Yeah, that's true. I have a digital electronics design background, so I guess that I approach it a bit different than a pure analog or RF design engineer.
lol, the same problem again, the trigger input was designed for 5V, it needs a diode limiter to 5V after the capacitor of the trigger input jack. Eurorack "standard" is 10V for ADSR, gates and so, everything will be out of wack if someone using modules from like Doepfer or so and injects 10V. you should definitely add a pot to GND for the velocity input, because it also was made for 5V, but somehow it's 12V in your schematic. I have made a usefull circuit to create a perfect square 1ms, 2ms or what ever trigger signals and it is very important (i also have some Roland TR drum sounds made on vero board on a big plank of wood but i used a PIC18F1320 @ 40mHz, it sounded very weird without the correct trigger signal duration), it's kinda digital, but doesn't require software. Can upload it for you, so you can have a look at it.
The trigger and accent inputs are identical to hex-inverters design, I didn't spend much thought on the shape of the trig-signal, but you are correct that it should be a digital trig signal limited to 0-5V, with fast rise- and fall-time. I'm using my 909 modules with a DIY MIDI-to-trig-interface and I use 74HCT595 (VCC@5V) for the trig outputs. The trig high duration is 4ms.
@@MeeBilt Ya, but you're not building modules for your self if sell them. It should work in any case for other people using different modules from different "brands" to avoid claims, don't trust online schematics. i have uploaded a new circuit video containing a trigger circuit. maybe it will suit you. Why use the 74HC595 for gates? use the 74hc174, much better, just like Roland did in the TR-909 or CD74HC574E with is 8 output. cascade them like Roland did, probably a bit faster than bit shifting, almost instantly, 3 clock cycles. with 3 x 74hc174 you need 6 data lines and 8 with the CD74HC374. You need a diode to 5 volt to limit on the trigger input before heading to the transistors or havok will occur.
@@MeeBilt You could remove Q80, Q81 and the stuff around it from the Hi hat circuit, move R480, R475 between Q85 and the inverted opamp input of IC 67B, copy or move the 1.2nF and 100pF on the same spot as on the Q80, apply a static positive offset voltage of 2.5V on the NON inverting pin of IC 67B, play with value R483, but you have to play with 1.2nF and 100pF, witch is a type of anti aliasing filter, till it sounds right. You safe 9 parts or so. You could also do what i tell, but double IC 67B and have 2 anti aliasing filters, like before.
@@AnalogDude_ Thanks for the suggestions. If I decide to make a product that I can sell I will definately revisit the design and have a look at the overall BOM and maybe add some features. I'm also considering to redesign the trigger circuit (maybe remove the button completely). But I will complete all 909 voices before I do that, so I can consolidate the BOM and learn more about which part of the design that I can reuse between different modules.
@@MeeBilt I'm making a pcb with the 74hc174 part and the (roland) resistors, i did a simulation and also with a normal R2R array, than i added the "VCA", opamp and Q80, Q81. using a plain R2R array avoids special values of resistors and does the same job, but takes up more room, unless you use resistor array's. For the VCA part you could use a LM13700 instead of Q85. it's probably best to use a AT27c512 for the cymbal and crash and use 2 x 74HC174, but that requires a MCU and a few lines of code to jump between the addresses of the crash and cymbal in between (mixing), like done in the tr-707. dropping the button is a good idea, but i guess before that you should make a Sequencer.
Nice job! Always fascinating to see how others design PCBs. Everyone has their own tips and tricks and style!
Yeah, that's true. I have a digital electronics design background, so I guess that I approach it a bit different than a pure analog or RF design engineer.
lol, the same problem again, the trigger input was designed for 5V, it needs a diode limiter to 5V after the capacitor of the trigger input jack. Eurorack "standard" is 10V for ADSR, gates and so, everything will be out of wack if someone using modules from like Doepfer or so and injects 10V. you should definitely add a pot to GND for the velocity input, because it also was made for 5V, but somehow it's 12V in your schematic.
I have made a usefull circuit to create a perfect square 1ms, 2ms or what ever trigger signals and it is very important (i also have some Roland TR drum sounds made on vero board on a big plank of wood but i used a PIC18F1320 @ 40mHz, it sounded very weird without the correct trigger signal duration), it's kinda digital, but doesn't require software. Can upload it for you, so you can have a look at it.
The trigger and accent inputs are identical to hex-inverters design, I didn't spend much thought on the shape of the trig-signal, but you are correct that it should be a digital trig signal limited to 0-5V, with fast rise- and fall-time.
I'm using my 909 modules with a DIY MIDI-to-trig-interface and I use 74HCT595 (VCC@5V) for the trig outputs. The trig high duration is 4ms.
@@MeeBilt Ya, but you're not building modules for your self if sell them.
It should work in any case for other people using different modules from different "brands" to avoid claims, don't trust online schematics. i have uploaded a new circuit video containing a trigger circuit. maybe it will suit you.
Why use the 74HC595 for gates? use the 74hc174, much better, just like Roland did in the TR-909 or CD74HC574E with is 8 output.
cascade them like Roland did, probably a bit faster than bit shifting, almost instantly, 3 clock cycles. with 3 x 74hc174 you need 6 data lines and 8 with the CD74HC374.
You need a diode to 5 volt to limit on the trigger input before heading to the transistors or havok will occur.
@@MeeBilt You could remove Q80, Q81 and the stuff around it from the Hi hat circuit, move R480, R475 between Q85 and the inverted opamp input of IC 67B, copy or move the 1.2nF and 100pF on the same spot as on the Q80, apply a static positive offset voltage of 2.5V on the NON inverting pin of IC 67B, play with value R483, but you have to play with 1.2nF and 100pF, witch is a type of anti aliasing filter, till it sounds right. You safe 9 parts or so.
You could also do what i tell, but double IC 67B and have 2 anti aliasing filters, like before.
@@AnalogDude_ Thanks for the suggestions. If I decide to make a product that I can sell I will definately revisit the design and have a look at the overall BOM and maybe add some features. I'm also considering to redesign the trigger circuit (maybe remove the button completely).
But I will complete all 909 voices before I do that, so I can consolidate the BOM and learn more about which part of the design that I can reuse between different modules.
@@MeeBilt I'm making a pcb with the 74hc174 part and the (roland) resistors, i did a simulation and also with a normal R2R array, than i added the "VCA", opamp and Q80, Q81. using a plain R2R array avoids special values of resistors and does the same job, but takes up more room, unless you use resistor array's.
For the VCA part you could use a LM13700 instead of Q85.
it's probably best to use a AT27c512 for the cymbal and crash and use 2 x 74HC174, but that requires a MCU and a few lines of code to jump between the addresses of the crash and cymbal in between (mixing), like done in the tr-707.
dropping the button is a good idea, but i guess before that you should make a Sequencer.