheet of paper worms its way through the paper path, shimmying, rotating, and generally misbehaving along the way (from a pixel peeping perspective).

The best accuracy on this printer is in the top left quadrant of the page, which consistently got less than 0.25mm distortion.

This isn't something anyone would ever notice on normal printed output.  But the precision is definitely not as good as the two inkjet printers that I compared with.  Both the inkjets I tested had less than 0.1mm error across similarly sized artwork.

So this is yet another thing to consider when you choose your process -  i.e. there is an absolute limit to the precision that a laser printer can do, so if you are trying to make a very large double sided PCB (like 8" x 10"), you will probably struggle to get things lined up accurately across the whole board (your printer may be better, or worse, than mine...  try the test with a large thin frame!).

Then there is the issue of large black areas not receiving enough toner...  I brute forced that,  by overprinting the same page twice,  with slightly shrunk artwork on the second pass (to avoid registration issues)!  Works a treat, completely eliminates the problem of "not enough toner" on large areas.

1. Why persevere?  - toner transfer is still simpler and faster than photolithography.   Just keep in mind that if your board is larger than 4-5 inches, the lack of precision can become a real issue when aligning front and back.  For some boards, this won't matter (single sided, for example).

2. Why not just get it made in China? -  it is super cheap to have small PCBs made in China.   But try getting a quote for an 8" by 11" PCB with 2oz copper...


Basically, if you make PCBs at home...   there is an argument for being able to do both toner transfer and photolithography, depending on what you are doing!

[m3vuv]

i had similar issues with my hp laserjet cp-1515n printer,even sanding the copper clad with p600 wet or dry and cleaning with acetone the toner mostly wouldnt stick to the board,brought a cheap copy black toner cartridge now it works every time,just my experiance.

[m3vuv]

i use hcl and peroxide mix for the etchant,when it stops etching i have a small aquarium pump bubbling air through the solution,i use a small plastic takeaway curry container,i put the solution in the microwave for a min first,then use it,normaly does a 3x6 board in about 10 mins.

[KL27x]

i use hcl and peroxide mix for the etchant,when it stops etching i have a small aquarium pump bubbling air through the solution,i use a small plastic takeaway curry container,i put the solution in the microwave for a min first,then use it,normaly does a 3x6 board in about 10 mins.

But you can't do that without peroxide, right? I don't understand why reality is different for me and for the rest of the internet. According to Adam Seychell's figures, you need 1.85L of cupric etchant in order to etch a 1oz single-sided 3x6" board. My reality matches this. How much etchant do you warm up in your curry container?

And an aquarium bubbler won't regenerate that "in real time" as you etch, IME. It will take multiple hours.

[KL27x]

My final attempt at a fish bubbler was a 4 output bubbler with 4 bubblers in 3-4L tank. This was terrible. I etched scores of boards with it, successfully. But in practice I was constantly adding acid and peroxide to get boards to finish etching. And etch quality was not significantly better than etching with acid peroxide.

[m3vuv]

the amount of liquid is normaly about 1/3 of a pint,i start of with peroxide,then after a couple of boards just use the pump to bubble air thru it,i normaly heat the mix to about the same temp you drink coffee at,i find it slows as it cools tho so normaly pop it back in the microwave halfway thru,i seem to get good results,when i finish i put the solution in an old coffee jar and reuse it,seems to last forever.

[KL27x]

the amount of liquid is normaly about 1/3 of a pint,i start of with peroxide,then after a couple of boards just use the pump to bubble air thru it,i normaly heat the mix to about the same temp you drink coffee at,i find it slows as it cools tho so normaly pop it back in the microwave halfway thru,i seem to get good results,when i finish i put the solution in an old coffee jar and reuse it,seems to last forever.

Gotcha. That's really not much different than using freshly mixed acid peroxide, in my experience. It's good to do it like this, to keep your waste concentrated, but it's not really giving you significantly higher etching accuracy, IME. FYI and FWIW. Hence, the thousands of people who are intentionally making cupric (with no way to practically benefit from it), are just making more waste for no reason. Seems like out of every thousand people who do it, 990 of them just created ionic copper waste for no reason. Cuz aquarium bubbler=cool.

Your bubbler is window dressing.* I wonder if xaver60 will also provide further details to his magic brew. He states no peroxide in the last 3 years and etches in 7 minutes. But I imagine he must be putting a tiny test strip of board into a beaker of this stuff. Scaled up for a decent sized board, you also have to scale up that beaker. Then you find the practical issues.

There was a thread from maybe 3-6 months ago. Guy built a gorgeous vertical bubble etch tank, 3L large, all plexiglass, with CAD-modeled parts. Filled with cupric. On his first few square inch test board, the 3L of cupric turned dark and slowed down, then he added peroxide. When he upgrades the regeneration/bubbler, he can see what cupric really does.

*Peroxide is a really powerful oxidizer. It's very unstable and really wants to donate that oxygen. Which in this case, it does to the elemental copper. Cupric can't etch copper oxide, at all. This is the reason that you need some excess of acid in order to etch with cupric; the acid is necessary to remove any copper oxides that will incidentally form due to dissolved oxygen and/or airbubbles in the solution. When you put peroxide in the solution, the copper is very aggressively oxidized, and you block the cupric reaction almost completely, for practical purposes, no matter how weak you keep the acid. The acid pathway is the only way that copper oxide is going to be removed.

[moffy]

I watched this video:
Truly beautiful detail from a flatbed inkjet printer. Don't know how to get one at a reasonable price.

[Domagoj T]

That looks very nice, but how does it perform with double sided boards? For me the ability to make them is crucial.
I suppose you could make templates with registration marks where you drill holes in bare PCB and have pins on the printer bed.

[KL27x]

Hmm, that's a good question. If you don't move the magnets and just flip the board, that only works if you were perfectly centered and square to begin with. Seeing as all the work that went into this, I would guess he can figure out a solution, if he hasn't already.

I make boards like this regularly, no problem, with toner transfer. This looks exactly like one of my boards, routing and clearances and trace sizes. 0.2mm is roughly 8 mils, which is my default. The traces in the bottom center/left of his board look a bit rough, though. I mentioned this earlier, but I don't remember if it was in this thread. About not leaving wide spaces between traces, unless necessary. Even if there's extra unused space available, I would have bundled those traces together to maintain 0.2mm spacing!

I didn't try 0.1mm. When I tried 6/6 (0.15mm/0.15mm), it worked fine over 99% of the board. For a small board it might work. The problem was a break only every so often (could be seen in the print, itself, prior to transfer). So that short test pattern at 0.1mm/0.2mm looked good, but it might not be the end-all.

Otherwords, this looks good. Execution is impressive. Results not that much. The setup/alignment and extra-custom things you have to do might not even save time but add it, until this is more refined. And I can already make that board 30 times out of 30, no problem. Cash money on it. Sure one might go bad, but I'll bet on it.

[moffy]

KL27x, glad to hear of the quality you can get from toner transfer. Going to buy a new laser printer tomorrow, the KYOCERA P2040dw to replace my older Brother. You can get toner replacement for $40 per 7,000 sheets (5% coverage). I'll see how it does for toner transfer.

[SilverSolder]

Hmm, that's a good question. If you don't move the magnets and just flip the board, that only works if you were perfectly centered and square to begin with. Seeing as all the work that went into this, I would guess he can figure out a solution, if he hasn't already.

I make boards like this regularly, no problem, with toner transfer. This looks exactly like one of my boards, routing and clearances and trace sizes. 0.2mm is roughly 8 mils, which is my default. The traces in the bottom center/left of his board look a bit rough, though. I mentioned this earlier, but I don't remember if it was in this thread. About not leaving wide spaces between traces, unless necessary. Even if there's extra unused space available, I would have bundled those traces together to maintain 0.2mm spacing!

I didn't try 0.1mm. When I tried 6/6 (0.15mm/0.15mm), it worked fine over 99% of the board. For a small board it might work. The problem was a break only every so often (could be seen in the print, itself, prior to transfer). So that short test pattern at 0.1mm/0.2mm looked good, but it might not be the end-all.

Otherwords, this looks good. Execution is impressive. Results not that much. The setup/alignment and extra-custom things you have to do might not even save time but add it, until this is more refined. And I can already make that board 30 times out of 30, no problem. Cash money on it. Sure one might go bad, but I'll bet on it.

To my mind, inkjet printing directly on a possibly slightly warped PCB is less "ideal" than to printing on gel-coated transparency (which requires no modification of the printer, and can be re-used again and again, possibly saving work/time later?).

[moffy]

What papers do you guys use for the toner transfer? Have ordered my printer but unsure about paper.

[tooki]

What papers do you guys use for the toner transfer? Have ordered my printer but unsure about paper.

I have yellow Chinese toner transfer paper from eBay or Ali. Works well. My results have been far better than what people who use magazine paper and other such hacks report.

[SilverSolder]

I get nice results with the Pulsar paper.  Their customer support is good/helpful too (small business) by email or phone.

[KL27x]

I have used successfully made boards with plain printer paper and magazine/catalog paper. Value of your time has to be very low for this to be worth it, though. The only thing I've read about which I never got to work was photo paper. Was a guy (in Denmark) who swore up and down about a specific brand and sku of photo paper. I never figured out the special sauce to make that work, but perhaps the formula was different in America or maybe it had changed.

If you make more than $1.00 an hour at your day job, you are probably better off buying transfer paper. From what is easily available in the US, it seems like there's 3 choices.

1. PnP Blue: this is the most sophisticated stuff. It is a lamination of 2 different plastic sheets. The side you print on is rough-textured to help prevent the toner from squishing/smearing. Where the toner sticks, this entire top sheet tears away from the backing sheet, covering the toner transfer with a layer of thin plastic. It's around a dollar a sheet in 10 packs. But you can buy 100 packs for like 65 dollars, if you use a lot of it.

2. Pulsar: this is the most expensive in the US. It's a little more than a dollar per sheet in 10 packs, purchased conveniently from Mouser or Digikey. It's a very thick and stiff paper, almost a thin cardstock, coated on one side with dextrin. Print quality doesn't appear to be affected by the dextrin, and after you transfer, you dunk the board in water for 20 seconds; the paper peels away leaving just the toner. According to the manufacturer, it is sensitive to humidity and can go bad after so long. I have never experienced this, but I don't get high humidity. Since there is no retail bulk discount, there's no major incentive to stock up on it, anyway.

3. Yellow paper off eBay. I'm 99% sure this is printer label backing paper, sold on eBay as toner transfer paper. It is dirty cheap. But the 100 pack I bought came all rolled up in a tube. So the paper is very thin and curly and I dunno how other people use it. I imagine they have a system to uncurling it, or maybe even tape it to regular paper... but that would defeat the cheapness. It would be nice if you could just put the 100 sheets in your printer feed tray. Print, cut out what you need, and toss the other 90% blank paper into the bin. But at least with my printer I don't see that working. 

There are pros and cons, depending how you do the transfer. I've used over 100 sheets of PnP Blue. That was a lot of boards; results are great, but I experienced a significant amount of failures/redo's either in transfer or after the etch. I'm not suggesting that someone else can't get it to work 100%. But for me, I had failures probably in the lower double digits % for relatively small SS boards; the double-sided boards is what really made this bad for me. The method I have settled on for the last many years is too hot for PnP; it melts. I use Pulsar. Now, if I need one board, I make one board... not 2-4, just in case. I haven't failed/redone a board ever since, probably 4 years, now, and a 10 pack of pulsar per year. I bet I could get similar results with the yellow backing paper, but for me it's not worth the hassle for the cost savings. (TBH, that 100 sheets of yellow paper I got in the snail mail, I looked at it. Then it went straight into the garbage can.  )

So it the best paper depends. What your process, and why you're doing it. If you want to save money over ordering a board from China, we might be in the wrong century for that.

[moffy]

Thanks for your advice. I will try and track down one of the three papers listed by "tooki", "SilverSolder" and KL27x.

p.s. Managed to order some of the yellow paper. Should be here in a few days along with the new printer. Thanks again.

[Domagoj T]

Try to find some vinyl as well. It's really good and you should probably be able to get some offcuts for free if you ask nicely, but it's not expensive even if you need to buy some. It's the stuff used in sign making.

[tooki]

1. PnP Blue: this is the most sophisticated stuff. It is a lamination of 2 different plastic sheets. The side you print on is rough-textured to help prevent the toner from squishing/smearing. Where the toner sticks, this entire top sheet tears away from the backing sheet, covering the toner transfer with a layer of thin plastic. It's around a dollar a sheet in 10 packs. But you can buy 100 packs for like 65 dollars, if you use a lot of it.

Where do you get this for $65 per 100? I only found it for over $100.

[KL27x]

completely edited for helpfulness:

^Direct from Techniks.com. But it appears my memory has failed me or the price has changed. Probably the former.

As of today, the Pulsar is 13.71 for 10 sheets, at Mouser US. At $105 for 100 sheets, PnP is still cheaper. In bulk, anyway. So hmm. Unless I check my bank statements, I would say I probably misremembered the price.

But in smaller quantities, PnP looks to be more expensive.

[rfguy2020]

KL27x, glad to hear of the quality you can get from toner transfer. Going to buy a new laser printer tomorrow, the KYOCERA P2040dw to replace my older Brother. You can get toner replacement for $40 per 7,000 sheets (5% coverage). I'll see how it does for toner transfer.

Any experience with Kyocera?

[moffy]

I have spent quite a while just setting my Kyocera printer up. Had to make an extension table to fit it in the back room. It is scary fast, and the first print when it comes out, it looks like there is a puff of smoke! It also has so many features that it made me nervous setting it up. But I have direct WIFI and USB printing configured, and the driver package is first rate. Will let you know when I do my first pcb transfer. I also found my old A4 Creative laminator. Any advice on how to use it?

[tooki]

KL27x, glad to hear of the quality you can get from toner transfer. Going to buy a new laser printer tomorrow, the KYOCERA P2040dw to replace my older Brother. You can get toner replacement for $40 per 7,000 sheets (5% coverage). I'll see how it does for toner transfer.

Any experience with Kyocera?

I’ll report back. I’m about to print some transfers on the Kyoceras at work (one B&W, one color).

P.S. can confirm that Brother won’t work. It doesn’t adhere to the PCB at all, but does hold onto the transfer sheet rather well.