This probably is no real surprise to anyone but there is a world wide shortage of ICs at the moment, a month or so ago it was pretty much just micro controllers, but now it seems op-amps and even voltage regulators are in short supply. To cap this off a number of IC vendors are seeing this as a chance to “end of life” a few parts as well.
Add on top the complete train wreck that is brexit and the additional costs of buying and selling goods outside of the UK and things go from bad to worse.
And to cap it off, we’re also now seeing supplier issues in the UK, shortages of materials, courtesy of Brexit, and increased costs for parts and manufacturing. Add that in to the previous mix and that finding suppliers who are reliable and consistent in the UK has become almost impossible.
What does this add up to? well, a perfect storm for UK manufacturers like Dove Audio, creating a lot of stress and uncertainty.
The result? Building modules and synths has become somewhat of a chore, dealing with suppliers, HMRCC, taxes, shipping, supply issues and has taken the ‘fun’ out of building synths.
So we’ve taken the decision that we’re going to be pressing pause on things for a while, we’re not closing up, we’re just going to put everything on pause for a while. We’ll be keeping an eye on IC supplies and all the other fun things going on, but for the moment we’re taking a break.
Things are marching on with the MuSeq sequencer for the MU style modular synth. The pre-production PCB is assembled (couple of extra mod wires).
MuSeq has a number of modes and features, so I’ll explain them here;
2 Channels of sequencing, A & B (channel A shown above)
Channel A has the following modes;
Note (1V oct but scaled to 12 notes)
Step Length (variable from 1/32 to 1/1 with triplets)
Gate triggers (1, 2 or 3) commonly called “ratchetting”
Channel B is linked to Channel A, but still has the following modes
Note or CV (512 value resolution)
Gate Triggers (1,2 or 3) same as channel A
you can change the first step,
the last step
direction (forward, backward, pendulum, Fwd/bck and Random currently)
Syncable from MIDI or internal clock currently, will be adding Sync24, USB MIDI sync and also a pulse input from other modules in your system
Digital output (x2), one for Run/Stop and one for clock out (from 1 beat per bar down to 32 beats per bar)
Inputs (x2), can be used for Tempo, Channel A note, run stop, clock input, pattern change.
Bank play mode, this lets you chain up to 8 patterns together, with the added advantage that you can enable/disable patterns on the fly and set a number of repeats for each pattern. So you could have pattern 1 play 3 times, skip patterns 2 & 3 and then play pattern 4 once, then skip patterns 5,6,7 & 8 and go back to pattern 1.
Loading and saving of 8 banks of 8 patterns to an SD Card in text format, making it easy to share patterns.
There’s a way to go with the software but most of the features listed above are already in. I’m confident by the time the first production run arrives I’ll have got everything in.
One of the main goals for this was to make things smaller, I love sequencers, but loosing 8U of rack space to get 2 or 3 channels of fairly limited sequencing didn’t really appeal. MuSeq is 3U wide and has two channels and of course you can get up to 64 steps per channel with variable gate lengths without sacrificing a channel to do that!
I’ve endeavoured to keep the “analogue sequencer” feel as much as possible, but sometimes I think adding some digital into the mix makes for a truly powerful sequencer.
Below are a few images from the prototype showing the various “pages” which are selected by the knob just below the tempo knob.
The NJM2069 is a fabulous sounding VCF/VCA and I really wish you could still buy these new. not only is a great VCF, but it has a VCA built in and a mixer. It’s a very simple chip to use as well requiring very little in the way of external components. If I could get a good supply of these it would make building poly-synths so much easier!
Anyway, on the PCB you’ll see I bought out the extra pre VCF inputs and the CVs for controlling the levels. Just keep in mind you *must* keep the CVs within -5V to +5V (though they operate from 0V to 5V), going outside of this will damage the chip.
I’ve also added a “Ko-fi” page, so if you feel like buying me a coffee (or in all honesty more likely some chocolate) then you can do so on my ko-fi page here.
PCB routing all done and prototype PCBs have been ordered.
We’re having parts fitted by the PCB supplier, though there will be a few we have to do by hand (for example, the through hole parts).
Couple of changes to the PCB; 1) It’s smaller so as to ensure it fits in both the Box11 and portable cabinet styles. 2) A rotary selector switch for the mode select (was previously a pot) 3) An extra switch to enable “bank play” mode (just below and to the right of the tempo knob 4) Moved the jack sockets up a little so there is the correct amount of space for the MIDI sockets. 5) An extra chip to allow updating via USB. 6) Offset and scale presets for both analogue CV outputs, so you can be sure you’ll get 1V/Oct regardless of any power supply problems.
The code base will be the same as before, so there should be minimal work to bring this board up.
The next steps will be assembling testing the first PCB when it arrives (early February) and then designing the new front panel (a few things have moved, so the holes will need to move accordingly) and then shipping to beta testers.
We’ll post a video once we have the prototype up and running.
Relaying the PCB for the next version of MuSeq is well underway.
There are a few changes; * gone is the extra display and rotary encoder as it really wasn’t useful. * Added a switch for mode selection, just below tempo knob. * allowed more space for the DIN connectors for MIDI/Sync24. * shortened PCB so it actually fits in the case. * a new way of updating in the field has been found. * going to attempt to add USB MIDI in, this may or not be included in the final version. It’s very much dependant on what the impact is to the performance of the sequencer.
Routing the PCB is going well, but it’s going to take a while to complete. My aim is to finish the PCB before the new year.
I’m taking some time over the holidays to consider the future and direction of Dove-Audio. But a few things are certain, MuSeq will be resumed, and it’s going to need a new processor which means a lot of rework so will be a while before there are any real updates.
2020 has been a tough year for everyone, without exception. At Dove-Audio we’ve felt the impact, sadly sales this year, 2020, have slumped which means that I am probably going to need to do another Kickstarter to raise money for MuSeq, something I had hoped to avoid. It’s also going to have a knock on impact into the what we can develop, prototype and put into manufacture in 2021.
Delta, the poly synth, is going well at this stage but still has a long way to go. It is making sound and you can ‘play’ notes on it from a midi controller, but there is a way to go and there will need to be a revision of the analogue PCB.
The design we did for Tangible Waves for a wavetable oscillator came out really well, with 7 wavetables and two oscillators (de-tuneable by up to an octave) it sounds better than we could have hoped for. We’re pleased to have been able to work with Robert on this and we’re looking at a Tangible Waves format WTF module.
A few people noticed something odd to the left of my Q119 sequencer from synthesizers.com. This is my trigger expander, basically I wanted something that would let me trigger noise shots, or toggle end points on the sequencer.
Now, synthesizers.com very kindly put a header on the back of the Q119 which brings out the key signals you need to do this, so adding an external trigger module was simply a case of reverse engineering the pin out and making a PCB.
This shows the PCB, and what I wanted to accomplish was that this was something people could make themselves rather than something I was going to sell. Why? well a number of reasons, I don’t honestly believe many people would be interested in buying and also the cost of the switches would make it quite an expensive module to sell, for what it does.
So I made it through hole, this was a huge pain in the butt, routing PCBs with surface mount parts is so much easier because you can run tracks under pads, with through hole, every track has to go between the holes. Add to this the fact I wanted to keep it 1U wide.
This PCB ended up being 4 layers, but is still quite cheap (4 layer PCBs from china only cost a couple of dollars more than 2 layer versions).
You’ll need to be careful about the order in which you solder parts on, I’d suggest that the second to last is the switches and the final parts are the LEDs (so you can position them at the correct height. You’ll also need to solder ground and output wires to the three connectors which are not mounted on the PCB.
In the ZIP file of the project I’ve included the schematics, gerbers and KiCAD files you need so you can make your own, or tweak the design as you see fit.
If I was to make these again, I would pretty much certainly make it surface mount to make it easier to route and this would make it cheaper to manufacture in quantity.
Another quick update for MuSeq, This time I’ve added the ability to set the “clock out” rate, so you can go from 1 clock per 4 beats all the way to 32 clocks per 4 beats. A short video is below
There are a few things left to do, so here’s a list;
Clock in from front panel (feed it say 8 clocks and get 24 PPQN resolution)
Self updating, i.e. you can update from SDCard
Sync 24 Out
Some internal tidying up of sequence loading/saving, particularly when in Bank play mode.
Redesign of PCB and Front Panel
The big hurdle is going to be the self updating, it always is. As it essentially requires you to boot into a bit of code to check for updates, then clear itself and load the main code, add in the ability to make it so you can’t brick it and things become quite difficult quite quickly.
The remainder should be straight forward, then it’s a case of getting prices and looking for a couple of beta testers before it can be shipped.
We’ve finally found a suitable place to relocate to and we will be moving the end of August, this means we will temporarily not be be able to complete and ship further orders until around the middle of September.
This has taken a huge pressure off me personally, as have a couple of major life changes in recent weeks. My mood has lifted significantly in recent weeks as a result.
In short, I’m going to drop my current project list down to just three things and try and focus my efforts on those, they are;
MuSeq – My MU sequencer, there are a number of hardware changes that are needed, so I will be looking at these first. I did post a recent video with an update on some extra functionality that has gone in on my YouTube channel.
Delta – The basics are there, the analogue VCFs and VCAs work and the oscillators are all working. I’m considering a slight change to the VCF to give it a fourth filter mode and there is likely to a processor change (which is why it’s on headers) . I am pleased with how small it is, 110mm by 110mm for four voices.
PJ5CPU project, this is a bit of fun for me, basically it’s a RISC microprocessor made from TTL logic with a single instruction per clock cycle. I’m hoping to achieve 1Mhz/1MIPS from it, but it’s really just for fun and is not aimed at becoming a commercial product.
There are other things I would like to do, for example more modules for Tangible Waves, but for the moment I need to try and keep my focus.
Whilst Dove-Audio remains a hobby/cottage industry things aren’t likely to happen quickly, but I will post updates when I can.
Just wanted to give a quick update about things going on here at Dove Audio.
Things on Delta have halted for a short while whilst I finish a couple of projects, the NJM2069 Module was one of those and the MuSeq is the next.
MuSeq is significantly behind, basically Dove Audio is just myself, Paula, with the odd bit of help from my partner Jon. I hit a major bug with the MIDI handling in that the library I was using (which seemed to do everything) wasn’t interrupt driven and you had to poll it. Which meant that when syncing to MIDI clock it would sometimes miss clock when switching pages.
Reworking, or refactoring, this took a while. But it’s working. This blow kind of took the wind out of my sales, hence finishing the NJM2069 module first. Now the NJM2069 is out, I am working on the MuSeq again.