BTW, I would love Elektron would publish articles like that, diving a little more deep into the why and the how of these instruments… (yeah, this specific article is more marketing than anything other, but you know what I mean)
1 Like
Absolutely!
as a novice regarding these sort of things i found it really interesting to read a more in depth explanation of the actual components.
1 Like
Looks like a cirrus adc/dac again
1 Like
FPGA means higher clocking then DSP correct?
Usually, yes (but not always). Biggest difference is the workload they can handle. DSPs are more like traditional CPUs in that they accept general code (you’ll write C code for them normally) and can handle anything a CPU could handle (although they are much smaller and slower usually, than the one you have in your computer), while a FPGA are just hardware gates that you happen to be able to control via software. DSPs are usually limited by physics + the software, while a FGPA is just hardware signals so they usually run much faster (only limited by physics).
The “New” Oxford Oscillators are FPGAs it seems yeah.
5 Likes
Don’t tell Stimming.
7 Likes
I’d love this as well, but based on the following answer, it seems unlikely to happen, sadly :’(
1 Like
no company ever does such things… for obvious reasons
This really explains the higher price, did not expect 2 Coldfire’s and two Spartan’s.
It’s also packed so no battery mods for this one.
The Nanya NT5TU128M8HE is a 1Gb(= 128MB) DRAM chip.
There are two, so 256MB memory total. Maybe room for some samples?
The two Coldfire’s are the same as the Digitone, probably for digital sound generation.
I guess one of the FPGA’s could be for USB communication, at least that’s how RME and MOTU also use them. It allows to debug USB communication from the interface side, and thus making stable communication with crappy USB controllers possible.
edit: probably not though, since DT doesn’t have one.
FPGA’s can be used for anything digital that you can imagine, but they’re not magical or very fast.
They’re actually pretty expensive and slow compared to general purpose CPU’s.
Where they shine is on performing very small specific tasks, which can be much faster than on a general purpose CPU. But if you turn an FPGA into a normal CPU, performance is abysmal.
For example, just an Atari or Amiga CPU needs a pretty powerful FPGA to implement.
I doubt that they used C for this, my guess is on assembly.
The fire at AKM probably didn’t help.
6 Likes
Sorry, I didn’t want to imply I know what language they used to program the thing with. Eventually it’ll be Assembly for the DSPs indeed, although annoying to write by hand when C can (in most cases) do just fine, but judging by past Elektron staff replies here on the forum, I’m doubtful we’ll ever know for certain. For the FPGAs it’ll be Verilog HDL with 99% certainty.
Wow, looks like Elektron went all in…, with the chips…
okay, very bad jokes aside, anyone know if a second set of outputs is possible? I almost talked myself into the Syntakt today but realized there is no way to get a separate out for the Kick without sacrificing stereo out for the rest (standalone mode, obviously OB is an option but not for what I need).
I feel like Elektron still hasn’t come up with a proper SideChain solution. I ended up buying some SuonoBuono nABC comps to get the sound/workflow I wanted but seems like the syntakt won’t fit the bill.
I just wanted to make sure that people don’t think adding stuff or fixing bugs is just changing a couple lines of code.
A while ago I had a conversation on this forum that made me aware of the way that DT handles memory management. Can’t remember what it’s called, but it only made sense in the context of assembly.
For the FPGA’s it’s HDL indeed.
1 Like
I could be wrong but I remember someone saying the Rytm has FPGAs in them and they are just for rerouting the analog components for the different machines so no actual digital audio processing going down with them like creating oscillators.
3 Likes
Routing analog signals with a digital chip would be impossible, but using one as a digital controller for the analog components makes sense.
If you look at the PCB of the Digitone it looks like the FPGA is used to bridge/control/combine the two Coldfire’s. Maybe it allocates voices to each one?
On the Syntakt PCB there’s one in almost the same position as DN.
So, two Coldfire’s, each with 128MB memory, bridged by an FPGA. Another FPGA for control of the analog components, and a flash drive. That’s quite a technical accomplisment.
Totally forgot, for the USB audio part, there are a lot of hints towards usb-audio/Ploytec Microchip reference designs like this: USB2.0 Audio JCT2.10 lowest possible latency soundcard
1 Like
Ess said about the Spartan in the DN exactly “handles the communication between the CFs”.
The other Spartan probably implements digital to analog interfaces like in the Analog series.
1 Like
11 Likes
That is where the elektrons come out.
5 Likes
Funny! 
hmm, drainage for unwanted aliasings ? hope it works 
I think someone at Elektron read the EML 101 manual:
“FILTER MAINTENANCE- After every 100 hours of operation apply a sine wave
to the output of the FILTER to back flush the trapped overtones to unclog
your filter.”
11 Likes
lol is that real?
Sounds like it’s taken from an air purifier manual 
1 Like