so Antoine are you saying the 100k effects it detrimentally? this to be honest would explain the squaring off at very high signal. which I think is a good aspect to this vca.
Yeah, assuming ideal circuits it doesnāt do anything (the amplifier tries to output a current, not a voltage), but @antoine.pasde2 points out that with 100k you hit the voltage drive capacity limits before any other limit.
in an ideal world Antoine and Fredrik what would you like to see on this? I have a version im sitting in front of which is a 4x4 cv matrix, and a larger 10x4 cv matrix. is there anything you would suggest id adjusted in the circuit? the zener diodes went a while back. I have since updated them, and said on the site not to plop them in.
im gunna have a look again on breadboard, back to the data sheet and see if it saturates, maybe the 100k is the good stuff. I dont understand why it would be useful to have hotter signals its very useful for things like distortions that donāt have clipping circuits not blowing your head off. but like I say in the above comment Iāll re visit it before I put these vca matrixās up. Just out of interest.
The problem (*) with synth stuff is that you cannot treat it as audio gear, so that āsurely this is wrongā reflex isnāt always helping Iād say if you like what youāre hearing with 100k, keep it, but with a note that builders can use a lower value or a link instead. Or the other way around; design with a link (or small resistor), but mention the option to use a 100k to get a tiny bit of sweet clipping.
*) not a problem, really
would be useful to have hotter signals
Which reminds me ā is there a Kosmo definition of āhotā? @antoine.pasde2 wrote ā-10V to +10V typicalā (20 Vpp) in the draft here:
but seems you may be designing for closer to 10 Vpp in practice, maybe? (also see the many discussions about the 3340 output levels).
(I guess we should all be responsible audio adults and talk dBs instead, but I find that a bit confusing tbh and not at all helpful when looking at circuitsā¦)
ok fredrik cool! its great that your challenging these things, thankyou. the thing is im going into this treating it as audio gear, the main focus point is the end product coming out of the speakers. Which yes to get to that point many philosophies and approaches can be takenā¦ The thing that attracted me to synthesisers is there is no right way. even the most absolute wrong way can exhibit interesting outcomes. and yes when designing these things I tend to go from what it should be and take away or add till it seems right to me. Which im sure to some people may not seem correct, like this character. This module is meant for one of the last in the line, the closest to the speaker.
but yeah the 100k ill play with later on and then do a note to it! im going to do a vid on the more serious channel about all these things regarding this synth! so ill talk about it ill adjust this and see what changes it represents practically in the vid. as im doing the vid to show the mod of bypassing the normalising of the jacks.
thanks. please keep criticising and analysing these circuits.
well this is all very up in the air. Antoine will be please to know I listened about his comments on the oscillators. all future oscillators will have an adjustable offset for unipolar and bipolar moving forwards! however yeah its seeming 10vpp is the aim. 20vpp is quite high is it? the 10vpp leaves space for things like feedback dynamics, and distortion to take it the rest of the way. and go into clipping territory nice and early on. which in the end of the day gives an analog synth its goodness.
however 20vpp would make just as much sense, as the 1222ās are unipolar, you put a bipolar oscillator in with a unipolar oscillator, it would end up being 15vpp, the great thing about this is if you clip the merge of these 2 oscillators after you get very uneven clipping which sounds rather gnarly in a good way!. so yes I think the standard may be subject to evolve. I have found distorting oscillators can be quite difficult to get a nice character, and if they are offset and unsymmetrical with each other its much easier to get that frazzle like in a guitar. I have a digital vco coming that has an offset so the waveform can be bipolar or anything inbetween, and after that a dual vco that is doing what the fart box does
Oh, itās definitely āaudio gearā in that sense ā I was more thinking of the Hi-Fi āthere can be no distortion anywhereā ideal (which is usually just āI like this distortion better than that distortionā anyway, no different from here). If it sounds right, itās probably ok
thats funny about hifi. āI like this distortion more than that distortionā. I always questioned āhifiā im glad you said that. glad thats out in the open!!!
I havenāt done it yet, but would bypassing the first 100k require the removal of the 100k feedback resistor? does this not mess with the gain of the circuit? not had a chance to mess with it as of yet.
No, that one controls the output level. The configuration here is a current to voltage converter (basically, the VCA outputs a current, the current cannot go anywhere other than through the 100k, which results in a voltage across the resistor, the opamp attempts to keep the āinput at 0 V since thatās where the + input is, so it pulls the output low, and you end up with Vout = āIinĆ100k)
(what Antoine pointed out that I missed in my initial analysis is that with enough resistance on the way to the converter the VCA cannot produce the output current itād like to; the output resistor gets in the way.)
The alternative is to shunt the voltage to ground through a resistor and get a voltage that way, as itās done in the datasheet, and add a voltage follower after that, but Iād say this is more elegant
ok that explains the dynamic clipping! which is fine good for this module but yes I will for the matrix mixer which is equally for voltagesā¦ Quite excited about the matrix module, as I have made it in such a way thats its modular within itself, each row is a separate board, so you can make a matrix module as big or as small as required, then I have 2 different panels 1 that houses 4 columns or one that houses 10 columns. but because its in a setup that suits the mixing of control voltages too. I will revisit the above.
aaah! yeah nice! they working good???
Works perfect in isolation, and in small configs. Still missing some modules for my larger (still rather vague) plans so remains to be seen if this is a good idea in the long run, but as Iāve mentioned earlier arbitrary and rather silly constraints is a huge part of the fun for me (*). Iām the first to admit that theyāre not all the most practical way to do things
*) main constraint being āhow can I organize this kind of circuit so the important bits fit on this tiny PCB without going all SMD, and how can I make the circuit and the PCB more generally useful while doing thatā; Iām especially fond of PCBs that can be used in multiple ways, or for that matter orientations
Im just glad someone went farther than i did with the modular modular idea. Especially when its someone that can actually see it through n stuff.
OK, it looks like @fredrik and you sorted it all out while I was sleeping. Good.
The 100k resistor could cause clipping for high level signals and high gains, but if you like it that way in this design, leave it like that, it can easily be replaced for those that donāt want the clipping.
If 10V peak-to-peak is to be the recommended ānormalā signal level (which makes sense to allow for headroom etc), Iāll add that to the Kosmo Specification Wiki
Glad to see that we agree that the 2V Zener is not necessary on the signal input.
Given the voltage divider before the linear control voltage input, a 2V Zener diode is not necessary there either.
While analyzing the circuit to determine if a 2V protection Zener diode was necessary on the CV input, I noticed that the switch labeled āINVā doesnāt actually invert the effect of the control voltage as I expected it would, but it just adds an offset equivalent to about 5.6V on the CV input jack so it will just increase the output volume (and can be used with a negative going envelope to cause a drop in volume, but the envelope would need to have been inverted before being applied to this moduleās CV input).
What is the expected voltage range out of an envelope generator?
Or said differently, what is the expected CV input range for this module for a normal envelope?
What does the #1157 Mini ADSR put out?
What do Eurorack envelope generator modules typically put out?
the inv switch yes doesnāt technically invert its function, inv probably is a tiny bit misleading, its function. well its a couple of functions.
the module this module was replacing in my synth was the module at the bottom right of the synth cabinet at the end of 3 voices. 1 of those voices comes out of 1 of the vcaās and into another one of the vcaās with the inv usually turned on, this offsets it to like you say 5.6. the reason 5.6 well because 220k was a common round about value. It serves 2 purposes practically. in this mode an inverted envelope generator acts as a side chain going into it. but also going into it is an accent envelope generator too.
the other purpose is originally In the old vca mixer it was set to full power, in a gig in order to figure out where a problem is in the synth voice live I would need to bypass the vca, initially when I flicked the inv switch on this it would go full volume and blow peoples socks off. so I changed it to about half power (originally 100k resistor, then halved it pretty much, funnily enough I seem to recall in a van in a driveway in Hamburg! but also to add with the 100k at the time it was making it louder than what the envelope generator was powering it so it was making it loud loud.
the 1157 puts out 10v. Eurorack I have seen some put out 10, some out more than that! most less. so rather underwhelming ones closer to 5v. its all over the shop. the 1157 has an inverted output too.
usually going into the vca with inv mode in my instance there are two envelope generators going into a mixer, one inverted and one none inverted, one for side chain wired to a sum of the kick and snare input, and also an accent trigger for a bit more volume whilst keeping under 10v.
then the vca before that vca is wired with 1 envelope generator
as for the other 2 voices going into that mixer vca, thanks to the normalled jacks, are isolated from the above 2 vcaās and these two voices go to the mix out. (well not anymore actually they both go off separately now as its evolving)
I hope that makes sense. its a sort of middle ground. to allow for sidechaining and accenting. and also trouble shooting on completely closed VCAās without blowing peoples socks off.
in the dual VCA this is just replaced with an offset knob, instead of a switch.
I did have a plan to do a video on these VCAās when I put them out but one thing lead to the other and it didnāt happen. then it sort of slipped my mind, however because of a mod I spoke about further up the thread (for bypassing the normalling of the jacks) ill be doing the video this week hopefully. Iāll talk about all the stuff we have been speaking about too. and possible mods. including increasing this inv/offset/bypass switch if people feel its not setup for their needs.