This is a lovely piece of work; I think it's missing the point to speculate on possible profit margins, or the magnitude of a kickstarter project; the thing is the point here. Something so well turned out makes me want to make things, not buy them.
Your comment precisely describes the "Hacker Spirit" that I expect to see around HN. These are the kinds of projects that I gladly upvote.
Hacking (often just for the fun of it) is so much more rewarding and in the end provides so much more value than the usual techcrunch-style drama that often occupies most of HN home page.
A killing? As a long-time head-fi member I disagree. There have been plenty of cool or highly artistic as far build and casework amps and they tend to be very limited hand-built affairs (usually less than 100 units produced). At any scale, if you're selling in the hundreds to thousands, head-fiers will require that a $100 amp be at least sonically competitive with other $100 amps, $300 amps with $300 amps, etc. They might pay a 25% upcharge if something about its build or casing is unusually unique, but generally for portable amps they pay more for portability/slimness, not because it simply looks cool. Usually in these lower ranges you're competing with Chinese built stuff that is phenomenally good for the price. You can get pretty high-quality portable amp/DAC combos around the $100-200 price point from companies like iBasso/FiiO.
That this is a fully point-to-point design encased in resin is cool, but it's not $300 cool unless he's building it with top-notch components, which I don't see here.
I must diagree. This is no ordinary amplifier, and I don't think you can simply compare it to an amplifier with a "unique case".
This is what I would call, if you'll forgive me, transparent design. Obviously I'm not talking about the just the case. The mechanism of action of modern electronics is hidden behind 6-layer printed circuit boards and integrated circuit packages. Opening up an iPhone gives precious little insight into its workings. You wind up googling datasheets instead of following traces.
This is a tactile product. A 6-year old could pick up this amp, ask a few pointed questions about what they see inside, and actually have a prayer of figuring out how it works. I know I would sit there for cumulative hours just turning it over and watching the image of the components refracted through the faces. With everything stripped away, the noise becomes the signal – the imperfections in the solder joints, slight tool marks on the leads, the banded resistors that have gone out of style.
Yeah, there are people that would pay quite a bit for this, even if they're not audiophiles.
Where did I say there aren't people who would pay for this? I said he couldn't make a killing. Depending on the price-point... if he offered it at $500, there's probably a market of 25 people. $300, probably 50. $150, probably a couple hundred. I've owned about 20 headphone amps over the years that cost between $50 - $2500. Some of these amps have serial numbers in the double digits, many hand-built PTP tube amps, and the designers tend to be pretty transparent about the numbers sold, as many are headphone enthusiasts themselves and they come from the DIY-world then transition into small businesses. Very few of the small shops make anywhere close to a killing - most are just barely making a living doing what they love. There have been tons of Kickstarter-esque pre-orders threads on head-fi, where a well-known designer has a goal of selling something like 100-500 amps before beginning production and offers early-bird discounts. Usually the product in question changes the game in some way whether it comes to sonics/power/battery life/portability (if we're talking about portables).
It's a poor assumption that novelty-factor is high on the list of an audiophile. As cool as this design is, what it does does not add any functionality or utility to the amp. At the extreme upper end sure novelty may become more of a factor. But even when you get to kilobuck amps the component and labor costs are still quite high, ie. the designer of a $5000 amp usually has $2500 sunk into the amp, and he might sell a dozen of them. And to do that, the company has to have a ton of cachet in the industry.
There is a very large and active community around headphone audio if you're interested in this field... in excess of a quarter million members with about 10M posts - http://www.head-fi.org/. I've been an active member there for over 7 years and into this hobby for over a decade. I've hosted a meet with about 30 members w/ almost $100k in equipment. I personally have owned over $20k in equipment over the years. I think I can speak with some authority on this.
"As cool as this design is, what it does does not add any functionality or utility to the amp."
Well, it looks like you could drop the amp off a bridge, run over it with some trucks, and attack it with a blowtorch, and still have it work fine afterwards, as long as you didn't hit the jacks directly with the blowtorch. That may not be functionality you're interested in, but I think it probably counts as added functionality!
It might LOOK that way, but I suspect it's more likely the opposite is true. If this device landed on a hard surface it would be a very high deceleration, as there is no way for it to deform and absorb the energy. A device of more traditional construction can deform and actually stress the components much less. If the outer case could deform even a few mm on impact (which most consumer plastics would instantly rebound from), that could mean the difference between a 100g deceleration and a 2000g deceleration.
This kind of encapsulation, filling the empty spaces in the device with a thermoset resin, is called "potting", and it's been used for decades to improve the shock, vibration, and impact resistance of electronic circuits. It's true that the actual deceleration experienced by the components is much greater, and potting is not without its problems (a friend of mine told me about a problem he tracked down at his company in the 1970s, where potting in too-hard epoxy caused many of their circuits to fail in the field due to a sort of thermal shock), but it does generally work.
The issue is that acceleration in itself doesn't damage chips, capacitors, resistors, etc. What happens is that the things that hold those components in place — their leads — experience large forces from trying to hold those components in their relative position. Those forces are generally greatly reduced by potting.
There are probably exceptions. It wouldn't be surprising if MEMS gyroscopes and accelerometers were more subject to damage after potting.
Buy a mint-tin cmoy amp and you'll see the same thing. The circuitry is very simple. Wires from the input connect to an op-amp chip, which is connected to the output. Some resistors feed the output back into the input, controlling the gain. Finally, a big beefy capacitor sits between the power supply and op-amp to provide some extra current for sudden loud parts of the music. That's it. Everything cool happens inside that little IC.
Agreed - if I wanted something that'd make kids interested in what's going on, hiding all the "magic" in an opamp chip isn't how I'd do it. Something super simple, where I could explain exactly what each component did, and more-or-less how they do it (with the obvious handwaving requirement for explaining transistors to just about anybody).
Could be built for approximately zero dollars by just about anyone who's the sort of person who's got transistors "lying around" at home. I can easily imagine explaining how that works to the satisfaction of my "test" 9 year old. Not so much with the opamp version...
Oh come on, the only thing transparent about this design is the resin. It's an integrated circuit connected to a few passive components. The only way you'll have a chance of understanding it is by reading the datasheet.
You aren't really addressing his main point about market size other than saying there are people that would pay quite a bit for this, even if they're not audiophiles. I have no idea what this does, or why I would want it, and I'm fairly confident I'm not alone there.
How many people want or even know what this thing does?
How many of those people will think this product is so much better than what they have that a clear version is worth spending $100s of dollars on?
It could be that the market is huge and I just do not understand it, but you didn't really help us understand how big it is either. To me, "a killing" seems like a stretch.
There is a forum/website out there call "GearSlutz"[1], that has a huge & healthy community based around, well, gear.
The amazing thing I've learned from lerking there from time-to-time is that some folks are actually seeding, promoting and selling their "cottage-made" products there.
The thing is, these gear "slutz" are hungry for well-made/hand-made, "analogy" products that simply don't cost a fortune -- think Value Compressors & Amps etc. -- I think this guy could do well to go and check it out...
An example of another product that was "launched" on this site (serial #001 was sold on that very forum), is a beautiful little "8bit sound box" called Biscuit, hand-made by some guy in France:
Unless he can find cheap workers, the labor to produce it would make $100 each a steal. No way does it take less than an hour start to finish to build that.
I'd start pricing at around $350 and see how that flies.
If you invested in some jigs to hold the components while soldering, then I could see significant time savings. The time to cure the resin really doesnt count. I woukd estimate that with the correct jigs, you could complete the soldering in half an hour tops. I count roughly fifty solders on thus thing, most of them pretty simple.
You also overestimate the per hour cost of semi skilled labor. The skills involved arent that involved or difficult to learn (especially with the right tooling).
Of course you need fixturing: that's a given. Remember to account for the time and materials making up fixtures.
I would price the labor at around $15-$30 per hour minimum. You aren't going to farm this out to an assembly shop (not enough volume) and there isn't enough money in it to hire employees and deal with that overhead, so you're down to finding a FOAF or a student who can solder and is reliable. Factor into that that the person must be supervised.
Next, this isn't just assembly that has to work; it has to look good, so the assembler needs the skill and attention to detail to provide that. That means either you get lucky or you have a lot of rework, or you watch him closely. Soldering is not enough: need to pay attention to the cut ends of those thick conductors. The component lead terminations must look good and the components themselves should be fairly regularly placed and not look sloppy. Remember, you're selling primarily on looks. Also, if this assembler isn't always available, you have to make enough inventory to handle orders that may come in until he can make more if you don't have the time to do it yourself.
Volumes are uncertain, but the parts are cheap, so let's assume he buys enough to make 100 units at a time. The selling price has to reflect the risk that all those parts won't be used. Buy higher inventory levels and your parts cost goes down, but risk of "dead stock" goes up.
Having a small electronics mfg business on the side, I live this on a daily basis. Luckily I don't sell to price sensitive hobbyists :-)
But in the end, the price I'd try to sell it for has nothing to do with how much it cost to build: it's cool looking. That's where the price comes from.
Learning to solder mostly requires practice, so you could probably train someone to do it. You just have to make sure they don't destroy a lot of expensive parts learning. But you're right about all the other stuff.
I wouldn't be surprised to see something like this on Kickstarter, though.
You underestimate the power of well designed production, and how good people can get at a repetitive task they do all day.
I count about 20 components, about 40 solders. Allowing 5 seconds per bend (using jigs), 5 seconds per solder, one minute for assembly (again using jigs), one minute for finish... I'd say about 7 minutes labour per unit. Jigs, etc, could run to 100 hours all up, they seem very simple.
I'd feel comfortable budgeting 15 mins/unit for runs of 1000+.
I've performed as well as supervised manual circuit board assembly. I know what it's like to get in the zone of doing something like this all day.
My experience teaches me that it takes on average soldering a few dozen units of anything unusual before the assembler gets the hang of it. Unusual being anything besides stuffing a PC board and manually soldering. This is not likely to ever be a high volume product, and I probably (admittedly only knowing the size of the market from what other posters have said) would not be comfortable doing production runs of more than 50 at a time, so you lose a lot of the benefits of doing long runs.
That said, I think your estimates are a bit optimistic. Once the entire time is accounted for: from kitting up the run, setting up the work area, actual assembly and testing, my guess would be closer to 30 minutes per unit.
Slightly OT, but if you can point me to any good resources for setup and manual assembly of cable harnesses, it would be much appreciated. It's an area I find myself getting involved in and I don't have much experience, so it's pretty much "learning by error."
Until MegaCorp comes along and slaps him with a patent infringement lawsuit, of course. I'm pretty sure there are a large number of patents around amplifiers of various kinds.
Most, if not all of those patents are expired (20+ years old). Once a patent expires, it's in the public domain and can't be re-issued or re-assigned. So yeah, overly cynical in this case. :)
This is actually a pretty simple op-amp circuit. The tricky amplifier stuff is all in the OPA2107 chip; everything else is either part of the power supply, or a straightfoward arrangement of resistors and capacitors to get the op-amp to do the right thing.
The really cool part here is the assembly. That thing is very pretty.
This worked out beautifully. I hope we see a lot of new electronics projects built like this in the next few months.
His choice to pot the electronics in transparent (apparently polyester) resin reminded me of this unfinished design provocation of mine, "The Egg of the Phoenix: a computational time capsule": http://www.canonical.org/~kragen/eotf/
Specifically, I was thinking that potting solar-powered electronics in transparent resin would be a good way to ruggedize them, so that your electronics might have a chance to keep working for decades or centuries.
Does anyone know what the lifespan of the resin is? Is it longer than 6 months? For example, see the link here that states this resin has a 6 month lifespan. Wondering if you can get them for longer lifespans measuring in years.
If you're lucky it'll have some amazing cable lacework[1] holding the various bundles together. Sometimes I've thought that the insides are actually better looking than the case. It's a shame that it's dying out in favour of cable-ties and velcro, although the effort involved in both learning the skill and creating/reworking bundles means it's seldom the most economical approach.
I think it still lives on in some Ham radio groups teaching materials and standards though, as well as niche environments or reliability critical applications like (aero-)space
Agreed, but I think the look rather outweighs maintenance. It’s such a simple device that you’d hardly care to maintain it. You fry a capacitor, it’s one of a couple dozen parts; the thing probably wasn’t terribly expensive to begin with, so you just frown and replace the whole shebang.
Even fewer would have the equipment to fix it and the determination to source the part. A fraction of those would have the skill to pull off a successful replacement.
Anyone know about the thermal conductivity of the cured resin? If it's a thermal insulator I wonder if the capacitors might burn out after prolonged use because there's not enough heat dissipation.
Very nice, but seeing that made me wonder whether this can be improved as follows: instead of soldering, drill holes in the copper wire that are slightly smaller than the thickness of the component wires you want to attach. Then, heat or cool both enough for the component wires to fit through the hole (thermal expansion factors will hopefully be different enough for this to work), put them in and let the device go back to room temperature.
the speaker in a headphone cup works as follows: current flowing through a coil of wire that is fastened to the speaker cone generates a magnetic field, which interacts with the magnetic field of a fixed magnet, moving the cone and so generating sound.
the cone movement, and hence volume, is proportional to the strength of the magnetic field, which is, in turn, proportional to the current through the wire and the number of turns in the coil.
there is a tension here, which is not obvious. the problem is that the coil is fastened to the cone, and so moves with it. for a responsive cone, we want as light a coil as possible (a heavy coil will give the speaker cone inertia, making it less faithful to the music signal). but if we make the coil lighter by reducing the number of turns we reduce the volume; if we make it lighter by making the wire thinner we increase the resistance, lower the current, and again reduce the volume.
one way to work around this is to supply a higher voltage to the coil. that can counteract the higher resistance of a thinner wire, and so provide the same volume with, hopefully, better sound quality. but the output voltage of many sources (particularly portable players) is limited.
so, finally, all should be clear: people use amplifiers like the one here to increase the available voltage so that they can use higher impedance, and arguably better sounding, headphones.
If you're using headphones with a high impedance, the power output of device like an iPod or whatever may not be high enough to deliver enough volume for your liking.
Although there's better ways to express it, you might be right. I wouldn't want this thing next to my RF power amps, cellphone devices or power tools. Turning deaf at a turn of a nearby equipment it's probably not nice. Nonetheless it looks pretty, but not as pretty as the glow of my now defunct old tube radios.
Yet the maker says exactly the reverse in his blog...
"Although there is no metal shielding as you would have in a conventional chassis/PCB the amplifier exhibits no unwanted noise or RF interference as you may associate with an open chassis design such as this it is dead silent even though it is next to my mobile phone and WiFi router."
Just because it's next to the device doesn't mean it'll induce current in that position or at the right frequency to generate audible noise. Testing is much harder than that. I bet someone will build a clone and find out the hard way.
The function of this amplifier isn't to achieve the best amplification at the lowest cost -- it's to be a curiosity piece that actually does something. It seems to perform this function admirably.
