Author |
Message |
LetterBeacon
Joined: Mar 18, 2008 Posts: 454 Location: London, UK
|
Posted: Wed Oct 07, 2009 3:18 pm Post subject:
|
|
|
Thanks for the reply!
Just to be 100% clear -I should be able to hard wire this to the CV Out of a sequencer PCB and, with the Glide pot fully CCW, it would be as if there's no Glide circuit in there at all.
Sorry to state the obvious but I want to be absolutely sure as I'm designing the front panel at the moment. |
|
Back to top
|
|
|
j.dilisio
Joined: May 19, 2009 Posts: 200 Location: baltimore
|
Posted: Wed Oct 07, 2009 4:38 pm Post subject:
|
|
|
Thats how it works over here. You could try it first before commiting.
And you can always substitute a different circuit later if you're not happy. _________________ DRONEGOAT |
|
Back to top
|
|
|
Mooger5
Joined: May 02, 2007 Posts: 199 Location: Portugal
Audio files: 8
|
Posted: Wed Oct 07, 2009 7:01 pm Post subject:
|
|
|
That´s logarithmic portamento, right? I found this circuit online for linear. Didn´t try it yet.
|
|
Back to top
|
|
|
Tim Servo
Joined: Jul 16, 2006 Posts: 924 Location: Silicon Valley
Audio files: 11
|
Posted: Thu Oct 08, 2009 2:48 pm Post subject:
|
|
|
Hey Mooger,
Yes, that circuit should produce a linear glide instead of expo. It comes from putting the cap in the feedback loop of the output buffer op-amp. A triangle core VCO uses the same arrangement to get linear slopes. I've also done the same thing with envelope gens to get linear slopes instead of the more common exponential slopes (linear slopes work nicely with expo VCAs, while expo slopes are typically used with linear VCAs). You'll probably need to use a bigger timing cap with the linear setup, as the timing won't be the same.
Tim (hittin' the slopes) Servo |
|
Back to top
|
|
|
Adam-V
Joined: Jan 29, 2007 Posts: 300 Location: Australia
Audio files: 1
|
Posted: Thu Oct 08, 2009 8:55 pm Post subject:
|
|
|
Mmmm trapezoid generator!
Cheers,
Adam-V _________________ Digitalis Effect | Fractured Symmetry (www.spiralsect.com) |
|
Back to top
|
|
|
Mooger5
Joined: May 02, 2007 Posts: 199 Location: Portugal
Audio files: 8
|
Posted: Thu Oct 08, 2009 10:39 pm Post subject:
|
|
|
Tim Servo wrote: | Hey Mooger,
Yes, that circuit should produce a linear glide instead of expo. It comes from putting the cap in the feedback loop of the output buffer op-amp. A triangle core VCO uses the same arrangement to get linear slopes. I've also done the same thing with envelope gens to get linear slopes instead of the more common exponential slopes (linear slopes work nicely with expo VCAs, while expo slopes are typically used with linear VCAs). You'll probably need to use a bigger timing cap with the linear setup, as the timing won't be the same.
Tim (hittin' the slopes) Servo |
Hey Tim,
Yeah if this circuit has a different configuration and is labeled Linear, then the other had to be Exponential:) What would happen if one of the buffers were in opposite phase to the other, btw?
In practical terms, do you know if the expo portamento time in the first circuit starts fast and slows down increasingly while it approaches the destination note, or the other way around? I imagine the former to be more familiar sounding, like the trombone, fretless bass or Ondes Martenot player gliding between say C2 and C3 moves initially fast by instinct until slows it down to find the right pitch; all in a fraction of a second, but still noticeable. (I´m not saying synthesizers should sound like this or that, btw).
With linear it´s always the same time constant, right? Which one would take the same time from C1 to C5 as it would take from C4 to C5?
Sorry all if this sounds off-topic!
Regards |
|
Back to top
|
|
|
Mooger5
Joined: May 02, 2007 Posts: 199 Location: Portugal
Audio files: 8
|
Posted: Fri Oct 09, 2009 3:00 pm Post subject:
|
|
|
I think I grok the diode for the optional asymmetric portamento: as it is, it´d bypass the resistor at positive voltage changes, so the circuit would perform only "downward" portamento and "upward" by reversing the diode?
If that´s so, could it also work as saw/ramp waveshaper? |
|
Back to top
|
|
|
Tim Servo
Joined: Jul 16, 2006 Posts: 924 Location: Silicon Valley
Audio files: 11
|
Posted: Sat Oct 10, 2009 12:02 pm Post subject:
|
|
|
Hi Mooger,
The glide in the expo circuit should start quick and then slow as it gets closer to its 'destination' voltage. The output will look a lot like the typical envelope shapes you see (kinda like a shark's fin).
OTOH, with linear glide you get a constant RATE, but not a constant time. In other words, linear glide doesn't start quick and then slow down. A linear glide from C1 to C5 will still take longer than a linear glide from C4 to C5. Getting a glide that was a constant TIME would be extremely tricky. You'd have to have a VC glide and some circuitry that would compare the start and end voltages and produce a CV that was proportional to the difference. It could be done, but this is the sort of thing that would be much easier with a digital synth or microprocessor-based circuitry. If one of the buffers was out of phase (non-inverting, while the other was inverting), I don't think the circuit would work. The feedback going from the second op amp back to the first is what keeps the output stable once the output voltage = the input. Might be worth playing with if you have this up on a breadboard. You might want to put a small value resistor (<1k) after the glide time pot. This acts as a current limit. Look up some other glide circuits on the web, or check out Barry Klein's "Electronic Music Circuits" book for more info:
http://members.cox.net/barryklein/em.htm
BTW, You might want to verify this, but I recall the Minimoog uses a linear glide, while the ARP Odyssey uses an expo glide.
And yeah, on the Asymmetrical glide, the trick with the "steering diode" is the basis of most of the Tri/Saw variable waveshape oscillators you see around. You could use two diodes and put a glide pot in series with each one to get separate up and down controls. Now you've got a circuit that is roughly the same as an AR envelope gen. It's funny how many synth circuits depend on the basic lin or expo integrator circuit (glide, envelope gens, LFOs, even VCOs). Check out Ken Stone's "Utility LFO" for an LFO example:
http://www.cgs.synth.net/modules/cgs58_lfo.html
Tim (integrating the flux capacitors) Servo |
|
Back to top
|
|
|
Mooger5
Joined: May 02, 2007 Posts: 199 Location: Portugal
Audio files: 8
|
Posted: Sat Oct 10, 2009 5:45 pm Post subject:
|
|
|
Thank you for taking the time to reply, Tim.
Looks like I mixed up the nomenclatures while reading elsewhere about the lin and expo glides properties!
So the expo simple design is somewhat compromised by the non-linear response of the cap and the lin compensates it with some sort of "negative feedback" loop...
I´ll experiment with the following:
a) a positive feedback loop; b) a "two-pole" expo; c) audio through the lin; just to see what the heck comes out of it
You´re right about the Mini and the Odyssey. It´s common knowleadge, I think. Been checking a few schematics and the Ody is definitely expo. The Mini features a 1uF grounded cap, but there´s two jfets for up and down glide; too complex for me to understand.
The best news is the expo glide in my Rogue is very much like this simple one. There is no first buffer, but the second is a single dedicated IC. Hoping it´ll sound more "Minish", it´ll be a matter of replacing the single IC with a dual, cutting a few tracks and using the existing components to turn into a lin glide
Thanks for the Klein book tip.
Quote: | It's funny how many synth circuits depend on the basic lin or expo integrator circuit (glide, envelope gens, LFOs, even VCOs). Check out Ken Stone's "Utility LFO" for an LFO example:
http://www.cgs.synth.net/modules/cgs58_lfo.html |
There´s a thread about a Universal Synth Module in the Developer´s Corner I think. Something like a Modular Module. I find it a great idea, even if only for educational purposes.
One again, thanks very much for the enlightenment
Regards |
|
Back to top
|
|
|
Mooger5
Joined: May 02, 2007 Posts: 199 Location: Portugal
Audio files: 8
|
Posted: Sat Oct 10, 2009 7:33 pm Post subject:
|
|
|
Tim Servo wrote: |
And yeah, on the Asymmetrical glide, the trick with the "steering diode" is the basis of most of the Tri/Saw variable waveshape oscillators you see around. You could use two diodes and put a glide pot in series with each one to get separate up and down controls. Now you've got a circuit that is roughly the same as an AR envelope gen. |
I see now Nicolas´ LFO skew control uses the same principle. Fascinating stuff. |
|
Back to top
|
|
|
Spacebones
Joined: Jul 12, 2011 Posts: 15 Location: Utah
Audio files: 2
|
Posted: Wed Jan 30, 2013 7:12 pm Post subject:
|
|
|
I'm having an issue with this glide for some reason. When I feed it square waves, it doesn't seem to work at all... I set it up such that the output of a 40106 oscillator is connected to the input of the glide circuit, and the output of the glide is connected to an led so that I could see visually if the glide is working. When I give it power, the led just stays lit, and doesn't oscillate between lit and off.
However, I tried putting the 40106 output into the first input of a little R2R DAC, and fed the output of that into the glide input. At this point, the led modulates like I had hoped it would with the oscillator alone. What gives? I'm using a TL072 in the glide if that makes a difference.
Any insight would be excellent!
[EDIT] Nevermind, apparently putting some resistance (1 meg) between the oscillator and the input did the trick... Can anyone explain to me why that is? |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
Posted: Mon Nov 18, 2019 9:32 am Post subject:
|
|
|
Hey there, I have built various versions of this classic glide circuit and I always have the same issue: at say 4V-5V DC (a few octaves above middle C for most synths), 1V/octave VCOs go slightly flat.
I figure there is some lag/slop in the cap? So it never gets back up to source voltage exactly?
OK this is Not a big deal if all VCO's are tracking the same CV (although things might sound a few cents out of tune for high notes)
But the other day I compared VCOs with this slew in parallel with others that got straight V/oct. if you have one VCO tracking slew and others not, for me in 2 circuit variations the slop is noticeable.
Does anyone else have this issue?
It started to bug me recently because I can't figure out a good way to fix this....smaller cap with bigger pot maybe? I would think some kind of high frequency additive compensation would be really complex.
Do most synths have this problem with VCO slew and I have never noticed? _________________ Visit my AUDIODIWHY blog and website |
|
Back to top
|
|
|
ixtern
Joined: Jun 25, 2018 Posts: 145 Location: Poland
|
Posted: Tue Nov 19, 2019 5:23 am Post subject:
|
|
|
Please post exact schematic of your circuit. Without it I can only guess.
1. Don't use electrolytic capacitors as they have some leakage current. Use film caps.
2. For such circuits try to use JFET-input OpAms with low input current, low offset voltage and low temperature drift. If TL072 - then TL072B. TL052 are little better, LT1057 also. Pay attention to PCB design (guard ring near cap and OpAmp input) and PCB cleaning (no after-soldering fluss).
3. Don't use output buffer 1kOhm resistors outside feedback loop! |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
Posted: Tue Nov 19, 2019 8:34 am Post subject:
|
|
|
ixtern wrote: | Please post exact schematic of your circuit. Without it I can only guess.
1. Don't use electrolytic capacitors as they have some leakage current. Use film caps.
2. For such circuits try to use JFET-input OpAms with low input current, low offset voltage and low temperature drift. If TL072 - then TL072B. TL052 are little better, LT1057 also. Pay attention to PCB design (guard ring near cap and OpAmp input) and PCB cleaning (no after-soldering fluss).
3. Don't use output buffer 1kOhm resistors outside feedback loop! |
1. Yes, thought of that. electrolytics kinda suck and here's an example? I might have some axial 2.2uF poly film 50V, but a lot of these I already have are like 630V and huge (used for tube amps) and those are too big to be practical.
Also I might try 1uF cap metal-film and a 2M AT "donpechi'd" pot (dual gang 1M ATaper wired as a rheostat). Can't see why taht wouldn't work.
2. I am Using TL072 but not sure of the flavor. From the TI spec sheets of all the different flavors I am reading input impedance (is that why you suggest?) is 10^12 ohm (i.e., "forget about it") so I am curious why you recommend one over the other--seriously I want to learn more about this--I have always considered the TL07xx family to be interchangeable at the low frequencies we work with.
Good suggestion on flux removal. Haven't done that.
3. Ah, that makes sense. I had to think about this for a bit! The output resistor protects the IC but also makes the 100K 1% resistor at VCO input (that most VCOs are going to have--mine all do anyway ) into a 101K resistor which changes input current and knocks things out of tune. Interesting! Ha! I;ll bet that's the issue! if I removed that I'd have to be careful about shorting the glide circuit output to ground, but I'll try it--be careful patching but have your instrument stay in tune, right? Good suggestion!
(For the attached sch: pot would be Audio taper 1M, wired to "POT1" wiper to "POT4". SW is optional, a SPST switch to bypass the entire glide path if wanted. I lifted this straight out of electronotes but may have added the 1K resistor?
Thanks!
Description: |
|
Filesize: |
26.65 KB |
Viewed: |
632 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
_________________ Visit my AUDIODIWHY blog and website |
|
Back to top
|
|
|
ixtern
Joined: Jun 25, 2018 Posts: 145 Location: Poland
|
Posted: Tue Nov 19, 2019 12:01 pm Post subject:
|
|
|
R3 1k resistor for sure is wrong connected. If you really want it as output protection, connect inverting input (pin 2) after the R3 resistor, not before. This way accuracy will be maintained. Only output voltage swing may be little reduced but should be enough. If not then change 1k to 220R (still inside the feedback loop).
Input 100k resistor is also too small if driven not directly from OpAmp output but e.g. through 1k resistor like in your circuit. Try bigger value (1MOhm or more) although it may then become sensitive to the electric noise if connected by longer wires or so.
While I don't like TL072 in control voltage circuits? Because it may have large offset voltage, up to 5 mV, and 5 mV is a serious hearable error in pitch. And you have two such Op Amps (offset voltages may add or cancel or mix). Of course it may not if you have a luck. TL052 (newer version of TL072) or LT1057 have lower offset voltage and lower temp drift.
Input impedance is not important, input bias current is important and what I also dont like about TL072 and any other cheap JFET OpAmp is that input current is unpredictable as it heavily (exponentially) depends on temperature.
Think of electrolytic capacitor and OpAmp input as resistors. If capacitor has at 1V 10 nA leakage then you may think of it as 100 MOhm resistor. You are loosing 1% of accuracy (for 1MOhm pot).
If OpAmp has 1nA input current for 1V (not uncommon for TL072 at high temperatures) you can think of it as 1000MOhm resistor - you are loosing 1 mV for it. Not big but adds to other errors.
There are very small 2uF MLCC ceramic capacitors you may use (although I prefer foil ones). |
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
|
Back to top
|
|
|
ixtern
Joined: Jun 25, 2018 Posts: 145 Location: Poland
|
|
Back to top
|
|
|
ixtern
Joined: Jun 25, 2018 Posts: 145 Location: Poland
|
|
Back to top
|
|
|
cslammy
Joined: Apr 27, 2018 Posts: 206 Location: USA
Audio files: 1
|
|
Back to top
|
|
|
|