Beta blog: development log with a bit of random yak shaving


#1

Hello and welcome to the beta development log. This is more of a work diary. I like to think deeply about the design / manufacturing / user experience and have always enjoyed standing by the water cooler and sharing work details.

Because this is an open source project, we’re able to discuss the design rational in near real-time. I’m really passionate about engineering, design, and great documentation and it’s really important that I share that so other people can learn and participate on the same level. Hopefully, branching off and making something even weirder and cooler.

There are thousands of silly obstacles to success, and the most valuable ones aren’t talked about in books. The shortcuts and wisdom that makes hard things easy is in the heads of engineers and designers everywhere and I want to promote a culture of sharing design methodology.

I’m also excited to learn about my design shortcomings. Designing in a proprietary environment makes people secretive about their work. It’s hard to get much feedback or share industry best-practicies, because proprietary developers don’t share design on the same scale as, oh everyone on the internet. I’m not poo-pooing proprietary anything. I am saying, however, that we are all blessed to live in a world where we can put our work in public, knowing that it is not perfect, and we can work together to find out what that perfection is.

So, with gusto I welcome you and thank you for your time and your passion.

You are welcome to comment on this thread and ask questions, but I request that it be on a topic in this thread. General beta discussions are in another castle.


#2

LED Woes

I was hoping to close the case on the LEDs, but I got a little too spendthrift. I know exactly how I got here too.

I spent the first 14 years of my career working on really precise electronics. Microvolts and Picoamps mattered. Keeping down digital noise was critical. So, I’ve never really learned how to skim the bottom and design digital for as low cost as possible. Also, I would say that losing time by selecting the wrong part usually keeps me from setting myself up to pull out my hair prematurely. But, since I’ve been spending so much time with start-ups, I whipped out that shotgun and just bought a bunch of inexpensive level shifters and started sciencing. Last week was my fourth and last low-cost foray. It’s just too glitchy and too susceptible to line noise. What I should have done was get some nicer dev boards just in case. For now, I’m waiting on the postman for that.

In the downtime, I got this forum setup.

Cases

I’m starting to get quotes back for the cases. They’re gorgeous, and they are definitely going to keep the final price low. The only problem is the lead-time and MOQ for prototyping. I’ve got CNC here at Believotron World HQ :smile:, but waiting 14 weeks to get back a prototype is a looooong time.

There is also a heavy risk if the prototypes don’t go well, so I may have to have a discount sale for the happy mutants.

What this means is that the first hardware that is going to ship is either full DIY, acrylic cases, or CNC-cut cases.

I’m also working on adding support columns. I’m scratching my head to find something swift and inexpensive to account for the slope. If you have any ideas, I’m all ears.


#3

This is a repost from the old blog

##We’re in beta!

Fresh off the heels of Knobcon, the largest synthesizer convention in the world, we are proud to announce that we are launching the public beta of the Believotron Wanderlust synthesizer.

The beta is free and open to participation. Beta hardware will be available for pre-order mid-october. If you’re intrigued, sign up for the beta.

The alpha really outperformed on many different fronts, thanks to the robust design of the Axoloti controller. I was expecting sloshy knobs, or capacitive touch triggers that had about 20 ms of jitter. I was blown away by how crisp and regular the triggers were. Once we realized that we had a serious jambox on our hands, we whipped up various menu chains to prove that you can get things done without menu-diving. And then we stopped.

We stopped, because as we stared across tables and corkboards of sketches, workflows and ideas, we realized that menu control is deeply personal. If you have a favorite or least-favorite synth, you’ll know what I mean. Some instruments sound profound, but require a strange ritual of click-clacking a gesture across the magical meta buttons. And that kills the jam. So we stopped.

We stopped, because at this point, it’s foolish to design a general purpose, reconfigurable synthesizer without you. Without an instrument in your hands, no flowchart or state diagram is going to predict what makes you productive, joyful, or in a deep synth vibe.

So! Start doing your synthesizer dance! We’re flipping a board and should be ready to take reservations by mid October. If you haven’t signed up for the beta, head on over.

##Hardware Status

  • Knobs: The knobs work wonderfully. Able to roll as many as you have fingers and dexterity. No noticeable clipping or jumping. Was able to connect to a 16-step sequencer to use knobs as pitch control. Volume knob is invaluable. Wish there was room for more knobs.
  • Padcaps: Work flawlessly. Really great reaction time. Its using a naive polling method, rather than using the irq bits, and it still triggers accurately. You could literally get three octopi and they could touch all the pads simultaneously, or with some sort of magical mollusk pattern and it would trigger. Whether routing to menu systems, triggering notes, or samples from the SD card, the padcaps work well. The piano isn’t as expressive as a mechanical keyboard, but I think the functionality for rhythm, simple melodies, and arpeggiation make it an extremely versatile interface. I like it better than the squishy buttons of the small jamboxes from the last 15 years. Those feel good, but they squish up your rhythm. Modern fingers have become accustomed to precise control on tablets and smartphones. The multi-touch on this thing is a dream. We’re extremely lucky we didn’t mire ourselves in six months of capacitive touch tuning. The future is wonderful!
  • XY Joysticks: I picked up these two joysticks on a lark. They’re really tiny on my fingers. I haven’t found a top for them, so they’re a little pokey. They work flawlessly. I’m not sure they’re travel range make them the best, but they’re located right next to the keyboard for pitch / expression bending. I’m excited to hear what people think about these.
  • LEDs: Funny thing happened on the way to the LED store. Against my own nature, I ordered a reel of LEDs from a brokered source. They were the wrong model, and I think they may even be a strange variation on that model. I’m going to get actual samples and put them through their paces, because these LEDs have some really remarkable colors. The original APA102Cs worked just fine for the Makerfaire demo, which means we’re going to have an LED bake-off! I’m also planning on replacing the voltage translators because they had issues with some 3.3V logic systems.
  • SCH: The schematic is a confusing mess. I used a CAD tool that has a lot of positive traits, but was a real nightmare to make clear schematics. I’m going to transition to another CAD tool and redraw the schematics so it actually makes sense. It is monumentally important that both a stranger to schematics and a pro can read, understand, and figure out how to modify. The maze of boxes and wires that I got from this software is frustrating and reads like a plate of spaghetti had an aneurysm.
  • PCB: It needs small changes. Well, it needs big changes, but I’m squeezing out one last release on the old CAD tool. The Beta-0 PCB silkscreen won’t be extremely readable, but Beta-1 will pierce clarity into the backs of your eyeballs. I’m very picky about PCB readability, but copying the files over to a new CAD tool would delay hardware a couple of months. The general feedback is to get something in your hands now and incorporate all the changes into a B1. I’ll make up for it by making a very, very high resolution infographic that explains what everything is and a writeup on how design decisions were made. Until then, if you have a question about a specific part, take a photo and circle what you’re curious about and we’ll talk about it in beta. I want this to be a very open design process. If someone has a question, it’s probably a really interesting question that other people would benefit from hearing asked and answered.
    ##Beta-0 PCB Changes:
  • Realign mounting holes. (Upper right is misaligned)
  • Enlarge potentiometer holes. (Footprints. Am I right?)
  • New voltage translator
  • Remove extra jumpers

##Software Updates

I’ve got a lot of admiration for the Axoloti patcher environment. Creating new objects, adding ports, adding code is starting to come together. It’s a very robust framework, and it’s very easy to build synth models quickly. I’m interested in pushing the boundaries and making dynamic routing and meta menu controls easy. We’re going to be tracking our goals and progress on a public taskboard. We’ll let you know in the newsletter

Beta Milestones

The beta is a work in progress. We will be reacting to whatever the community chooses to highlight and will add time to the schedule for new ideas that are great. We will start moving things to later phases if they are holding back the beta. Dates and schedule will be recalculated weekly. Hardware will be pre-ordered monthly. The first shipment will be phased so we can focus on a small group of new users each day.

Phase 0 (Now)
Small changes to PCB
Pre-order Beta-zero hardware (Oct-15th)
Ship first Beta-zero hardware
Phase 1 (Software / Usability testing)
Develop UX based on user feedback
Develop metal case styles and circulate
Develop new SCH / PCB
Pre-order Beta-one hardware
Phase 2 Beta Shakedown
Design out MFG / DFM problems
Pre-order Final Release
If you have any questions, please feel free to contact us.


#4

tldr; LED Dev delayed. Waiting on a shipment of samples.

The LED level translator development has hit a wall again.

Unfortunately, the samples that I sourced were all either bi-directional translators or single rail translators. Bi-directional translators are great for I2C, but there is something about connecting to my system that starts transmitting data back on the MOSI (Master Out) line. The single rail translators are supposed to take a lower voltage input and adjust it to the 5V rail. My experience with these is there is just enough slosh in the digital signal transmission that the clock is really jittery and the data line is noisy. This is the first board I’ve done in a long time without impedance matching and I can see it’s effect. You can get a winning serial bus on a 2-layer board, but you would need a Co-planar-waveguide trace-matching.

So, I rang the doorbell of every major semiconductor mfg and sourced a 2-channel dual-rail external direction level translators and they should be here on Thursday. This includes a translator I’ve used before, so it should go well.

As frustrating as this all is, once it’s done and I’ll show you videos of the LEDs in action. They’re relatively low cost and I think they ad a LOT to the synth.


#5

The LEDs work!

I wanted to give a quick update before I go down the rabbit-hole of designing the flashy-LED interactions.

The short version was that my knockoff parts were actually reasonable knockoffs, I was just dealing with troubleshooting the voltage level shifter at the same time. I was also temporarily thinking I had parts with a completely different protocol, because essentially what I was doing was working with sloshy clocks and overcoming it by writing a kludge that kept me in the dark for a day.

What really helped was to return to an original single LED wired up “dead-bug style” and confirming that it worked with the new voltage regulator. I’m a big fan of always keeping around your original hardware prototypes and referencing them when things go awry.

Here is the original APA102C sourced from Adafruit (No longer carried as individual chips)

Here is the counterfeit knockoff. You can see the “dead bug” legs in the background.

Now it’s “fully possible” that the LED factory has just refactored their design, or an upstream manufacturer forced a change. Maybe it’s a manufacturing improvement and my “knockoffs” are old stock of an old design. Tested, but true. What is more likely, based on the high number of wire-bonding and flip-chip operations I’ve seen, is that it isn’t a hard problem to solve and duplicating a form-factor / marketing reach is inevitable.

Still, I’d love to find the original designers and support their design. None of my normal distributors or contacts have been able to verify. I guess LEDs are a high commodity business. If you know or know someone who can help me track down the original designer / mfg of APA102Cs, I’d love to find out. I’m also interested in any other LEDs out there that support clocked data on SPI / I2C (No hard-timing requirements, please, the world has enough problems :slight_smile:)

I’m still going to keep sourcing LEDs, but I think if I had originally had the correct voltage translator, I wouldn’t have noticed the difference.

Pricing

I’m still waiting on some price quotes. We’re also going to have to re-design some of these files based on DFM reports. We’re still on-track to have pricing ready by the 15th, so based on the history of every project, ever, I’m going to push that back to the 20th. Even if we don’t have final case pricing, we should be able to offer DIY and acrylic case pricing.

Next Steps

The best way to find out the roadmap and what’s going on now is to head over to the public work-board.

For now, we’re going to be making a rough-naive framework for the LEDs. We fully expect to throw it all out after a month or two, but for now we need a basic framework, otherwise we’ll be yak-shaving in the dark. Literally.


#6

#Pricing for Beta
Quick update on pricing. I’ve got all the models, I’ve got all the quotes. I know that at this point in a project, there is at least 20% extra cost that will rear its head somewhere. I also know that other things will find a way to be less expensive. Which means it’s time to get hardware into your hands to figure out what works and what needs tweaking, otherwise that 20% cost or savings won’t be known.

I’m going to take the weekend to meditate on the pricing. My goal is to offer up solutions that are as low cost as possible, and offer up enough different solutions for the DIY crowd without creating a confusing mess of options. The prices are good right now, but there are too many choices.

Here is a rough summary of the variations we’ll be offering:

  • Free: It’s open source. Just download the files and off to the races!
  • PCB Only: You know what to do.
  • Full part kit: Just add SMT and Through-hole soldering
  • TH Soldering Kit: Solder a handful of knobs
  • Fully Soldered (With or without Axoloti)
  • Synth with enclosure. (Acrylic, 3D Printed, Metal, Plastic)

#7

PCBs are out for fab!

Hi synth lovers!

Since we’ve finished ironing out the bugs on the LEDs, we just finished updating the change backlog on the PCB and have sent away for boards. They’re green this time instead of blue, because at low volumes, a color change moves us from a hands-free process to one that I pay an engineer to use the hands free process :confused:, but the final boards will be blue and white. We got a really good deal, which let me drop the price of the system by quite a bit.

If blue is your color, let’s get a blue team together and we can do a group buy. Just start a thread in the forum for features. That’s what it’s here for.

#Beta cart WIP
Remember when I said we could have pricing done by the 15th, but I was going to pad it until the 20th because hunch? Well, I was right on the time for the wrong reason. I thought that getting updated quotes and cost modeling would dragging on and on, and we would waffle at the prices. It’s a delicate balancing game between 100 and 1000, but the numbers really worked out well for everyone.

The problem was that we wanted to offer up a range from full DIY kits (like solder all the components on yourself), all the way up to fully enclosed, ready-to play synths. Industry wisdom says to offer the fewest SKUs as possible, but pretty much every Eurorack enthusiast asks if there will be a kit. Some people want to make their own cases. Some people just want to turn the thing on.

I think the beauty of the way that culture and technology have warped the product development process, is that we get to engage with the people who will love our instruments so much sooner. Instead of waiting another 2 years and emerging with what we think people want, we get to talk to you now. And there is wide enough interest in the hardware hacking, software, arduino, musician, artist, performer, teacher, goofball communities that we really need to offer up a wide array of solutions to see where people take them.

The Wanderlust is primarily a controller for an Axoloti, the best open source synthesis engine. It’s never going to leave that behind. But there are also connectors for Raspberry Pi, Adafruit Feather, and you can use headers and wire jumpers to prototype with any microcontroller system.

I fully expect someone to prototype a non-music application using the controller, fork the project and poof! There’s a new thing. I also fully expect that someone starts with our controller and then leapfrogs us with something even more amazing. Poof!

I’m sure that a group of performers are going to huddle around the controller and wish that they could split them off into free and separate controllers.

These are all radically different places to come from. I’m used to mono-marketing. Convincing engineering directors the wisdom of a certain path. Displaying beauty to a consumer. Listing off deep, deep lists of technical details to a logical mind. Communicating with artists about truth, time, and expression.

I have never had to do these things simultaneously, but it’s pretty fun! There’s an old expression “when you get to 95% complete, you have to shoot the engineer and release the product”. Ours now is “when you get five different market cultures together and you sort of aren’t totally confusing, send the information designers home!”

Bear with us. We’re trying to get this yak-hair out the door!


#8

The beta-0 pre-order is live! Now’s the time to get your hands on a synthesizer!!

We have ready to play instruments, as well as DIY kits to get playing for less.


#9

#Parts, shipping, delays.
Luckily, I have good humor about these things.

We have a program in the States that redirects shipments to your primary address to an “access point” when you’re not home (usually a mobile or convenience store). In a hilarious “Who’s on first” moment, my phone number got associated with the shipping location. Calling customer support was a strange loop of circular calls. I call them asking where the package was. They call me asking where the package was. Luckily nothing got lost, but by the fifth loop I was in tears, laughing.

What is a problem though, is one of the main components got gobbled up and spat out in a DOUBLE MERGER. NXP bought Freescale, and just last week Qualcomm announced that they are buying NXP. The part is on an extended life plan by a third-party chipless fab, but because we’re bootstrapping, we postponed buying parts (Which 18 years of experience has told me is a bad gamble). Someone slurped up all the parts on the market and we may not get parts until Late January. We’re working on getting brokered parts, but that is another thing that has failed catastrophically in the past (counterfeit parts, boo).

Less of a problem is our larger print-bed 3D printer being delayed. I made the mistake of buying a prototype add-on that doesn’t ship until the end of November, and the printer got delayed to match the ship date.

If I can get the Freescale-cum-Qualcomm parts in-house, I can use one of the hundreds of 3D printers in Chicago to make shipments by Dec 15th. I will keep everyone posted.

Other than that, I’ve been repairing our Alpha boards so that more software development can happen in parallel. Our prototype stencil was off by a hair and many of our LEDs and other chips weren’t soldered on correctly so we only have a few fully working boards from the Alpha until the Beta boards and parts show up (brand new steel stencil!).

The good news is that we’re likely going to spin a thin LED board so we don’t have to worry about reworking a surface mount connection so close to the knobs. The good news is that the extra PCB and labor cost should make up for the difference in scrap, so the total price should go down.

As always, product development is slowed by mundane things. The complex technical challenges of EE and ME design are small in comparison to moving parts around the world. Always good for a chuckle.


#10

Good news!

I got a lead on the capacitive touch chip, so we don’t have to wait until late January to build these beta boards.

Now it’s just a waiting game. I’m going to continue to work on the Alpha boards and developing patches.


#11

PCBs are here! They look great, and they work perfectly with the potentiometer footprint. We were clipping the leads off all the pots to get them to fit in the supplied footprint.

We are still waiting on the 3D printer and the capacitive touch chips. Once they show up, we will hand solder 3 prototypes and see whether the boards have any fatal flaws.

Today we’re opening the hood of the software and re-architecting everything. We are hoping to have a pleasant out of the box experience by the end of the beta, but for now we have to make as many improvements to first-time usability issues. We’ll document them here as they pop up.

We are also making big progress on a secret project. I hate to be mysterious, but we’re building a lot of cool tools here to help out in the development and manufacturing, and one of them is close to being done. Let’s just say that one of our home-made yak-shavers is saving us a lot of time and we want to make sure everyone with a 3D printer, some motors, electronics and fasteners can use the same technology at home.


#12

Quick update. The combination of PCB + stencil gave us really poor yields. Every board required rework and the solder paste gummed up and caused solder bridges.

Instead of pouring a lot of rework time into these boards, I’m going to refactor the surface mount parts onto smaller module PCBs. I’ll still take a crack at salvaging these PCBs, but I’d rather spend the next week designing the boards than populating board after board.

I declare the Wanderlust beta-0 deprecated. Long live Beta-1!


#13

We’ve pushed the yak shaving discussion to the main blog today: http://believotron.com/blog/2016/11/23/rip-beta-0-long-live-beta-1

New design, new boards!


#14

Knob-Knob module is go!

The new Beta-1 prototype boards showed up, and an Engineer’s life is beautiful again.

Above is the 16-ch knob board in the integration fixture. We had 100% yield from the knob-knob board. We’re re-ordering more boards to put together some finished prototypes. It’s looking like the “Knobs-Knobs-Knobs” synthesizer is going to be ready first! Our only gripe is this new batch of PCBs lost the ease of installation from the previous manufacturer, so we’re going to look into how the information on the drill / plate is communicated and interpreted. This is common when switching between manufacturers to find a difference in process that results in a different manufacturing experience.

Touch-Touch module is pitch perfect!

Well, we had perfect pitch. Just the wrong pitch :confused:, so much as I used the 0.65 mm pitch spacing part instead of the 0.4 mm pitch part. I remember reviewing the datasheet and reading 0.65 mm, but I have no idea where the disparity came from. If you watch Better Call Saul, the scene with Chuck at the copy shop is pretty close to where I’m at with this one. (In the scene, he’s certain his brother has used this copy shop to switch two numbers in a legal document. He remembers this because the number is one off from the Magna Carta, showing a rigid and systematic mind). I’m totally fine with making mistakes on boards, but I always like to know the vector. In this case, I think I was not looking at the correct datasheet. I’m not sure how to correct for this, other than to bind datasheets to the part and create a 1:1 link. I’m pretty sure I went online, searched, and opened a datasheet. I had multiple copies on my desktop, but instead went online. I think I may start including the datasheets in my git repo, in order to keep the dimensions and other facts close at hand and tightly coupled. Hrm.

Regardless, the change took 10 minutes, and we’ve sent out for new boards. These are the small touch module boards, so we really only lost time.

Bright Bright Future

The LED boards are also here. I was in the process of trying to bake the LEDs at 60 C for 24 hours when I realized that my potato warmer wasn’t going to work.

Well, so much for a home-grown option. After a quick search online, I found a meat smoker that can controll temperatures below 60C, so I’m going to try that before going down a resistance wire science project nightmare. I’ll post pics and such after I review the smoker. We should have a prototype LED board working either late this week, or early next week.

Solder Station 2016!

We realized that applying solder paste via silkscreen was going to add cost (screens) and still be really inaccurate at our price range. A cheap screen is < $20. A real screen with a frame costs about $250, and a robo squeegee system is thousands of dollars. In order to reduce NRE costs like this, we’re working on an automated solder robot. The first stage is to fine tune the solderpaste dispenser. We’ve sourced an inexpensive glue pump from Taiwan to get a handle on the process. It was messy and created a lot of solder bridges, but I have shaky hands. I think that we can fine tune the process and skip the screens all together. I’m pretty sure that if we were to invest each screen purchase into more solder robots, we could build a process that matched a reflow oven’s throughput. We’ll see. For now, it’s a lot easier than squeegeeing, or applying it with a manual syringe.


#15

Quick update with a more extensive end of the year in-depth look.

##Led Board
The LED boards are functioning. We’re still having a lot of fall out. The meat smoker was a huge flop. A delicious, delicious flop. I didn’t realize it was an outdoor smoker, and that it requires a 10 foot clearance. Oops. Luckily I have a friend who is a chef, and we will smoke meats. I also have a sneaking suspicion, it will do a good job on vegetables!

We wound up getting an actual science oven from a local Chicago supplier. I was really surprised that there was any economic viability in US heating chambers at that price range, but they were definitely cheaper than an import. I’m delighted to support my hometown!

We managed to dial in the new solder station to hit the LED pads and we got to 100% soldering. It’s just, that, there were still broken LEDs, which we think will go away when we bake out the humidity.

#Keyboard Board
We’re still having a lot of difficulty with the solder station and fine-pitch parts. There are a lot of servo-controlled projects, and they all seem to perform well, so we are going to ditch the compressed air and get a tiny servo dispenser set up ASAP to get these parts on the board.

We’re also switching to T5 solderpaste and trying both low and high temperature solder. We’ve sourced a new set of tips that include conical shapes. Because solderpaste can be compressed, it doesn’t behave as a fluid, and can’t actually make it out of the thin metal tubes unless you heat it. Finally, we’re trying an electrical heating element to keep the paste at a more liquid state. We’d done this using a heat pencil, but want to try an element wrapped around the syringe.

In the meantime, we also spun up an alternate design that uses an off the shelf MPR121 breakout board, so we can get betas into peoples hands while we work on dispensing smaller amounts of solder. We should have full prototypes in early January and beta kits / assemblies going out in February. After that, we’ll be making revisions and updating one board at a time.

I’m super excited and I can’t wait to let you all play with these!


#16

Tip from a long term assembler and/or repairer of surface mount components, BLOB the solder on, sorta, don’t worry about solder bridges until the parts are soldered in place properly. THEN take a short piece of solder wick and wick the excess solder away, should be a clean soldering job ( maybe with a tad of practice )

chuck hejkal

long term electronic repair and design tech, and now retired, but willing to help with troublesshooting or design bugs, remote or in person.


#17

That’s awesome Chuck! What sorts of things were you working on? I cut my teeth as a test engineer, so I always get excited to meet new techs.

We zeroed in on what was causing the shorting: I went and visited a friend who has a desktop pick’n’place and he showed that we needed to use a much more conservative amount of paste per pad. The temperature profile difference between the desktop IR oven (T-962) and the industrial reflow ovens as well as the difference in the paste means a desktop assembly needs smaller apertures. Hooray for science!

Thank you for the solder-wick tip! I have really shaky hands, so anything smaller than 1 mm pitch becomes tricky for me. I found the solution was to use tacky flux. I’d only ever used the thin stuff. It gets everywhere and makes boardwash take forever (the capacitive touch sensor boards need to be cleaned, otherwise they go randomly haywire). I found Chipquick SMD219 and SMD4300TF10 to behave better than any flux I’ve ever used. It’s sort of like magic how fast it cleans up shorts and cold solder joints. It’s still a pain to clean though :wink: .


#18

#Quick Status Update
We just did a fit on our plastic enclosure and decided to reduce the size of the boards by 5.08mm from 37U to 36U in order to improve our DFM. We’ve looked at hundreds of off-the-shelf enclosures and most of them are unsightly boxy affairs. Luckily we found one that we really like, but unless we trim 1U off the design, it’s back to the drawing board. We also have a metal design, but the cost difference is huge, so we’re getting the plastic model out the door faster.

It’s really important that we start off with an off-the-shelf enclosure so 1) We can iterate rapidly. The point was never to have one model of synthesizer, but constantly inventing new devices and ways of playing music. The capital and engineering lock-in that happens with plastic injection tooling is the thing that causes most projects to hiccup. Just about everything else can iterate quickly, but tooling isn’t one of them. 2) So others can participate. We want to zero-in on a form-factor that works for a lot of different communities so that they can buy a low-cost enclosure and make PCBs independent of us. We never wanted to own or defend a form-factor, and we welcome people using and modifying our frameworks.

We have PCBs today that function well and we will be assembling these in the next few weeks to put together demo synths for video and local Chicago circulation. Once we get these new slimmer, boards we will start the Beta-1 order/fulfillment process. We’re currently in the middle of the affect of Chinese New Year on PCBs (It doesn’t matter if you buy them locally; this time of year is the time to get to the back of the line unless you have a huge order coming through). So, I’m going to guess conservatively that we can ship by the end of March. We will put the store back online when we are confident of a ship-date.

We’ve done a lot of refactoring to help avoid delays from…more refactoring :slight_smile: . Seriously, though, breaking the boards up into fractional modules and standardizing on a form factor has sped things up a lot, and will allow so much more flexibility.

As always, you can watch our progress, or make your own boards from our github repo:
https://github.com/Believotron/BelievotronWanderlustHardware. It’s a little chaotic, and underdocumented, so if you have a question, please ask away.


#19

#Enclosures

Hi everyone! Welcome to all the folks who found us through the Jameco article. Glad to have everyone here.

The synthesizer continues to be fully functional, but we are still waiting to get the final dimensions of the PCBs and an aesthetically pleasing enclosure.

We thought we had it made, but the tolerance stack-up caused the top module to interfere with the Axoloti core. We’re currently seeking out a new OTS enclosure to replace this one. We have a whole lot of other backups, but they’re all a little bland and square. We’ll still be able to ship prototypes, but depending on how our research goes, we may wait for samples and possibly delay until the new cases can go through DFM.

We wanted to show you how the case is looking, even though we’re going to have to ditch it, because we think it was starting to look good!

Here you can see an early CNC run. You can see really bad melting of the case caused by using a bit that had been milling aluminum. Keep your materials separate!

We’ve mounted the modules to the top of the synth to show how it would look on top of a Eurorack case. The top row is unpopulated, but there is room for another 32 LEDs for step sequencing or whatever your flashy needs are.

#Onwards

On the plus side, modularizing these boards into their specific function and layout has been a real benefit. When you’re prototyping you often get a situation where something that worked one moment doesn’t anymore. It’s really convenient to be able to remove subsystems to identify the problem. Often what shows up is something that can be designed out. It’s definitely divided and conquered the problems. 10/10 would split again.


#20

Enclosures
We’re on our eighth enclosure design (not eighth iteration, the eighth design :confused: ), having to start from scratch with a new manufacturer. One of the more recent quotes came in at double the max cost and the other quote is MIA, so we’re trying another service. If this doesn’t work, we’re going to work with laser cut acrylic for the beta and work on an injection molded design.

Sequencer
We’ll record some videos, but we have a really nice demo synth with four simultaneous instruments on a 16 step per page sequencer. It turns out there is an awful lot of things you can do with a sine oscillator and a simple ASDR envelope. Combine multiple oscillators for even crazier sounds.
Each page can be loaded / saved to the SD card, allowing for more pages of data than you are likely to be able to use. There is a quick and easy way of scrubbing through the notes with a scrub-knob, and I find it’s not only useful to make patterns, it’s also very tactile and fun. There is a lot more functionality to be put into the synth (and a lot of bugs) to squash, but the basic functionality is coming together.

In the process of doing this, we ran out of SRAM multiple times, eventually combining a lot of GUI patch elements into a large, single script. The more I poke around with the Axoloti patcher, the more I’m really impressed on the entire system. It’s a fairly robust C platform, with the ability to quickly build from naive function blocks, and an easy approach for combining them when there is SRAM to save. It’s a tricky thing to build a platform that works well at the high level and the low level and the patcher is one of my favorite software stacks.

Touch Pads
The new CPT112S chip and PCB board is arrived and working 98%. We have one channel that is triggering erroneously, so we’re going to try and reroute that trace to avoid the capacitance stackup. Aside from that it appears to be just as responsive as the MPR121, but only more testing will tell.