OT: thoughts and other brain-droppings on low volume mfg

[OB Dave]

I didn't want to hijack the Mojo 61 thread but I did want to answer this question because it's one I get from time to time and I'd rather that discussion happen here than on the Mojo thread.

 

On a side topic: there's been some love for the Roland FA-07 lately.

Would it ever be in the realm of possibility that you do a single production run of the OB-1 in black?

Probably not.

 

At the time the DB-1 went out of production, the retail price was $349, which frankly is way too much money for a box with two buttons and nine sliders. God bless every single one of you who bought one, though. And even that that price, the combination of margin and volume wasn't enough to justify going forward. I had pretty much saturated the market at the point point I was able to hit. When it came time to start planning for the next production run, costs had crept up further and I really couldn't see raising the price on a product that already cost too much.

 

And the pushback I got on price was not unreasonable. I get it. Take a look at the Korg Nanokontrol. List price is 84 bucks, streets for $60.

 

http://c1.zzounds.com/media/fit,2018by3200/quality,85/NanoKONTROL2_black_slant_634305189002800000-dd87f14385cb124016e66d18499041e5.jpg

 

It's probably an A-mark item, so Korg has to be able to produce these things for about 15 or 20 bucks. That's assembled, tested, in the box, shrink-wrapped with the user manual and ready to ship. How the hell do they do it? I was paying $1.66 for my faders, in 10,000 piece quantities. Now, granted I picked a quality fader because I'm picky about how I wanted the thing to feel, but still, the cost of the fully complete Korg assembly is roughly the cost of nine raw parts on mine. No knobs, no enclosure, no circuitboard, no profit (!), just the linear potentiometers. Damn.

 

It's the sort of thing you can only achieve by manufacturing at very high volumes, in China or offshore. And you pretty much have to have boots on the ground, because any undetected error in the assembly process will cost you whatever profit you would have made for the whole production run. It's why Uli Behringer moved to Shenzhen. There was a telecom company who shall remain nameless who was manufacturing cellular phone handsets in San Diego. The assembly process is largely automated, with pick and place machines that place all the surface mount devices on the circuit board. These parts all come on big reels. They sorta look like Super 8 film. Well somebody had mistakenly loaded the wrong reel into one of these machines, so the machine was placing one wrong part on the circuit board. The machine only ran for about two hours before the mistake was caught, but now those boards either had to be reworked by hand, or thrown away. This one 2 hour error cost the division its entire profit for the quarter. For the quarter. Let that sink in a moment.

 

So the high volume manufacturing thing is not for the weak of heart. Now granted, there's more margin in musical instruments than telephone handsets, but still....

 

So that's the bad news. For a niche product like the DB-1, the high-volume build-it-overseas thing is just way too risky.

 

The good news is that the barrier to entry for low volume manufacturing is fairly low. I'd have to go through Quickbooks but I think the startup costs for the DB-1 were less than $20,000. About $4500 of that was the regulatory testing required for electronic devices, and a few thousand more to get the knob made. Everything else was material cost. Having no employees and working out of the house certainly helped. You can buy the parts in any quantity you want, the trick is figuring out what quantity to buy to hit a decent price point. It's a balancing act managing the cash on hand. And electronic assembly cost is in part driven by volume because there's setup time there (somebody has to set up the pick and place machine, load the reels of parts in, and so on). So you want to do the runs as big as you can afford to. The DB-1 had a bunch of parts that had to be soldered in by hand, and that drove assembly costs up too.

 

So the trick is to come up with something with OK profit margins to offset the low volumes. In the case of the DB-1, the only reason it penciled out at all was, at time time it came out, the Nord Electro was dominant. There was no Sk1/Sk2, no Numa, no Crumar, no CX-3, no PC3, no Kronos, no Fantom, etc. Those hadn't happened yet, and if you wanted tonewheel organ and some other voices in a slab keyboard, the Electro was the only game in town. It was an expensive instrument, so selling an expensive accessory for an expensive instrument seemed doable, because the accessory in question was the only one of its kind and it solved a problem much more elegantly than any of the workarounds. Fast forward to today, and there are lots of clonewheel options available, of which the FA-07 is one of many, and in my opinion, not even the best. Not to throw shade on it, I just don't see it holding the commanding position that the Electro once did. If you really want drawbars, there are many excellent clonewheels that have 'em. That wasn't the case 7 or 8 years ago.

 

 

[AnotherScott]

Thanks for that interesting insight from behind the scenes.

 

I could envision one way this could work, though... a kickstarter campaign. i.e. maybe it would make sense to do a run if you'd have commitments from, x00 people who were willing to buy one, that many pre-sold even before you manufacture.

[ElmerJFudd]

There's definitely a type of customer that prefers boutique quality built unique feature specific items over mass produced plastic stuff and don't mind the made in the US (or assembled) mark up (within reason). But yes, not being able to assess interest accurately and calculating cost and sale price is very tricky. As Scott suggested Kickstarter might be the way to go for any project you imagine doing in the future. MI being very specific market and keyboard players being just a % of that.

 

The thing with the Mojo61 or Sk, or Electro is that there are players that won't consider gigging a clonewheel or even these models that also offer a few ancillary sounds because they're lacking as jack-of-all trades boards... especially in the programmable synth area, but also in the other stuff one covers in a night like brass, strings and other orchestral stuff, even percussion, arps, loops, etc. Dragging a Hammond or even a clonewheel around is a luxury for any gigging player that isn't doing a straight ahead jazz, blues, gospel gig. Personally I won't carry more than two boards and the lower has to be a weighted action stage piano. The FA is a decent choice for upper tier but it lacks faders and if adding faders to it, drawbars would be preferable. A lot of people choose the Kronos or Montage, PC3K etc. with their organ engines and faders because they need a board that does a lot more than the typical clonewheel. That said, still hoping GSi will do a Mojo73 with the full Gemini dual DSP and synth engine and include more controls. Or we can dream about the older synth engine from the Nord Stage 2 trickling down to an Electro 6.

 

 

[GRollins]

How the hell do they do it? I was paying $1.66 for my faders, in 10,000 piece quantities. Now, granted I picked a quality fader because I'm picky about how I wanted the thing to feel, but still...

 

I'm able to talk the lingo to electrons--get 'em to sit, beg, roll over, etc. When I got to the point where I had a couple of keyboards and wanted some control over what was going on, I started sketching out a few ideas about how I wanted a keyboard mixer to work. Let's see...use a discrete JFET differential biased with a CCS to create a balanced signal, don't really much care about having scads of gain, since I'm not planning on using mics through the thing, so with the gain comparatively low, I can get a good S/N ratio pretty easily...started working the whole thing out in my head, right? Had a few bells and whistles, but other than the fact that I was planning on doing a discrete circuit vs. opamps, it wasn't really going to be all that weird.

 

Then I started pricing the pots. And knobs. Don't forget the damned knobs. And jacks, both XLR and 1/4". And...oh, crikey...the pots, knobs, and jacks alone were going to cost me more than I could buy one for, and that's not even charging myself for the design or doing the layout and etching a PCB.

 

In the end I went higher than I needed to--bought an Allen & Heath ZED-14--but had I set myself to building something along those lines (with something like 100 pots and sliders), I'd still have paid more for the parts than the unit cost me, and that's if I'd gone cheap on the pots. Coming from a background in high end audio, I'd prefer to use some sort decent pot, like an Alps instead of an Alpha, but that would only tilt the scales further. Now, granted, my design would presumably last longer and perform better, but...

 

I'm picking my electronics projects much more carefully these days.

 

Grey

[OB Dave]

Hahahaha, yep. It's the mechanical parts that do you in. The DB-1 started out as a software project, so naturally I started with the microcontroller. Lots and lots of options here. I picked a decently fast and flexible one. They're two bucks apiece. The passive components are essentially free. I think the resistors run about seven bucks per thousand. No problem! And then there's all the electromechanical parts. The silly MIDI jacks are 50 cents a pop. There's four of them. Just little blobs of metal and plastic and together they cost as much as this amazing little microprocessor. I think the DIP switches were about a buck. It starts to add up pretty fast.

 

The single most expensive line item was the sheet metal enclosure. The metal company did a beautiful job with those, but they weren't cheap.

 

The kickstarter model is something I've considered, but I haven't thought of anything with a wide enough market appeal, yet not so wide that Yamaha would come along and squash me. And also, in the case of the DB-1 it really started with: I want this thing to exist, how come nobody makes one? I never set out to manufacture them, I just wanted one for myself and it just sorta snowballed once word got out.

[GRollins]

If you have a market for, say, 10,000 of them (and know that you'll sell that many), it's easy. You can go for quantity breaks--really plunge in and commit. But when you're only looking at, say, 100...and you're not necessarily all that sure of that number, it's a lot tougher.

 

Two scenarios: You're ready to go onstage. On the other side of the door there are 20,000 hysterical fans stomping, screaming your name...or...there are 2 pathetic drunks and a bored bartender, none of whom care a whit about you or your stinking band. Sometimes it's easier to go big.

 

Yamaha, Korg, and Roland have 20,000 screaming fans. Even if they turn in a "bad performance," the fans are loyal enough that they'll be back for the next gig. Joey and the Fence Post Humpers are going to have a much harder time of it--particularly if they flub the night. No mercy. They don't have that reservoir of good will to draw on.

 

And if you've got a niche product that will only sell to a limited audience in the first place, you're never going to get to the 20,000 adoring fans stage. You'd better like the small venue experience; it's all you're ever going to have. $50-100 a night and haul your own gear. If your fingers got slammed against the door frame bringing in the PA, tough, the show must go on.

 

Grey

[Threadslayer]

Having no employees and working out of the house certainly helped.

 

This, definitely. Imagine your unit cost if you had to factor in the overhead expenses associated with having employees*.

Good for you for persevering and delivering a quality product even for a while. It ain't easy.

[ElmerJFudd]

This is very reminiscent of the boutique guitar pedal business. Thankfully in that arena there is still a scene, interest in not using big name manufacturing to get your sound. But yes, there's always fear of a Behringer or BOSS swooping in to copy desirable components, user interface elements, and aesthetics. These guys managed to build a following doing boutique guitar specific trade shows. But even that is fading to social media, company splash page to gauge interest and kickstarter which perhaps does it all and asks for a commitment at certain price points including early in.

[Markyboard]

It's quite a revelation going from making something for yourself to providing it and to a small base of potential customers. An easy way to see this is building simple cables with TRS or XLR connectors. Many of us do this for personal use, and like you choose to use good connectors and cable. Now think about selling those hand built cables; not just the part cost but factor in your time and shipping cost. It quickly becomes evident that the 10-20 ft cable (which btw all cost pretty much the same to make by hand) has to sell for $20-$25. Even at this ridiculous cost the small potential profit just isn't worth it.

 

Great post Dave- thanks! Btw no way is this OT imo.

[stoken6]

It quickly becomes evident that the 10-20 ft cable (which btw all cost pretty much the same to make by hand) has to sell for $20-$25. Even at this ridiculous cost the small potential profit just isn't worth it.

 

It comes down to how much the niche is willing to pay.

 

Dave's OB1 is more expensive than the Korg Nanocontrol, but has a couple of unique advantages: primarily, real knobs and 5-pin MIDI. Similarly, I paid about $20 per cable for a few custom-made leads - exactly the right length, right-angled connectors of the right colour just where I want them narrower-gauge cable making the loom easier to handle etc.

 

I was willing to pay extra for my niche needs.

 

Cheers, Mike.

[Joe Muscara]

I'm glad some people here realize the realities of this. I seem to remember threads on here where people were complaining about the costs of items like, "I could do it cheaper!" It might seem so until you really get into it.

 

I bet the same people are the ones who way underestimate how long a project will take.

 

Great thread, Dave!

[Rusty Mike]

I bet the same people are the ones who way underestimate how long a project will take.

 

At least twice as long as my wife thinks it will.

[GRollins]

I bet the same people are the ones who way underestimate how long a project will take.

 

At least twice as long as my wife thinks it will.

 

Fooey, my projects always take at least twice as long as I think they will.

 

I always tell my wife two times. The time I estimate...and the time I'm betting that it will actually take, which is always a minimum of twice the first estimate. It's a standing joke between us.

 

The sad reality is that sometimes things take twice (or more) the time of even my 2x estimate.

 

Grey

[Joe Muscara]

I've lived by this for a long, long time.

 

[video:youtube]

[GRollins]

So that's what I'm doing wrong...not multiplying by four to begin with...

 

Grey

[AnotherScott]

[Synthoid]

I've lived by this for a long, long time.

 

[video:youtube]

 

I was just thinking of that!

 

 

[pjd]

Hi Dave --

 

That was a terrific write-up. Manufacturing ain't for the faint of heart.

 

One of my old buds started a very successful software company, then decided to extend the company and product into hardware. It damned well near sunk his company.

 

-- pj

 

Music technology blog: http://sandsoftwaresound.net/

 

[GovernorSilver]

Sometime in 2016, I first heard of the SOMA Lyra 8, a synth designed for dark ambient/noise/experimental stuff by some Russian guy. He planned to make this synth by hand, for each order, so we were told to expect a long wait. I got on the waitlist anyway around Sept. 2016.

 

Then the Russian guy entered an agreement with a Polish factory. All of a sudden, customers got their SOMA synths far earlier than expected, thanks to the Polish factory kicking in production. Thus, my Lyra 8 arrived last month instead of sometime in 2018 as originally expected.

 

Final price wasn't that much higher than originally projected 530 Euro price. I think though the costs were kept down by all the manufacturing being done in Eastern Europe. My new synth is built like a tank.

[Jazzmammal]

All of this is simply proof that people never really cared for quality it's just that 50-70 years ago all the factory equipment necessary to produce cheap stuff hadn't been invented yet. That's why a 50's era toaster or hair dryer was all chrome, built like a tank and are still in use today in some cases. Think the original B3. That was how they they had to build them based on the materials available at the time. The individual components were big, heavy and expensive requiring a large, well built cabinet or housing. Since B3's and toasters had to be built that way, they were very expensive so in order to entice people to pay that much the overall design and build quality had to be there. They weren't really marketed to the lower end masses, they were marketed to the upper middle class.

 

We all talk about how great the quality was then and how they "don't build like that anymore" but in truth if the general public had a choice then it probably would have been the same as now. A cheap article that is "good enough" for 80-90% of the population wins. Someone else comes up with a way to do it cheaper while maintaining that 80-90% they win and so on. The term for that is dumbing down because every iteration of that process loses another 10% of the original functionality and quality. Our favorite musical items are getting dumbed down to the point of uselessness in many cases. We know it but the younger generation doesn't so they don't care.

 

Now, it's not worth it to try to fix something that cost under $300 or so. It's cheaper to give it to the recyclers and buy another one because we all know it's life expectancy is only a few years anyway.

 

Bob

[Markay]

Not my field but wasn't 3D printing designed to facilitate affordable low volume manufacturing?

[WesG]

As long as you want to manufacture simple, low-tolerance hunks of plastic..it's a long way from making electronics in metal enclosures. Probably 100 years away.

[Markay]

Well according to this site they supply 3D printers that use a nickel alloy suited for the following:

 

"Applications include:

 

High termperature applications such as :

- Components for liquid fueled rockets,

- Rings, casings and formed sheet metal parts for aircraft and land-based gas turbine engines,

- Cryogenic tankage,

- Fasteners and instrumentation parts."

 

Overkill for the MI industry but surely lower cost suitable materials can be used.

 

And this manufacturer has a 3D machine they claim makes "100% Additive Printed Circuit Boards".

 

 

[AnotherScott]

We all talk about how great the quality was then and how they "don't build like that anymore" but in truth if the general public had a choice then it probably would have been the same as now. A cheap article that is "good enough" for 80-90% of the population wins.

Yes, we tend to choose, not just price, but also convenience over quality. That's why we so often listen to music via MP3, why the original DVD took off over laserdisc, why people are willing to not just supplement but fully replace landlines with inferior sounding cell phones.

[Synthoid]

people are willing to not just supplement but fully replace landlines with inferior sounding cell phones.

 

Hey, my phone sounds great!

 

 

[WesG]

With all due respect, none of your examples are even close to printing a finished product.

 

3D printing a blank PCB is of virtually no interest. There is almost no cost in small-run PCBs. I've been doing one-offs at home since the 80s. A mini CNC machine would do an awesome job.

 

Show me something that can print a PCB with passives and silicon, and we'll talking about actually manufacturing something.

[GRollins]

It's been my experience that 3-D printer companies make grandiose claims that have little or no basis in reality. By the time you get a 3-D printer that can actually do anything like what they claim for the lesser models, you've spent so much money that it's more cost effective to do it the traditional ways. And let's not even begin the discussion about reliability and getting repeatable results. The real world waste factor is horrendous, once you factor in throwing away the ruined work pieces that look like some sort of bizarre alien artifact. The concept is elegant, no doubt, but the reality is something else entirely.

 

Maybe someday...

 

Grey

[AnotherScott]

people are willing to not just supplement but fully replace landlines with inferior sounding cell phones.

 

Hey, my phone sounds great!

 

The tech keeps improving, and I'm sure there are cases where cell phones sound as good as land lines, but it has generally been not as good, and yet that has not stopped many people from making the switch anyway.

[Synthoid]

That's true... games and other apps are quite tempting.

 

 

[WesG]

Ironically, my cell phone sounds better when used with Skype than placing direct calls.

 

The fact that the phone system has shelving EQs under 300Hz and over 3kHz doesn't help.