Toner Transfer

[BradC]

Ok, so I've been doing toner transfer PCBs for a few years now. I have a great HP2200d on my desk that serves double duty as a general purpose monochrome office printer, and when required I pull the gold-plated genuine HP toner cartridge out of the zip-loc bag and use it for toner transfer.

The *problem* is that these cartridges last about 5 years *tops* before the wiper blade takes a set and I get ghosting on the images. I've put up with this because it meant I only had to buy a new HP cart every 5 years and the whole thing is so damn reliable.

Now, I'm due a new cart due to ghosting (I can make a board, but it can only be smaller than about 8cm as that is roughly the drum circumference) and find the clusterfuck that is now HP (thanks Carly _there_really_are_no_words_for_how_I_feel_about_this_oxygen_theif_ Fiorina) has actually discontinued the cartridge. Experience says that cheap toner doesn't transfer as well (I can do 8/8 down to 6/6 with the genuine HP toner and Puslar paper). So I'm a bit up the creek without a (paddle/outboard/jet) right now.

So, what I'm looking for is a recommendation for a printer that does awesome toner transfer, has a reliable supply of genuine toner, a 500 sheet feeder + duplexer and costs fuck all (my 2200d+feeder was less than $50). A big ask? Maybe.

But seriously, let's take cost off the table. Does anyone have any recommendations? The 2200d makes an awesome office printer as I have a complete spare unit for parts, a full service manual and the office grade toner is 0.6c per page so it costs nothing to run. If I had to buy a dedicated "board only" printer for the 10 times a year I make boards, I'd figure something out and slide it past the wife.

It's just the aftermarket toner for the 2200d doesn't make good boards (or I haven't yet found one that does reliably). Maybe this is the push I need to move from toner to UV. Hrm.. build a light box, deal with dry resist lamination and producing adequate negatives, or buy a new printer than does awesome toner transfer... what to do?

Ideas? Anyone? Anyone? Bueller? Bueller?

[Buriedcode]

I think you'll get quite a few replies along the lines of "just get your PCB's manufactured, its cheaper".   I still occasionally make my own for single sided jobs using toner transfer, and whilst I can have a working board in a few hours, with all the hassle of it, and the arse that is drilling accurate 0.6mm holes, I end up with something that isn't that good, and actually costs more than getting 10 of the same boards from China in 3 days.  The only except is one-off large boards (200mm x 200mm +).

With that said I use a Cannon LBP2900i with generic blue toner transfer paper and using a laminator could manage 8/8 spacing with ease.  Bought a replacement toner cartidge of ebay - looked like a cheap copy, and got the same results.  Its one data point, but it works.

[Jwillis]

 I have had good luck with a Brother MFC 9440 CN I picked up at a second hand store for 50 bucks.I just use the refill kits and seem to work fine.Rollers and belts are easy to find as well.
I'll be retiring the toner method soon in favor of negative dry resist .But a good printer is still essential for creating the negatives.The box I'm building uses a digital microwave oven timer ,10 amp SMPS, UV led strips attached to an aluminum plate and a key pad I made (using the toner method) .Kept the fan to keep things cool .Unfortunately the timer seems to have a cool down for anything over 5 minutes.So I'm rewiring the solenoids to use the lamp circuit instead of the magnetron circuit. Lamp solenoid isn't rated high enough to handle the SMPS.
 Life is taking me away from my projects ,but I'll post some schematics and photos when completed and tested.Fingers crossed.

[In Vacuo Veritas]

just get your PCB's manufactured, its cheaper

[SeanB]

F8unny enough I get better results from refurbished cartridges than from the OE HP and canon cartridges. The C4096A cartridge is not one I use any more, but here the refurbished one is around $35 for a remanufactured unit from my favourite supplier, which has a new drum, new wiper and a nice dense toner fill in it as well. Bonus is the refill also gives about double the pages before it runs out as well.

[TrickyNekro]

Go for the UV method, you´ll never regret it, I tell you! You just need transparencies and an UV chamber (easy peasy stuff).

You´ll get better resolution than toner transfer and you won´t worry about rolling out the good cartridge from the basement.

 

[SiliconWizard]

I did that for a while more than 10 years ago. IIRC, I was using a Brother monochrome laser printer and tried various photo papers. This was working fine but tended to mess up my printer's drum, as the kind of paper adapted to this job is not really adapted to laser printers. Towards the end of this era, I stumbled upon a company that was selling a special paper for this use, and it was working rather well (much easier to peel off than any other kind of paper). Had to dig up to find this back: the company was called Pulsar, more info there: https://pcbfx.com/main_site/pages/start_here/overview.html

And yes, PCB manufacturing costs have plummeted so much that, well... yes, just get your PCBs manufactured. All the more that decent FR4 copper clad boards are not cheap. And the chemical mess... and getting anything else than single sided boards is quite a feat.

If you have access to a small CNC machine, that's another option. It does not need any chemicals (at least for the "etching" part) and no messing with paper and heat transfer, but it's very noisy and will get you copper/epoxy dust everywhere.

[SilverSolder]

I find that making large PCBs (e.g. 12" by 12" or larger) is still significantly cheaper as a DYI.   UV photoresist is best for large boards, the toner method doesn't cover large areas well.  For smaller boards, toner can work very well and has less steps than the UV method.

[BradC]

And yes, PCB manufacturing costs have plummeted so much that, well... yes, just get your PCBs manufactured. All the more that decent FR4 copper clad boards are not cheap. And the chemical mess... and getting anything else than single sided boards is quite a feat.

It's the Pulsar paper I use. Since switching to that many years ago now, I've had consistently good results. I can't say I've had any issues knocking out a quick double-sided board from time to time. I've attached a pic of one I did a while ago. Double sided, using 0.25 wire for vias with 0.3mm holes. That took me 2 hours from plot to power-up, and that's hard to beat even using a proto-service. Yeah, it's ugly but it helped me beat out the design failures pretty quick.

I'm sure one day I'll get around to building a light box and going the UV method, but aside from the printer issue I do get workable results with toner transfer *fast*.

My biggest issue with getting stuff fabbed is shipping. Sure Seeed does a 100x100 for $5USD, but it's either 3 weeks shipping for $10 or 3 days dhl for $20. So 2 to 5 times the cost of the board. Bear in mind these are USD, and our exchange rate is pretty crap too.

When you look at that, being able to knock out a one off proto in an hour starts to look more attractive.

[BradC]

So I just did a quick comparo on 2 boards between Seeed and OSHPark (and ended up ordering them through Seeed).

Excuse the differences in spec as they are not relevant to my requirements (ie HASL vs ENIG, colour and the like).
Both boards are single sided, but the standard service gives PTH so they are just trackless on the other side.
Board 1 125x18mm
Board 2 41x82mm

Seeed I get 10 of each, Oshpark 3.
I chose 3 day DHL with Seeed to get a parity cost comparison (I could have saved $14USD if I wanted to wait 3 weeks). Oshpark was the standard shipping (also takes ~3 weeks). An extra $23USD for faster shipping.
Both $44USD.

So $61AUD for a set of boards for 2 designs.

I did a quick comparo of OurPCB, but again the minimum postage is >$20AUD.

Even with expedited shipping, it's still at least a week. This is why I can afford to lose an hour or two printing, etching and drilling even a double sided board when I want it in a hurry.

[tooki]

Ok, so I've been doing toner transfer PCBs for a few years now. I have a great HP2200d on my desk that serves double duty as a general purpose monochrome office printer, and when required I pull the gold-plated genuine HP toner cartridge out of the zip-loc bag and use it for toner transfer.

The *problem* is that these cartridges last about 5 years *tops* before the wiper blade takes a set and I get ghosting on the images. I've put up with this because it meant I only had to buy a new HP cart every 5 years and the whole thing is so damn reliable.

Now, I'm due a new cart due to ghosting (I can make a board, but it can only be smaller than about 8cm as that is roughly the drum circumference) and find the clusterfuck that is now HP (thanks Carly _there_really_are_no_words_for_how_I_feel_about_this_oxygen_theif_ Fiorina) has actually discontinued the cartridge. Experience says that cheap toner doesn't transfer as well (I can do 8/8 down to 6/6 with the genuine HP toner and Puslar paper). So I'm a bit up the creek without a (paddle/outboard/jet) right now.

Bear in mind that the Canon EP-32 toner cartridge is the same as the HP 96A. (and since Canon almost certainly made the print engine, they're probably the OEM on the cartridge, too!) There seems to be ample stock of that, even if Canon has stopped making it.

How hard would it be to swap out just the wipers??

[BradC]

How hard would it be to swap out just the wipers??

I'll soon find out.

Given I have 3 genuine HP cartridges that are close enough to full and all exhibit the fault to varying degrees (one to the point it's near enough unusable), I've decided to attack it from another angle. So I've ordered a pack of after-market wiper blades (and some other contingency items) with the intention of seeing how much damage I can do, and mess I can make swapping out the wipers.

A couple of very interesting posts on some printer forums about these specific cartridges. Most of them indicate in about 2002 HP changed the design of the wiper blade (purportedly with the intention of limiting its life to try and limit refills). Unfortunately there are some seriously old NOS cartridges out there (one I got not long ago was made in 2007) with the result anything more than about 5 years old almost certainly will suffer this problem. Given the plethora of NOS genuine cartridges available, I figure I can afford to stock up as the age, and just replace whatever needs replacing to keep them going.

I do this sort of thing for the ink carts on my DesignJet 750, so I'm used to getting messy.

[IdahoMan]

I would like to try etching many items as well as PCBs (faceplates, blades, etc.), and other materials than just copper.

I have been trying several of the toner-transfer methods out there, but the transfer is very poor. I am etching with saltwater and current via a 6v/2a wallwart, but it eats away at the toner as well as the bare areas.

[KL27x]

I have been trying several of the toner-transfer methods out there, but the transfer is very poor. I am etching with saltwater and current via a 6v/2a wallwart, but it eats away at the toner as well as the bare areas.

This sounds like undercutting. Different etchants have different levels of undercutting. Electroetching is perhaps "too fine" and gets into tight spaces, causing the toner to separate, early. This is also a problem with HCl and hydrogen peroxide etchant. The thicker/deeper you want to etch, the more you have to pay attention to this.

Most of them indicate in about 2002 HP changed the design of the wiper blade (purportedly with the intention of limiting its life to try and limit refills).

Modern progress at work.

[jmelson]

You can get replacement wiper blades through fixyourownprinter.com for most printers.  I buy old carts and often have to put in a new blade to get them working for general use.  Still much cheaper than the price of new carts.

But, if you've ever tried dry film resist, you will probably never go back to toner transfer.

Jon

[jmelson]

I have been trying several of the toner-transfer methods out there, but the transfer is very poor. I am etching with saltwater and current via a 6v/2a wallwart, but it eats away at the toner as well as the bare areas.

This sounds like undercutting. Different etchants have different levels of undercutting. Electroetching is perhaps "too fine" and gets into tight spaces, causing the toner to separate, early. This is also a problem with HCl and hydrogen peroxide etchant. The thicker/deeper you want to etch, the more you have to pay attention to this.

I use dry film resist, in a heated FeCl spray etcher.  When everything works right, the board is etched in about 1.5 minutes.  Exposing and developing the resist is the critical step, any dry film that is left in the areas that are supposed to be developed away will lead to slow etching and undercutting.

Jon

[tautech]

I would like to try etching many items as well as PCBs (faceplates, blades, etc.), and other materials than just copper.

I have been trying several of the toner-transfer methods out there, but the transfer is very poor. I am etching with saltwater and current via a 6v/2a wallwart, but it eats away at the toner as well as the bare areas.

Toner porosity will be the issue and while a glossy transfer medium can help a higher DPI printer is needed to overcome porosity issues. Shorter etch times and/or heated etchant can help too.

[KL27x]

The way I do toner transfer, there is no porosity issue. Power planes come out solid. If you get the toner hot enough to completely melt, it fuses the dots together. You just have to pre-etch the board so that the toner stays where it's supposed to. And if you want the cleanest most accurate etches, you cannot use hack etchant like HCl and peroxide.

But, if you've ever tried dry film resist, you will probably never go back to toner transfer.

That might be true for the vast majority of people. But if you learn how, toner transfer is extremely reliable and probably a little bit easier.

[tautech]

The way I do toner transfer, there is no porosity issue. Power planes come out solid. If you get the toner hot enough to completely melt, it fuses the dots together.

Yes well that all depends on your printer, toner, dpi, transfer medium and transfer temp settings, but agreed that result is what we strive for.
Change any one thing like getting a new printer and all bets are off and back to the drawing board to rediscover a recipe that works well again.

You just have to pre-etch the board so that the toner stays where it's supposed to.

Actually a pre-etch is not required if the gloss is first removed from the copper. 000 steel wool or 1000 grit wet and dry followed by a cleaning with water or IPA works fine too.
Of course contamination with skin oils from handling is to be avoided.

[KL27x]

Actually a pre-etch is not required if the gloss is first removed from the copper. 000 steel wool or 1000 grit wet and dry followed by a cleaning with water or IPA works fine too.
Of course contamination with skin oils from handling is to be avoided.

I'm afraid you and I are not on the same wavelength. This is why so many people are better off with UV. In the surface, toner transfer is simple. But there are some minor nuances to get consistent and accurate results. And the signal to noise ratio on the subject is very very low and expectations and techniques and results vary quite a bit. A lot of the information is old. And there are a lot of recent videos and webpages about it. Mostly with old information. Even 20 years ago, people were saying that a pre-etch is useful. And other people were saying that it is optional or not necessary.

IMO, pre-etch is essential if you want to completely liquify the toner. Cleaning the board is essential, period. And I use steel wool for that. The toner will have no trouble sticking to the copper if you just clean with steel wool or sandpaper.

The pre etch makes the surface porous, so that it sucks up the liquid toner instead of squishing it around. After cleaning the board, I dunk it in the etch tank for a few seconds and then rinse it off in a bucket of water. That's all a pre-etch is.

The predominant school of thought is to just barely/partially melt the toner. This is a very fine window with failure modes if too cold or too hot. If you pre etch, you can crank up the heat until the board bubbles and delaminates, and the transfer will still be fine.

This wasn't practical even 10 years ago.* The melted toner would suck into the paper and leave a mess of fibers on the board. Or the PnP would melt and distort. Today, this is the way to go, using Pulsar dextrin coated paper. Welcome to the 21st century.

I'm not saying I can make better boards than people who do not pre-etch or do not use Pulsar. I'm claiming that I can make a board once a year, any size, thickness, mass, down to 8 mil traces, and it will come out perfect on the first try, every time. No particular attention to detail or timing or muscle memory needed. It is foolproof. Just do the steps.

*unless you are one of those guys that stumbled upon the mythical legendary Photo Paper. The one that they changed the fomula of and/or discontinued by the time you posted about how great it is.

[Gandalf_Sr]

I have used pcbway in China, I've had great results and very low pricing.  Messing around with toner is not even worth the effort these days.

I use (a fully licensed version of) Diptrace for all my designs.

[cdev]

I use used glossy paper.

Its the cheapest paper imaginable, you know the ads they send you in the mail? the really thin sheets of paper totally printed with color ads. Its glossy and thin and free. It works great. It comes off well, with a little gentle rubbing.

*unless you are one of those guys that stumbled upon the mythical legendary Photo Paper. The one that they changed the fomula of and/or discontinued by the time you posted about how great it is.

[KL27x]

^I tried several photo papers. Some said it was the glossy that worked, some said it had to be matte. I dunno. I spent $60.00 on photo paper and gave up.

Prior to switching to PnP, and ultimately to Pulsar, glossy clay-coated magazine paper was the best transfer medium I have used. But for me it's worth the cost to buy the Pulsar.

You can use more heat than with regular printer paper, but if you get it too hot, the glossy magazine paper does eventually stick and turn your pcb into a chia pet. And it suffers greater dimensional distortion when it heats up and dries out. So on anything larger than a couple square inches, my thru holes and vias never lined up quite right. But it works really well, if you want to try toner transfer for free. It has some advantages over PnP, even. Just nothing on Pulsar.

When I buy Pulsar, it's not the cost of the paper I see. It's the amount of time and effort it will take to use up that paper. The donuts don't make themselves.

[Gandalf_Sr]

I used to do cheap inkjet photo paper used in laser printer.  Now I see the cost of all the materials plus the massive amount of time involved as a complete waste of money and time.  Learn to use Diptrace and just have the PCBs made.  My first quote for 25 boards from PCBway was so low that I had to check with them whether the price they quoted was for 1 or 25 PCBs, the answer was 25.

[AE7OO]

Hello,

I use the cold method of Toner Transfer and have had zero problems getting 10/10 (I can do 8/8 for single sided with a ground plane) boards made once I worked out the various ratios to use for my transfer liquid.  I've got three marked cups I use to mix it up, 1 for acetone, and 2 for alcohol, one for cheap 50% dollar store stuff and the other for the better 91% stuff.   It's mixed about 8 to 3 alcohol to acetone using the 91% stuff.
On the printer side I use a Brother HL-L2305 using stock toner printing onto FedEx/Kinko's Glossy paper using the darker print unless I've got to print at 1200 for the finer stuff.
I clean the board using a fine grain sanding block and then do a wipe down using acetone.
I pour the transfer liquid down, apply the diagram, smooth down and clamp between two old PCBs for about 5 min, release the pressure and allow to air dry.  If I'm making a double sided board I make sure I've got at least two ref holes.  I then plop the board into a tray with water and wait for separation.  I'll then clean the board using alcohol.
I then bake it  for about 30 min at 125C and let cool in the oven(This is a toaster oven I also use for SMD work).  This takes care of it being porous.  Then I etch using HCL/H2O2.  I don't use fuming HCL, but rather the muratic acid meant to change PH in pools.  It may take longer but it is quite mild.
Then clean with acetone and drill any holes needed.  I use leads from old resistors for my through hole, with flux I can get solder into holes.

I've done many boards for STM32 mcu including 32, 48 and 64 pin types with nary a problem.  The board I had the biggest problem getting right was for a Nordic Semi nrf24L01+ which really pushed the size down.
 
In less than a hour I've got prototype boards to confirm design params.  Maybe your good enough to create a perfect board every time, but I know I'm not.  More than once I've ended up doing 2 or 3 boards before I get it right.  Then if I need more than one or two I'll send it off knowing it will work when I get back.

GB