Poppin’ Caps

February 24th, 2014

The FPGA graphics engine (started around the last blog post from 2012) now has screen capture support! Here are a couple action shots from recent synth work:

Poking Around

Anti-aliased scope plotting!



You may wonder why the color scheme is so hideously fantastic. That comes from the loving care that went into picking the perfect colors and definitely is not just because this is an early prototype where the color scheme has been ignored…

Noise Sweeps

February 17th, 2014

Click here for audio

I’ve been deep in hardware development for quite a while now, so the blog and my responsiveness to e-mails have suffered. The good news is that something new is almost ready to show. This audio clip comes from a POKEY engine running inside an FPGA. The sweeping is the sound of modulating the POKEY clock across several octaves. I’m also varying the pitch in the registers of the POKEY engine, which is what causes the sudden changes in timbre.


Cheap FPGA Graphics, First Attempt

December 30th, 2012

I need to find a better way to record video. An iPhone camera doesn’t do justice to the 70 frames per second of the screen.

This was a weekend project proof-of-concept for a low budget FPGA GPU. To be fair, it took a couple nights beforehand to get the display configured to recognize an RGB interface. That configuration is handled by a PIC at the moment.

The FPGA design is made of several engines running in parallel:

1) Background grid generator

2) Audio data plotter (including a UART receiver for getting data from a computer)

3) Geometry engines (one per shape)

4) Parallel RGB interface (this drives the LCD)

To keep the memory requirements tiny, there is no frame buffering of any kind for the geometry engines. The RGB interface steps through each pixel on the screen and passes the current coordinate location data to the other engines. Each engine returns a 1 (pixel on) or 0 (pixel off). The last step is a mixing block that defines the color and depth for the output of each engine. For example, the lowest layer is the background grid. Next up is the circle, and so on up to the audio plot. Only the highest illuminated pixel is shown for each location on the screen.

The geometry engines store the size and locations of their shapes in registers, so the only memory usage comes from the audio data drawn on top of everything. Audio data is being generated by Max/MSP and sent over a USB > serial cable. The FPGA double buffers it so that there aren’t any glitches due to synchronization issues.

A major downside to this general approach is that the amount of logic scales proportionally to the number of objects on the screen. If there are going to be 5 circles on the screen, this design needs 5 circle generators. One way to improve that is to use a faster clock in geometry calculations. Then each engine could handle multiple shapes per pixel.

Long-term goals: thick lines, anti-aliasing, sprites, and text rendering.

Tech details

The display is 320 * 240 pixels running at 70 fps. The complete design as shown uses 415 logic elements, 8192 bits of memory, and three 9-bit multipliers (for circle calculations). For a reference, the smallest Altera Cyclone IV FPGA has 6272 logic elements, 270 kbit of memory, and thirty 9-bit multipliers.

What does your soul look like?

July 30th, 2012

…because this is what a set of roughly 300 Nintendo NES CPUs looks like


It might be easier to think of in terms of square feet. In that case the number is four.


July 29th, 2012

The 2nd version Atari synth prototypes came back from assembly early this week, but something has them bricked.
The worst case is that there might be hardware damage (soldering too hot, static zap during handling, moisture problems), since that leaves little I can prove in a post-mortem. At $500 a run, prototypes are not cheap.

<geek notes>
The power supplies measure fine, and JTAG diagnostic tests on the FPGA pass with flying colors. The configuration pins all check out, too. Yet somehow the FPGA will not run. It is not simply a code issue, because the part won’t even run factory code from the manufacturer.
The 2.5V I/O banks appear to be driving out 0.45V intermittently, which can’t be a good sign of anything.
</geek notes>

So that is unfortunate.

To get back to something that *is* working, the 2A03 synth is coming along. It currently has voltage control of:
Pulse wave 1 pitch
Pulse wave 1 volume
Pulse wave 2 pitch
Pulse wave 2 volume
Triangle wave pitch
Noise wave volume

…and knob controls for:
Pulse wave 1 pitch
Pulse wave 1 volume
Pulse wave 1 duty cycle
Pulse wave 2 pitch
Pulse wave 2 volume
Pulse wave 2 duty cycle
Triangle wave pitch
Noise wave pitch
Noise wave volume

Click here for audio of one pulse wave modulated by a couple envelope generators

One of the downsides to using the actual NES CPU is that it has unobvious quirks. It would be far easier to make an FPGA clone of the 2A03 audio engine and just fix the quirks, but what’s the challenge in that?

The pulse wave pitch control registers are a good example. They are 11 bits wide, which isn’t anything special. On the other hand, any write command to the greatest 3 bits will cause a phase reset in the output waveform.¬†In a videogame soundtrack where each note starts with a volume of 0, it might never be a problem. In a synthesizer where oscillators are “on” constantly, it leads to clicking as pitch changes.

I am experimenting with workarounds. One trick is to use the pitch sweep hardware built into the part to trigger an increase or decrease in the upper 3 bits without issuing a write command. That has two downsides:
1) Worst case, it takes almost 20 times longer than just letting the click happen
2) It is more difficult to keep stable

It takes so much longer because it involves writing several non-pitch registers multiple times. That ends up lowering the sample rate on pitch control voltage. The stability issue is just that I see the pulse wave pitch occasionally lock up and refuse to change for several seconds. It is probably due to some corner case I have not found yet, but it never happens in the clicky version.

I cannot take credit for the clever idea to use the sweep unit, by the way. That came from the nesdev.parodius.com forum.

I made a demo to show the difference. The following audio file alternates clicky/smooth/clicky/smooth through some pitch modulation.

Click here for audio

2A03 Triangle Arpeggios

July 9th, 2012

Volt per octave control is kinda sorta working.

Click here for audio

It really needs to be calibrated, and it also needs some filtering to get rid of noise. Soon…

First CV Audio From 2A03

July 8th, 2012

I needed a break from working on the new Pokey.synth code, so I took this afternoon to get the DE0-Nano input board (mentioned in the previous post) working with an audio channel on the NES. Today’s wave: the triangle.

This is a Make Noise Maths plus a little knob twiddling feeding the triangle wave pitch of the 2A03. It’s not following the volt per octave standard yet, but it’s smooth and responsive like a modular synth oscillator should be.

Click here for audio.

Gold Plated

July 4th, 2012

So naked!


The past few months have been a blur. Who even knows what the boards in the top row are.

The board on the bottom is an input control board for interfacing an $80 DE0-Nano FPGA development board with modular synth controls. It has:

8 analog inputs

8 logic inputs

8 potentiometers

4 endless rotary encoders

8 LEDs

2 pushbuttons

2 toggle switches

Can an FPGA support all that at once? Yes, because FPGAs are great. I’m a little nervous about the analog input support. I am feeding the A/D converter with 5V power from the DE0-Nano, but the 5V regulation on the Nano is questionable. The board on my desk right now measures 4.71V.

Hax0rr3d, apparently

April 1st, 2012

Someone is in my base, injecting all my SQL.

This site was hacked at some point in the last couple weeks. I was alerted this morning when I tried to moderate spam comments. My browser pleaded that I turn back, for the site had become a wretched hive of scum and villainy. I am disabling blog comments to be safe so that there is no future risk of malicious code getting in through a comment.

HostGator was great about cleaning up the malware on the site.

If you have visited any part of skrasoft.com in the past couple weeks, please do a malware scan on your computer. I have no idea what the site was trying to install. I recommend Malwarebytes for scanning, but I may not give the best advice since I also get hacked.

Early Tests of Max/MSP Controlling 2A03 (NES CPU) Audio

December 12th, 2011

My last post mentioned that I had an interface working for controlling the 2A03 at a reasonably fast rate. Last week I wrote a serial data receiver (UART in this case) into the 2A03 control logic so I could send the 2A03 commands from Max/MSP. I technically could’ve used MIDI, but at 31,250 bps it is far too slow for all the data I want to send as quickly as I want to send it. I opted for 1,000,000 bps with the option to speed it up further later.

The chain looks like this:
Max/MSP -> FTDI USB-to-UART cable -> FPGA development board -> 2A03 adapter

It works just as well as I hoped. I basically have real-time control over all the NES audio registers from Max/MSP. Here’s a screenshot of the big ol’ mess in Max:

It serves as a quick idea prototyping/test platform for the NES synth module, so it changes too frequently to bother making it tidy.

This post is boring without audio, so how about a few test recordings? These are all straight from an NES CPU into the preamps on my sound card.

Click here for clip 1 – triangle wave FM

The pitch of the triangle wave is modulated by a sine wave in Max/MSP. There is a lot of jitter in the timing not only because of the serial transmission, but also because I have to convert the sine from an audio thread in Max to control data to use with the “serial” object. Control data timing is not designed for audio. It worked decently given the circumstances.

Click here for clip 2 – pulse wave delay

One of my favorite Nintendo sounds is two pulse waves configured as an echo effect. I am just playing with a few settings (delay time, second wave volume and duty cycle). The percussive nature of the voices comes from an envelope with an attack time of zero and a short decay to silence. The envelope and all other data is being completely handled by Max and dumped out over USB. The pops in the audio are from a known control issue. They will get better and go away as I tweak the control logic.

Click here for clip 3 – noise bursts

I couldn’t post audio tests without including the noise channel. This is a steady stream of short bursts (zero attack time and fast release time) with pitch and timbre being adjusted by hand then linked to the amplitude envelope in fun ways.