welcome: please sign in
location: x0x-heart


About the x0x-heart

The x0x-heart is the heart of the x0xb0x, surgically extracted from the x0xb0x to be transplanted into your designs. The x0xb0x is a faithful replica of the Roland TB-303, complete with sequencer and DIN/MIDI/USB. The x0x-heart is just the analog section of the TB-303, replicated with SMT components, and made small enough to fit into a Eurorack modular format. It even has a Eurorack power header on it.

The x0x-heart is designed to be small, inexpensive, and a close replica of the TB-303. It is not a perfect replica - some sacrifices were made to keep costs down - but it's pretty darn close. It also has no pots or controls, making it great for DIY synth building. All of the useful control points are pinned out on 0.1" spacings (except for J10, sorry!!) for easy prototyping. And, as a special bonus feature, digital control is given for Accent, Slide, Waveform select, Cutoff, and Env Mod.

User Manual

Tuning with the Pacemaker

The manual lists the tuning procedure for standalone operation. Unfortunately, when the x0x-heart was originally designed, there was no intent to make a kit out of it, so the when the Pacemaker finally came around, there were some really awkward design constraints. As a result, all the tuning pots are nearly impossible to access while the unit is assembled. The easiest way around this, is to run jumpers for J8, RES1 (set to mid), TUNE (set to mid), and CUTOFF (set to full up), and measure the frequency at the VCF output. its safer to do it this way as well, as its less likely that something will short.

Differences from the original 303

There are many differences between the x0x-heart and TB-303, most of which are due to components no longer being available, or not available in SMT format. In all cases, choices were made to keep costs down with minimal impact on the sonic quality of the synth. Here is a listing of some of the more notable differences:

  1. PMP4201/5201 matched pairs used throughout (in place of both singles and duals). These have very similar characteristics to the 2SC2291 and 2SC1583 matched pairs used in the 303, but have better matching characteristics. They are mediocre replacements for the 2SA733 and 2SC536, but it reduced part count to use the same parts throughout.
  2. 1% resistors used in place of 5% resistors throughout. The 1% resistors aren't significantly more expensive, and it makes the synths behave more consistently (which could be a seen as a drawback).
  3. Ceramic resistors in place of electrolytics and poly capacitors. This will have a small effect on the sound, but is unavoidable in an SMT design, as the poly caps in SMT are prohibitively expensive, are difficult to solder, and aren't as good as the THMT versions anyways. High voltage, and high quality dielectric ceramics were chosen to help eleviate any problems here.
  4. 3300PPM tempco thermistor for R100 instead of 3500PPM. This should actually give better temperature matching.
  5. 220k/0.1uF slide components instead of 100k/0.22uF. This should have negligible effect on the slide timing, but might increase offset voltage slightly. The feedback resistor on the CV buffer was also changed from 2.2k to 220k, which should eleviate any offset voltage problems (it will probably be better than the original).
  6. LM358 instead of AN6562. These are very similar amplifiers, and should have no effect on the sound.
  7. 2SK3666 instead of 2SK30A. These are very similar JFETs, with the same IDSS, and should have negligible effect on the sound. 2SK208 would be a better option, but these had low availability during the design phase (that seems to not be the case now). The main difference is that the 2SK208 has the same Yfs as the 2SK30A, whereas the 2SK3666 has double that value. The 2SK3666 also has a sharper knee in its ID versus VDS curve, in comparison to the 2SK208 or 2SK30A.

  8. Slightly different powersupply. It is a switching powersupply, with very similar frequency, output compliance, and output voltage. A definite improvement over the x0x. But, it can only take 5VDC MAX!!

  9. Digital waveform select. There is a bit of bleed-through that you don't get with the mechanical switch, but you can do so much with the digital control that it seemed worth it.
  10. Discrete BA662 implementation. Since this IC is no longer available, i reverse engineered its internal transistor layout, and just built my own on the board. details can be found on the x0x wiki page.

  11. Replacement of varius resistors with equivalent parellel/series combinations that add to the same value. This was done to save on part count.
  12. No headphone amp. This part is no longer available, and it seemed not critical to the design, so in the interest of cost it was cut.
  13. Level shifter ICs instead of transistor versions. ICs are cheaper, easier to layout, require fewer components, and work identically.
  14. Component designator changes. Not all components were used, and some were added, so they could not match in all cases. It is also important to note that the x0x-heart matches designators with the x0xb0x, and not the 303 schematics. The 303 schematics have errors in them (inconsistencies between schematic and PCB), and the x0x matches the 303 PCB, not the schematic. A bit confusing, but it seemed like the right thing to do at the time.
  15. Digital control of Cutoff and Env Mod. This can be turned off, with no effect on the sound. But, there were extra CD4066's, so why not?



x0x-heart (last edited 2017-01-24 10:41:06 by guest)