Vintage sampler (E-mu SP1200) re-creation and improvement

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

ryanvane

New member
Joined
Nov 2, 2014
Messages
3
Hey everyone, this is my first post here. I've had an idea to create a sampler with the same D/A and output stages of the SP1200, but with the following changes:

- Modern digital parts (besides the DA converter and SSM2044 filter chips) and increased memory and MIDI response speed over the original sampler
- Improved power supply (since I understand the original had some issues)
- No sequencer... functioning as a rackmount MIDI-controlled module only
- Saving to SDcard

I'm confident that I would be able to handle the digital components and microcontroller programming to make it function, but my knowledge of analog circuitry and power supply design is nearly nonexistent. Would anybody be interested in assisting with such a project? I have the original service manual and schematics.

Right now, this project is just an idea, but ultimately I'd like to make the PCB designs and source code freely available for anybody to build or improve upon.

Cheers

Ryan
 
I still have  a SP12.  The sequencer was so tight on feel.  The 8th note triplet was great.  The sequencer was a great part of the machine for the time period.  That said I still don't use it. 
 
The ADC and DAC are what makes this unit sound special. It adds punch to drums and basses like nothing else.

The magic is very much in the sampling process, the SP-12/SP-1200 uses the DAC to do the sampling.  The parts should all still be availible.

The input has a complicated TL084 based fixed active low pass filter arrangement.

I've long been thinking about building just the sampling part to have a simple box that can transform the sound, but I couldn't handle the digital part. ;-)

It would be really interesting to use the old DAC but clock it at a higher rate. The only reason for the low sampling rate was the low memory.

Anyway, I think this is a great idea.  :)
 
I would want to initially leave the sequencer out, because that isn't really the "magic" part of this machine. Converting this project into a full-fledged SP1200 replacement wouldn't be a very difficult task (compared to the rest of the project). Add a GPIO expander or two off of the main microcontroller, and replace the MIDI input code with sequencer code. Leave the sample engine code the same. Booya.

Living sounds, the thing about the sampling part is that you would have to have 8 output channels for accurate output, even if you're only interested in playing one sound at a time. You'd probably also be limited in how far you could raise the sampling rate, as well. The reason is that all 8 channels were fed by a single multiplexed DAC. The short version is:

- One channel's 12 bits of sample data is fed to the DAC
- DAC feeds another 8-bit DAC to set sample level (first DAC's output is used as voltage reference, 8-bit sample data sets level)
- A capacitor holds the analog output of the level DAC for its channel (my apologies for bad terminology... electronics is not my forte)
- Capacitor feeds output stage
- Switch to next channel, repeat

If each channel's sampling rate was 26.04 khz, then the DAC would have to run at 8 x 26.04 = 208.32 khz

The sample clock seems to be several steps removed from the crystal, as well. It goes through a bunch of TI counter ICs and through a PAL IC. The clock would probably be difficult to reproduce without knowledge of how the PAL was programmed by E-mu.

I wish I never sold my SP12. :(
 
You guys are making me want to dig out my unit and play with it.  I also still have an E3 also that just sets in the same closet.  I did so much work with those back in the 80's I could never part with it.  I guess I'm a gather.  You guys are hunters.  Good luck on the pursuit.
 
HI, Could you give me originak blueprints of sp 1200
I can't find it anythere((
sorry for my bad english(
 
Hi, Did you get anywhere with this? I am also researching recreating the D/A and output stage of this sampler.

For now I am thinking of using Raspberry Pi to control the DAC and multiplexing, because of its performance, I/O and other features that should make this convenient.
 
There are at least two projects of SP 1200 re-creations I know of. However, at least one of them is using modern SD-converters, so it will definitely not sound the same.
 
Thanks!

I figured out I'll start with getting a clock and then build around that.
I wanted to try something naive first. I did setup a HW clock using pigpio to 208320Hz, but the jitter on that clock would make it unusable.
So I got a board with Si5351 clock generator. After a few hours I cobbled together a code that was generating stable 208320Hz clock.
Now I wanted to setup a callback that would run a function on a rising edge, so I connected the clock to one of the GPIO inputs. I suspected this is not going to work, but had to try.
I setup a callback that it would flip a GPIO output every time it is called. Ideally I should see square wave with half of the frequency. Unfortunately it was struggling to go above 60kHz.
The code was in C on Rasbian.
I think I should look for a way to run it in realtime or maybe change platform or approach.
 
I have a feeling that probably Raspberry Pi is not going to be feasible platform to do this. I am going to try with STM32 and / or learn FPGA stuff.
 
I started dipping my toes into FPGA and it looks to me now that it will be the most appropriate way to get the timings and data flowing into the D/A, VCA and multiplexer circuits. It's going to be a slight learning curve, but last night I got the diode blinking on a FPGA dev kit  :D
 
Now that I look into FPGA stuff, it looks like it would be possible to load Z80 core and actually run the original OS. Problem with that is figuring out how all the I/O is wired up, memory map etc. I would have to look up how people reverse engineered old consoles, probably that would be similar.
So far I got the clock and signals for multiplexer running on FPGA. I also figured out how to run Picorv32 core (and C programs).
I would map a channel structure to memory, so the program would provide sample start, end, pitch and then the logic on FPGA would "play" the sample once program flips "play" bit. Once I get this working I'll get on wiring the AD7541 DAC and surrounding op amps.
 
Now I started playing with the original DAC and I/V circuit on the breadboard and that surprisingly seems to be working fine, but there is a lot of noise and glitches - I guess this will be gone when aliasing filter is added.
I have not run any samples yet, just attached a simple 12bit counter from Arduino board to make sure conversion is actually happening (i was initially setting up bits manually on the breadboard, but sending something off Arduino is more fun).
So the next part will be the attenuation with the 8-bit one and multiplexing the outputs.
With that I learned that maybe I need to use something else than a breadboard, as I had few times e.g. an op amp jumping out and some connections are intermittent. Since the DAC I/V was uneventful probably the rest will also work, so I will just look into creating a PCB for that section (and something to learn as I never done a PCB before 8) ).
For ADC bit, interesting is this part - 74HCT504 that is a successive-approximation-register. It's not available anywhere, but it seems to be the same part (although maybe working with slighty different voltages) as AM2504, which has datasheet available as is on eBay. I am going to try to recreate it in FPGA in the future as well.
 
Got some more progress with this - basically I got the DAC - I/V, digital volume control / amp, 8 channel multiplexed sample and hold, pitch algorithm working, then I built the main clock out of the digital ICs and now I am waiting for more op-amps to build the ADC section with successive approximation register.
 
Interestingly Rossum Electro-Music has lately started to sell refurbished original SP-1200s with some enhancements including SD card floppy disk emulator with software integration for $7500 a piece.

http://www.rossum-electro.com/products/sp-1200-renovation/

Trade in customer with SP-1200 have priority.
 
Back
Top