Finishing these will put me on the home straight…
Another 4 dual filter cards
3 Likes
maybe you’re playing a joke on us and it’s the same picture all the time 
pretty spaghetti !
3 Likes
8 boards built and no smoke yet, although I did reverse the +/-12v on the DCO analogue board until I layed them side by side and noticed.
6 Likes
6 voices are working, mainly due to lack of parts I cannot install the rest. Voice allocation is fine. Had a bit of a shock when the voices were not moving with the notes but then realized I had the wrong order of DCOs connected to filters. Not noticeable in unison mode, but pretty obvious in poly mode. Also blew up 2x AS3372e, one the wrong way around and another by connecting -5v to the filter out. I’m just a dumbass at times.
5 Likes
Morning all, been busy testing stuff over the past few days, I’m not 100% happy with the DEMUX section with these mcp4922 DACs, seems very glitchy compared to the internal DAC of the Teensy 3.6 on my last build. So I’ve been experimenting with using dual DACs to drive two DEMUX chips each, seems to be much smoother for the stepped CV wave into each DEMUX chip so that’s a big improvement for the cost of one DAC. But I still have a slight glitch on the sample and hold outputs to work out. I found a few issues in the wiring of the voice card, one output was a pin out on the edge connector. So when testing in poly mode I had a few voices missing here and there. I managed to get to 7 voices working, but voice 1 of card 1 was an issue. Slight sounds out of it, but not playable. Checked everything, swapped everything, same fault. Just now I found two wires touching on the VCF output, the insulation had melted between them and caused a short that was almost impossible to see until metering was applied to that area. Phew, 8 voices working.
Update: Fixed the glitchy S&H for the demux, the sound is a lot better now, no crackling in the background when the demux updates and a lot smoother controls.
4 Likes
Just testing the velocity for the filters and amps, note 7 velocity was broken, full +5v instead of the 0-5v swing, checked the hardware, all looks ok. Checked my code and it seems I forgot to initialize channel 7 of the 8 channel DAC, so it affects the CV for Keytrack and the velocity for note 7, doh !
What’s your opinion on using PWM and a capacitor versus an ‘ordinary’ DAC? Any experiences?
I’ve never used PWM, I guess it’s possible but highly inaccurate and noisy. I always use 12bits or higher for smoothness of controls.
1 Like
Could you elaborate on that?
I thought I had, you need a certain number of steps for smoothness and accuracy of the CV to go into the multiplexor. I have never used or ever seen any design that uses a PWM output to generate the required CV. 12 bits gives you 4096 possible steps and although midi is only 127 values for CC messages internally it’s better to have more steps so zipper noise is not noticeable. I’m actually only using 1024 steps, so 10 bits internally so I have to multiply this by 4 to send to the DAC. If you think a PWM output is suitable then go for it, I’d be interested to know if it works. I think for simple controls where CV is not required to be accurate it might be acceptable.
Often 2 pwm output are combined (but you could use more) similar to a Low Byte / High Byte configuration in a counter. The low byte is used for the absolute value of the low half of the amplitude, the high byte for the absolute value of the upper half of the amplitude. So this in theory gives you 16 bits in terms of steps while it costs only two one bit IO ports. You will need to take care of the aliasing I think, but I’m not sure I see why this would be inaccurate.
1 Like
I agree the more bits you can use in the own the better, but it doesn’t make sense as an SPI DAC requires only 3 connections and can be 16bit if needed with very little noise or jitter. Like I said, I’ve never seen anyone use a PWM output to drive a mux where a DAC is normally used, there must be a reason for this.
All 14 boards powered up in a “smoke free” environment. I’ve tested MIDI to the second DCO set and it generates sawtooth currently until in wire up the back plane and get the PWM under control and calibrated. I had to revert to a previous iteration of the power supply, the synth was intermittently failing to power on correctly on the controller side with the bank of 4 meanwells, I would have thought 2.5A at 5v was plenty, but every other power on was failing, so I went back to this triple Meanwell and a -5v generating single PSU. Did the trick, I know it’s untidy and dangerous currently, but it will be all encased and fused etc when complete. This is just a lash up as I e tried 4 combinations of power so far.
5 Likes
What a beast !! Are you saying it is using more than 2.5 amps ? There must some heat generated as well, no ?
I don’t think it draws more than 2.5A on the 5v rail, but it just fails to initialise every other startup on the other power supplies I used. This one starts perfectly every time. It’s rated at 6A I think for the 5v and 4A fir the 12v and 1A for the -12v
1 Like
Ok, I see. Probably not related (and you probably know better) but you could have a look at the thread PSU startup problems
I would have to check, but I think the PSU has started but the main controller Teensy 4.1 does not initialize. I will have to switch back the PSU to check.