DIY Synths General Discussion

mosc · Posted: Sun Apr 16, 2006 8:20 am Post subject:

Nice. I'd like to hear how your experiments with the tilt to MIDI interface goes.

blue hell · Posted: Sun Apr 16, 2006 4:45 pm Post subject:

Looks cool ! Doing with the tools you describe shows you've got good eyes and hands ! I'm getting too old for that, man I can barely see those 805 "dust grains" even :-)

BTW, what accellerometer did you use ? I'd like to see data sheets of it.

bigtex · Posted: Sun Apr 16, 2006 10:19 pm Post subject:

Macaba · Joined: Jul 13, 2005 Posts: 160 Location: UK

Macaba · Joined: Jul 13, 2005 Posts: 160 Location: UK

blue hell · Posted: Mon Apr 17, 2006 5:32 am Post subject:

Thx !

bigtex · Posted: Mon Apr 17, 2006 12:26 pm Post subject:

zipzap · Posted: Mon Apr 17, 2006 1:38 pm Post subject:

Jürgen Haible, that very extreme diy-guy has done some VCOs with linear detune. Read about it here
http://www.oldcrows.net/%7Ejhaible/jh5/jh5.html
I don´t know, but i think thats got to do with morphing special characteristics, going from log to lin or stuff like that.
When i went to school i should have learned about log and e and r2d2. I didn´t. What´s your idea behind this? whats the purpose?
Is it like some softsynths, where you can shape the responce (slope) of an adsr for example? That´d be cool.

bigtex · Posted: Mon Apr 17, 2006 1:42 pm Post subject:

I've got a few ideas, mostly involving waveshaping and nonlinear sequencing (sequences that change speed across a measure, like a bouncing ball sound, a very IDM kind of sound), but I just want to understand some basic bulding blocks before I go trying to design a circuit and get it all wrong.

zipzap · Posted: Mon Apr 17, 2006 2:27 pm Post subject:

sounds interesting. sequencers are some fascination. I´ve been designing a lot around delay circuits, for generating shuffle. That Pingpong thing could be nicely with a lm3914 sequencer i think (other thread).
Abot that amp, do i get this right: You take eg an tiangle wave, put it in the amp and get variable shaping of the rise and fall part, like linear or log curve?

bigtex · Posted: Mon Apr 17, 2006 2:32 pm Post subject:

zipzap · Posted: Mon Apr 17, 2006 2:50 pm Post subject:

Uncle Krunkus · Posted: Mon Apr 17, 2006 4:42 pm Post subject:

I might be completely off the track here, but as far as I remember, an exponential function is one which has an exponent, i.e. it is the function of one value raised to the power of the other. The relationship is not linear, as in a graph of it will not be a straight line. It will be a curve. So,
y=x^2
y=x^-2
y=log(x)
are all exponential functions.
So log is just a type of exponential function, of which there are many.

PS I'm not posting this to be a smart arse, I mainly want confirmation that it's right. If it's not I'm willing to learn. Smile

bigtex · Posted: Mon Apr 17, 2006 5:06 pm Post subject:

You know what? I'm not exactly sure... I don't really know the exact semantics of math, so I don't know if a log function is considered an exponential function. I don't think so, but I could easily be wrong (It's been a while!).

I guess it doesn't relaly matter what it's called..... just how it works, right? Smile

blue hell · Posted: Mon Apr 17, 2006 5:26 pm Post subject:

Uncle Krunkus · Posted: Mon Apr 17, 2006 5:39 pm Post subject:

Yeah,
I was gonna say that! Very Happy

(Thanks Jan)

blue hell · Posted: Mon Apr 17, 2006 6:00 pm Post subject:

Uncle Krunkus · Posted: Tue Apr 18, 2006 12:29 am Post subject:

ian-s · Posted: Tue Apr 18, 2006 1:07 am Post subject:

blue hell · Posted: Tue Apr 18, 2006 3:26 am Post subject:

Ah a speling error as well :-)

No I've never been waiting like that, must be stressfull I can imagine having everything ready and nothing happens knowing you might have to rush at about any time. Hope it will turn out well !

Uncle Krunkus · Posted: Tue Apr 18, 2006 5:03 am Post subject:

blue hell · Posted: Tue Apr 18, 2006 10:05 am Post subject:

zipzap · Posted: Fri Apr 21, 2006 3:12 am Post subject:

This is another small thing:
I was just playing with my sequencer and a filter. They use different PSUs.
The filter gets +15,0,-15v, the sequencer is running off a small 12v wallplug unit. The Filter can only respond to the sequencer if the two grounds are connected. If i compare the ground level between the two circuits with no connection i mesure about 3-6V. 5-20ma are running (the digital meter is not that fast...). thats 0,12w max.
Is that normal? Is that a problem (besides wasting the current)? Is this what a ground loop is?
The thing is also that i want to use an audio and a digital ground in the future. So then there should be no ground connection between the sequencer and the filter (only in the psu or the wall of my house). Why can´t the VCF respond to just the voltage coming from the seqencer? I mean it´s got it´s own ground to compare the voltage to. I don´t get that.

zipzap · Posted: Fri Apr 21, 2006 3:30 am Post subject:

Question 2:
In the shematic down there, can i just add more cv-ins via resistors to T16 (hard to read, it´s the one most bottom-left)?
Or should i use an summing OPamp?
thanks guys

moog.jpg

Description:

Filesize:

40.17 KB

Viewed:

286 Time(s)

This image has been reduced to fit the page. Click on it to enlarge.

Uncle Krunkus · Posted: Fri Apr 21, 2006 5:59 am Post subject:

I'm fairly sure that, to a point, you can just add CVs through resistors to T16. Have a look at the MFOS site and see how Ray adds CVs.
The VCF needs to see voltages which are relative to the 0V they were generated relative to. So you should have a ground connection between the sequencer and the VCF. If you got the ground through the PSU/Power Outlet chain then you would be building ground loops. The digital/audio ground is a per/unit concept designed to give the digital ground a faster short than through the analogue ground, so that the latter doesn't get contaminated. The ground to external units should come from the "star point" which is the junction between digital and analogue ground near the PSU. This sounds kind of weird, I know, because the first thought most people have is that if they are shorted together then there shouldn't be any difference between them. But it is based on the resistance of the actual ground connections (every wire has resistance) and the decoupling of different sections of the circuit, which gives the ground connection the ability to attenuate the high frequency transients on the digital ground. But only if they "see" the digital ground's smoothing caps first. i.e. if there was a faster way to ground through an analogue circuit on the same ground line, then these transients would be redirected through them. Thereby contaminating them with digital "spikes".
Does that make sense?
It's a bit esoteric, I know, but once you get your head around it, it makes sense.