Monday, January 6, 2014
Saturday, January 4, 2014
vocoder panel design
Here is my 99% done panel design
The functionality i need to build and verify before I get it fabricated are:
All the pots and jacks will be panel mounted. The sliders / led / patch panel will be pcb mounted. Here is the pcb for those components
the case will be an aluminum rectangular box folded from a sheet. i would like to have wooden side panels which mount onto the aluminum sides which are have one angled side. the angled side can be changed in order to either lye the case flat and have it sit on a 10degree incline similar to the ems 5000 vocoder. the wooden sides could also be rotated and allow for the case to stand vertically when desktop space is at a premium or you wish to sit it on top of a synth.
The functionality i need to build and verify before I get it fabricated are:
- aux speech input works
- v/uv threshold works okay and is worth having on the panel
- silence bridging AR controls are worth having
- whether to bring UV and V trigger outs to the panel
the case will be an aluminum rectangular box folded from a sheet. i would like to have wooden side panels which mount onto the aluminum sides which are have one angled side. the angled side can be changed in order to either lye the case flat and have it sit on a 10degree incline similar to the ems 5000 vocoder. the wooden sides could also be rotated and allow for the case to stand vertically when desktop space is at a premium or you wish to sit it on top of a synth.
a burst of vocoder progress
I have checked the V/UV and the synth / noise inputs and all is working
V/UV
I had a bit of a scare when testing this module. Voiced was locking on and it seemed that something was wrong. Luckily I tracked the issue pretty quickly - the 330k in one of the filters was 330r. I looked in my 330k bag and it was full of 330r resistors. I then checked to see how many resistors I'd have to replace on the board and luckily there was only one - phew! Large projects and systematic errors are not good - lucky break.
So after this fix it appears to be working. Unvoiced vocal sounds triggers the unvoiced channel. The user 'sound' on e-m has recommend different filters for the unvoiced but I kept it at Jurgen's design since it is based on a ems vocoder. I made no other component modifications the only mod is that I am using a pot for v/uv threshold as i think this is a neat feature to have on the panel.
I might even take the v/uv triggers out to the panel like on the doepfer v/uv module.
The only part that i'm curious about is that the voiced channel is always on even when there is no input. Hopefully someone can give me feedback on that.
Happy with this for now - I'll re-review it once I have some channels working
Synth Input
Synth Input is working fine. I had to replace a dead bf245 but otherwise no issues. 'Sound' on e-m recommend changing some resistors in this section which I followed but I had to revert the 1k8 / 10k which controls the noise amplitude back to 18k and 100k to ensure it was around +4dbu.
I am going to have the synLevel as a pot on the panel as the level should be the same as the noise input and the input may be an iPod, a module signal or a synth so it needs to be controlled by the user.
so where to now...
I am going to skip the slew and silence bridging modules for now. Time to get some channels working. Once that is done I can test the output and hear it for the first time. I will also be able to hear how the v/uv sounds as well as check the slew / silence bridging.
Once everything is tested I can finalise my panel design and send it off for fabrication. I am 99% sure on my current design but since it will be an expensive panel I'd like to verify my choices are all going to work and I'm not missing anything.
V/UV
I had a bit of a scare when testing this module. Voiced was locking on and it seemed that something was wrong. Luckily I tracked the issue pretty quickly - the 330k in one of the filters was 330r. I looked in my 330k bag and it was full of 330r resistors. I then checked to see how many resistors I'd have to replace on the board and luckily there was only one - phew! Large projects and systematic errors are not good - lucky break.
So after this fix it appears to be working. Unvoiced vocal sounds triggers the unvoiced channel. The user 'sound' on e-m has recommend different filters for the unvoiced but I kept it at Jurgen's design since it is based on a ems vocoder. I made no other component modifications the only mod is that I am using a pot for v/uv threshold as i think this is a neat feature to have on the panel.
I might even take the v/uv triggers out to the panel like on the doepfer v/uv module.
The only part that i'm curious about is that the voiced channel is always on even when there is no input. Hopefully someone can give me feedback on that.
Happy with this for now - I'll re-review it once I have some channels working
Synth Input
Synth Input is working fine. I had to replace a dead bf245 but otherwise no issues. 'Sound' on e-m recommend changing some resistors in this section which I followed but I had to revert the 1k8 / 10k which controls the noise amplitude back to 18k and 100k to ensure it was around +4dbu.
I am going to have the synLevel as a pot on the panel as the level should be the same as the noise input and the input may be an iPod, a module signal or a synth so it needs to be controlled by the user.
so where to now...
I am going to skip the slew and silence bridging modules for now. Time to get some channels working. Once that is done I can test the output and hear it for the first time. I will also be able to hear how the v/uv sounds as well as check the slew / silence bridging.
Once everything is tested I can finalise my panel design and send it off for fabrication. I am 99% sure on my current design but since it will be an expensive panel I'd like to verify my choices are all going to work and I'm not missing anything.
Thursday, January 2, 2014
vocoder speech in
This section is very similar to JH FS-1 so I used the same values. It seems that everyone else who has completed this section copied the first suggested values using THAT 1510 instead of SSM2019. I had some leftover SSM2019 from building my FS-1.
The values I used are as follows
Tone:
TONE Pot 50k linear
U2101 = SSM2019
Sp Level Pot - 10k log
Mic Gain Pot - 5k lin however I will get a 5k reverse audio for my panel
R2105 / RG2 = 10r
R2106 / RG1 = link
R2107 / RG0 = nothing (left open)
Limiter:
R2116 / RG3 = 22k (based on advice in e-m thread)
R2136 / RLIM = 6k2 (based on advice in e-m thread)
There was note that the U2101 would become unstable at certain settings and a large capacitor was used in place of RG2 but i didn't have that issue on first test.
Mods (not yet implemented):
I wish to have an aux input on the synth in. If you read texts on vocoders they suggest that either the synth or vocal input should be prerecorded, only one input should be live. Whether or not this is gospel is up to the critics but i would like the opportunity to insert a line level (or modular level vocal input - or perhaps a feedback input). I will follow the circuit used in the FS-1 input and send an input into U2102B. I'll document that once I have made some more progress on these sub-modules.
I am happy with the output thus far. The limiter light is coming on when it should ( i think ) the sp out is not distorted so time to move on.
Wednesday, January 1, 2014
jh vocoder progress
Finally some time to work on the vocoder has opened. At this point I have finished the panel layout for panel components and I have created a layout to mount the sliders. This slider pcb is being etched for me at the moment. All connectors and final components are in stock.
The path to the finish line looks like this:
The path to the finish line looks like this:
- test pcb
- psu
- synth in
- exc in
- all other modules
- one channel
- complete panel design
- aesthetics (fonts / linework etc)
- mount holes for case and pcbs
- case / faceplate fabrication
- wire up
Proud to say today I have powered up the board and the psu is functioning. I had a bit of a stumble trying to use a switching psu - I mistakingly bought a single supply instead of a dual. Not sure how I slipped there since I was specifically looking for a dual. Annoying. Luckily I had a 15v torroid in stock so I used the onboard psu. As far as I'm aware I believe I am the first to use this onboard psu.
Here is the heatsink with the TIP3055 and mica pads
Next up I will test the modules
Sunday, May 26, 2013
vocoder continued
Well this is turning to be a marathon pcb population. i've almost completed resistors and i've populated some other components outside the filters. After so long staring at resisters i think i have finally memorised resistor colour codes :)
And i reached a special milestone - i filled up fetta tin with resistor legs - i guess this was the project to do it
I have reworked my panel design also - this now includes a 20x20 ghielmetti patch matrix.
Thursday, May 9, 2013
Subscribe to:
Posts (Atom)