Advertisement

Blog

Can we get “printed” circuit boards which really are printed?

Printed-circuit boards (PC boards or PCBs) are literally the foundation of many of our electronic products. These copper-clad boards – often but not always using glass-epoxy FR-4 substrate, but increasing more exotic materials for the GHz+ spectrum – are both the structural support and interconnect medium we rely on the knot together active and passive components and their interconnects. While might be more accurate to call these printed wiring boards or PWBs, and there were attempts to change the standard designation about a decade ago, PWB never caught on.)

First, some historical background: The first PC boards were developed during WWII, using a ceramic substrate with conductive traces screened onto the surface; the leaded components were inserted and soldered into drilled holes in the ceramic. During the late 1970s and early 1980s, the PC board as we know it today came into production, again with through-hole components. Hard to believe but true: there were TV ads in the early days of PC boards from a major TV vendor (was it RCA? Or Motorola?) claiming the superiority of point-to-point, hand-wired “craftsmanship” in their TVs to the new-fangled PC board approach; but as transistors and ICs came into general use, hand wiring became a non-starter.

The use of PC boards with surface-mount technology (SMT) started in the 1990s, and lead pitch and track widths have been shrinking ever since, now down to a few mils (1000s of an inch). Double-sided and multilayer boards, some with over a dozen layers, added to the versatility. Ironically, many low-end consumer products, such as the PC mouse, often use low-cost phenolic as their substrate in single-sided layout with liberal use of jumpers to avoid the need for circuits on the bottom side; the boards are punched rather than drilled, which is quick and further lowers the cost.

Note that there is really nothing really “printed” about today's PC boards. Most are made using a subtractive process, where copper is chemically etched from unprotected areas; some use an additive process where copper is plated onto designated areas. There are also boards which are made using fine-dimension mechanical or laser-based milling (see the fascinating machines from LPKF, for example).

But what if we could really “print” or lay the conductive traces onto the substrate? Wouldn't that allow for fast set-up, board-to-board flexibility, and low-cost prototyping? That's why I was intrigued when I read in MAKE: about the Kickstarter efforts by Canada/China-based hardware startup Voltera to develop a low-cost printed circuit machine. It uses a silver nanoparticle conductive ink and places traces on a standard FR-4 board; there's even a way to place an insulating layer on top and then print another layer, for the electrical equivalent of a two-sided board. The machine is available for pre-order at around $2000 and will hopefully ship in 2016.

This printed-circuit board machine from Voltera on Kickstarter could be a convenient way to roll your own boards with flexibility and at low cost, although with some likely limitations.

This printed-circuit board machine from Voltera on Kickstarter could be a convenient way to roll your own boards with flexibility and at low cost, although with some likely limitations.

While this unit cannot produce 20-layer boards with super-tight pitch, I think even a basic non-chemical, do-it-yourself PC-board “printer” would be useful for verifying a sub-circuit's performance, evaluation of circuit designs, and more. Also keep in mind that not every design is a GHz-range RF product where tiny changes in component placement and track routing between the prototype units and final design will affect performance; there are many projects with lower-frequency operation or no RF functions, or where the RF role will filled by a module or IC which drops onto board.

I think it is great to have a unit which lets experimenters, hobbyists, and school programs make their own boards and get involved in circuit design. Too often, IMO, hands-on circuit engineering is frustrating due to the challenges of getting prototyping beyond simple “protoboards,” along with today's super-tiny discrete components and large IC with dozens of leads around and even under the package. As a result, much of the engineering experience is reduced to writing code for apps or programming assembled, purchased hardware (although companies such as adafruit do have some very nice kits and project material).

What's your view on DIY, non-chemical PC boards? Would one be useful to you, at a reasonable price, even if it could not do the latest PC-board track spacing, IC lead pitches, or complex multilayer fabrications?

Related

What PCB material do I need to use for RF?

What else can you do with our ubiquitous PC-board material?

9 comments on “Can we get “printed” circuit boards which really are printed?

  1. dick_freebird
    October 20, 2015

    Subtractive copper plated boards (if you only need 2-sided)

    are so commoditized (ditto etchant, the EPA aside) that an

    additive process would have an impossible time competing.

    Once you get to internal layers, things are no longer so

    clear – plenty of fine work inside the sandwich that must now

    be built up. Probably quasi-additive in some sense, but still

    the PCB houses attempt to squeeze out cost by panelizing

    and batch processing at steps where they can.

     

    Silver ink? That's adding cost. And taking you off the map

    as far as reliability design rules (let alone UL recognition). 

    Maybe fine for playing around, but what happens when you

    decide you want to run serious current + temperature on

    this print-head-friendliness-first composition trace?

  2. David Ashton
    October 20, 2015

    There would be certain advantages to “printing” circuit boards, depending on the process.  You could print certain tracks not only wider but thicker – and if you could make them thicker you needn't have them as wide for the same current-carrying capability.  And if the printing was robust enough you could get over the problem of tracks separating from the board due to mechanical stress (eg power connectors).  But depositing molten copper on fibreglass is probably not the way to go – you'd have to come up with some other material.  Not to knock the voltera machine, but it probably wouldn't cut it for high-current stuff.

  3. Bill_Jaffa
    October 21, 2015

    I think the real benefit here would for prototyping, to test out a circuit concept or get a quick board so you could write and debug the software. That's where prototypes wihc are not “geometrically” equivalent to the final board cause problems. So if this printed technique used real PC board substrate such as FR-4, and the tracks had the same geometry, that would be a big help in the product development/debug phase. I agree that it might have serious shortcoming for full production, but that's later in the cycle.

  4. jimfordbroadcom
    October 21, 2015

    This is definitely a niche product.  I previously worked at a company that had a board milling machine, and in about 2.5 years, I used it once.  The advantage was speed; we had a test board in about an hour, vs. at least 1 day for etched wiring boards (EWB's, as the military used to call them – I've been out of that circle for >20 years so I don't know what they call them today) if you want to pay big $$.  If you can wait a couple of weeks, OSH Park has boards for extremely low prices; I paid less than $90 for 3 boards this summer!  Other shops charge more of course, and finer lines and spaces and quicker delivery drives up the cost.  We find that the minimum order for say 10 boards is about $1500-2000, and for the finest pitch boards, we can spend upwards of $20,000 for a batch.  I think if you have to have your simple boards very quickly, or as the LKPF people like to point out, you're the FBI or CIA and absolutely have to have your boards built in-house to keep your secrets secret, a circuit board printer is the way to go.  Otherwise, it doesn't make sense not to use an outside fab shop.

  5. jimfordbroadcom
    October 21, 2015

    That's LPKF, not LKPF; sorry for the dislexia.

    I forgot to mention about the superiority of PCB's vs. hand-wired circuitry.  The way I understand it is that hand-wired point-to-point wiring is better for vacuum tube circuits.  It makes sense when you realize that tubes (valves, to the Brits out there) are high-voltage, low-current devices for the most part; they are characteristically high-impedance devices.  Transistors, OTOH, are typically high-current, low-voltage devices – low impedance.  So what happens when you try to put a tube on a PCB with groundplanes like we typically use with transistors?  Well, all that capacitance makes the tubes very unhappy (as the late, great Bob Pease would say [RIP, RAP]) and they oscillate and roll off the high frequencies and other assorted nasties (probably another Pease-ism).  Transistors take it in stride, as long as you don't get crazy with the shunt capacitance.  Now, trying to make transistor circuits with (relatively) large inductances to ground and between stages from the long, isolated wires doesn't work well, either.  Can't use sockets for anything but the slowest transistor circuits or IC's.  Not a problem for tubes, though.

    This is not to say that tubes can't be used on PCB's with their high-volume, low-cost, low labor intensity tendencies.  I think it's just a different philosophy of PCB design; keep one point for ground instead of a plane and route as much as possible on one layer, don't cross over signals if possible and make them cross at right angles if you have to make them cross (hey, we do that on transistor boards anyway), etc.  Anybody know where I can get surface-mount tube sockets?

  6. comfusn
    October 21, 2015

    It was Zenith that boasted of hand-wired craftmanship.

  7. Bill_Jaffa
    October 21, 2015

    All very good points. Tube-based design is increasingly a lost art (and it is “art”–all designs and layouts are) and their support compoments are also harder to find.

    Yet the audio world does do a surprising amount of tube-amplifier new designs, and tubes are still used for high-power transmitters and even TWTs in communication satellites. Of course, those transmitters do not use PC boards as we usually think of them at those currents and voltages.

  8. bobdvb
    October 23, 2015

    I think one of the trade offs is not just scale but remember that an additive method could have the following advantages:

    • Energy usage: subtractive production may be efficient but only at bulk, it is inherrently an energy intensive process to keep the chemical baths hot and to do the lithography.
    • Chemicals: The subtractive process uses horrible chemicals that create safety and enviornmental risks. I think some companies would prefer a cleaner method even if it was a bit more expensive.
    • Waste: Substractive designs might be more wasteful, or at least require expensive reprocessing to recover the copper lost.

    Sometimes the simple up-front cost isn't as big as the wider picture.

  9. eafpres
    October 27, 2015

    Hi Bill–first, thanks for a great article.  Here are a few thoughts, we did a LOT of prototyping in our labs when I was in an antenna company:

    1) A lot of our stuff was 2 sided, so it is “easy” to make a proto by removing copper with a purpose-built milling machine.  We had one that could take Gerber files and then machine the board.

    2) LPKF and others will sell you an entire benchtop line if you want–the ability to SMT etc.  Pretty spendy for most labs.

    3) Over the years there has been interest in certain applications, such as analog front ends, made using LTCC.  In some instances, you want to embed components inside the layers.  You cannot do that with rigid boards regardless of how many layers, unless you were super crafty and milled out a hole to make room for a component on another layer.  LTCC makes this a bit easier.  Here is an example from MuRata:  MuRata LTCC

    So, the additive machine means you can prototype such a design!  AFAIK that isn't really feasible with any other benchtop process.

     

Leave a Reply