Author |
Message |
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Fri Aug 24, 2007 10:41 pm Post subject:
|
|
|
Very cool samples - I especially liked the effects you got in vibscan. _________________ My Site |
|
Back to top
|
|
|
andrewF
Joined: Dec 29, 2006 Posts: 1176 Location: australia
Audio files: 4
|
Posted: Mon Aug 27, 2007 8:37 am Post subject:
|
|
|
Hi Ian
I'm up to the wiring stage for this cct (found a 6 pole switch), but wondering about using the buffer transistors on the LM13700s to drive LEDs.
I'm not very experienced with these chips but it looks like one could connect the buffer-in (pins7/10) to the amp outputs(pins 5/12) then buffer-out (pins 8/9) off to the LED's anodes via resistors.
would there need to be resistors between amp output and buffer-in?
Are these buffers independent from the op-amps?
cheers |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Mon Aug 27, 2007 10:52 am Post subject:
|
|
|
andrewF wrote: | I'm up to the wiring stage for this cct (found a 6 pole switch), but wondering about using the buffer transistors on the LM13700s to drive LEDs.
I'm not very experienced with these chips but it looks like one could connect the buffer-in (pins7/10) to the amp outputs(pins 5/12) then buffer-out (pins 8/9) off to the LED's anodes via resistors.
would there need to be resistors between amp output and buffer-in?
Are these buffers independent from the op-amps? |
Hi Andrew --
I'm really glad to see someone going ahead with building this module.
Your idea of using the buffers to drive LEDs is great! However, you should not connect them to the OTA outputs. The reason is that the buffers do not have a very high impedance, and they will rob current from the integrator. I played with the 13600 when it first came out and was never able to get it to work well using the buffer.
The good news is that the buffers are indeed independent of the OTA. All they are is a pair of Darlington-connected transistors, with the collectors connected to the (+) rail. So the way to do what you want would be to drive the buffers from the integrator outputs, ie, in parallel to the circuit outputs. I would suggest a protection resistor from the opamp outputs to the buffer inputs (anything from 22k to 100k would make sense), and then from the buffer outputs to the LEDs through whatever series resistance you need to turn on the LEDs.
According to the data sheet, the buffers will drive up to 20mA, but a footnote says to be careful not to exceed the chip dissipation limit. So you may want to work below 10mA.
Let us know how this works!
Ian |
|
Back to top
|
|
|
andrewF
Joined: Dec 29, 2006 Posts: 1176 Location: australia
Audio files: 4
|
Posted: Mon Aug 27, 2007 6:15 pm Post subject:
|
|
|
Quote: | I'm really glad to see someone going ahead with building this module. |
Always a pleasure!
I'm using TL074's for the op-amps (simply because I have dozens of them from dead mixers), which will leave me with 2 spare op-amps to play with. I was considering wiring up a simple mixer but do you have any other mods/extensions that would be worth trying?
cheers |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Tue Aug 28, 2007 12:15 pm Post subject:
|
|
|
andrewF wrote: | I'm using TL074's for the op-amps (simply because I have dozens of them from dead mixers), which will leave me with 2 spare op-amps to play with. I was considering wiring up a simple mixer but do you have any other mods/extensions that would be worth trying? |
No, Andrew, I don't have any special suggestions. But you can never have enough mixers, that's for sure.
I'll be interested to find out how the circuit works with TL074's. You may run into limitations at the low frequency end due to having more bias current and offsets. But it's a good place to start.
Ian |
|
Back to top
|
|
|
andrewF
Joined: Dec 29, 2006 Posts: 1176 Location: australia
Audio files: 4
|
Posted: Tue Sep 11, 2007 2:02 am Post subject:
|
|
|
Finished this circuit today, took me a week to find the jumper i'd forgotten to connect.... now its time to think about a panel for it.
Haven't had a big play around with it yet. I prefer to get things working and leave the fine-tuning 'til it is panelled and easier to jack signals into it.
Nevertheless it is capable of big sssssslooooooooooow sweeps - something like Klaus Schulze with narcolepsy
(it does fast ones too)
and with 6 or 8 phases to use, thats a lot of spacy twirling
i need a vc panner.
Thanks Ian, keep 'em coming!!! |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Tue Sep 11, 2007 9:28 am Post subject:
|
|
|
andrewF wrote: | Finished this circuit today, took me a week to find the jumper i'd forgotten to connect.... |
Congratulations! This isn't exactly a beginner project.
My version will run down to .02 Hz, but the lowest two octaves don't seem all that useful.
Can't wait to see how you do the panel.
Ian |
|
Back to top
|
|
|
tonyc
Joined: Oct 11, 2007 Posts: 16 Location: london
|
Posted: Wed Nov 07, 2007 7:08 am Post subject:
|
|
|
Hello
Thanks for you great circuit.
I am think of using the first design [SIN + COS] as a simple
Sine VCO.
Do you think it would be stable enough [with exp conv+tempco] ?
Is it possible to make even more simple without COS output, or
is this part of the circuit essential for its operation ?
Sorry for dumb questions, I am still learning !
-TC |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Wed Nov 07, 2007 9:17 am Post subject:
|
|
|
tonyc wrote: |
I am think of using the first design [SIN + COS] as a simple
Sine VCO.
Do you think it would be stable enough [with exp conv+tempco] ?
Is it possible to make even more simple without COS output, or
is this part of the circuit essential for its operation ? |
Tony -- The stability will depend a lot on the stability of your converter. But there is also the fact to consider that the OTA inputs produce additional drift. So maybe not as stable as non-OTA designs.
The two sections work together, so you can't just eliminate one. However if you do not need the quadrature feature, you could just make a good triange-core VCO and follow it with a tri-sin waveshaper.
Ian |
|
Back to top
|
|
|
tonyc
Joined: Oct 11, 2007 Posts: 16 Location: london
|
Posted: Mon Nov 12, 2007 4:50 am Post subject:
|
|
|
Hello
Thanks for the info.
Oh how bad is the drift ?
Although I have not seen any/many Tri core or sine VCO designs
that do not use OTA.
So perhaps not too bad . . .
I have also tried the Tri core idea with a Sine shaper, but looking on a spectrum analysis, the your sine VCO produces far cleaner and acurate sine output than even quite a good shaper circuit.
thanks again.
-tony |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Mon Nov 12, 2007 10:28 am Post subject:
|
|
|
tonyc wrote: | Oh how bad is the drift ?
Although I have not seen any/many Tri core or sine VCO designs
that do not use OTA.
I have also tried the Tri core idea with a Sine shaper, but looking on a spectrum analysis, the your sine VCO produces far cleaner and acurate sine output than even quite a good shaper circuit. |
Hi Tony -- My super tri VCO has a non-OTA core:
http://home.comcast.net/~ijfritz/sy_cir4.htm
An OTA shaper with feedback works quite well. I got harmonics more than 45 dB down with one version.
I haven't looked at the drift of the quadrature unit, since I plan to use it at low frequencies only. It could possibly be improved by biasing the linearization diodes on, since they compensate the tempco of the input transistors.
Ian |
|
Back to top
|
|
|
tonyc
Joined: Oct 11, 2007 Posts: 16 Location: london
|
Posted: Wed Nov 14, 2007 3:51 am Post subject:
|
|
|
Hello
Thanks.
I am liking the simplicity of the Quad VCO, so I will try that one.
Not sure I unsderstand too well about how to bias the diodes.
Do you mean the Zener diodes in the VCO circuit, or something
in the Expo converter ?
thanks
-tony |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Wed Nov 14, 2007 7:02 am Post subject:
|
|
|
tonyc wrote: | Not sure I unsderstand too well about how to bias the diodes.
Do you mean the Zener diodes in the VCO circuit, or something
in the Expo converter ? |
The linearizing diodes in the OTA chips. Unfortunately, I'm not sure how biasing these on would change the behavior of the oscillations. I'd suggest getting the circuit working as drawn and evaluating the drift to see if it is too bad for your needs. From there we could look at the linearizing diodes.
Ian |
|
Back to top
|
|
|
tonyc
Joined: Oct 11, 2007 Posts: 16 Location: london
|
Posted: Thu Nov 15, 2007 3:26 pm Post subject:
|
|
|
Hello
Ok thanks for the tip.
I will test the first circuit [sin+cos], and see.
I have breadboared already, and seems pretty good, but
does "drift" a bit, but I will work harder on the expo first.
thanks
-tony |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Thu Nov 15, 2007 5:07 pm Post subject:
|
|
|
tonyc wrote: | I have breadboared already, and seems pretty good, but does "drift" a bit, but I will work harder on the expo first. |
Woah, running already
Could you try something for me? If it's not a hassle, could you try connecting pins 2 and 15 to the (+) power suply with 22k resistors? This will turn the linearizing diodes on, and I would like to know if the system still oscillates. My board is waiting for panel completion, so I can't check this easily right now.
Thanks!
Ian |
|
Back to top
|
|
|
Nosferatu
Joined: Jul 27, 2007 Posts: 234 Location: Planet Rock.
|
Posted: Fri Nov 16, 2007 12:32 pm Post subject:
|
|
|
Quote: | pins 2 and 15 to the (+) power suply with 22k resistors? |
With quadrature VCF's based on LM13700 doing that they stop oscillating
i suggest imbalance caused by the diodes and resistors However this is
slightly different designs so it might just alter the shape. |
|
Back to top
|
|
|
tonyc
Joined: Oct 11, 2007 Posts: 16 Location: london
|
Posted: Sat Nov 17, 2007 8:41 am Post subject:
|
|
|
Hello
Thanks for the tip.
I tried the 22k's but they did'nt have much effect on
the pitch stability, but lowered the frequency a bit.
BTW The sine purity stayed the same.
The drift seems to occur more at higher end frequencies, or
perhaps just more apparent with my testing methods [low-tech].
When the CV is stepped to a new freq. the VCO responds quickly,
but then takes a few seconds to "settle down".
-tony |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Sat Nov 17, 2007 9:50 am Post subject:
|
|
|
Hi Tony --
Thanks for trying the extra resistors and for the report.
It looks like the drift is something that needs to be looked at more closely. I'd need to know more about your converter to go forward with this.
As far as settling -- well the loop response is very slow as a consequence of the small feedback needed to get good purity. You may have noticed that it takes several hundred cycles to start up! But once it is in quadrature it changes frequency very rapidly, as far as I noticed. Again, I was mainly interested in sub-audio (CV) frequencies, so I didn't look too closely at stability.
Ian |
|
Back to top
|
|
|
tonyc
Joined: Oct 11, 2007 Posts: 16 Location: london
|
Posted: Mon Nov 19, 2007 9:29 am Post subject:
|
|
|
Hello
I am using roughly the same expo circuit as in your [quad circuit] but with
an A1928 dual PNP transistor instead.
But obviously the VCO bit is the simpler [sin+cos] circuit.
I have tested the expo circuit with the Thomas Henry VCO1, which
seems to work pretty well.
I feels like the "settling" of the feedback you mention is perhaps the cause
of the small settling drift.
Thanks
-tony |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Mon Nov 19, 2007 9:58 pm Post subject:
|
|
|
tonyc wrote: | I am using roughly the same expo circuit as in your [quad circuit] but with
an A1928 dual PNP transistor instead.
But obviously the VCO bit is the simpler [sin+cos] circuit.
I have tested the expo circuit with the Thomas Henry VCO1, which
seems to work pretty well.
I feels like the "settling" of the feedback you mention is perhaps the cause
of the small settling drift. |
I also saw the settling drift, once Tony mentioned it. It turned out to be a problem in the converter, not the signal loop. I managed to get it pretty much fixed tonight.
The circuit needs 27k resistors in the lines feeding the OTA bias-current pins (pins 1 and 16). The circuit now tracks reasonably well and is pretty drift free up to an input bias current of .5mA. This agrees with previous work I have done with CA3080-based VCOs. The frequency of the 6-phase oscillator at this max bias is around 1500Hz. I checked the tracking from 5.0 Hz up to 1200 Hz, or about 8 octaves. That's about all you can expect with an OTA-based circuit. (Again the system runs to very low frequencies, 0.05 Hz or lower, but I didn't try to measure the tracking down there.)
Again, this circuit was developed for slow CV applications. For full-range audio the integration caps need to be reduced to something like 120 pF (for the 2-phase version).
Thanks, Tony, for putting this design through the wringer.
Ian |
|
Back to top
|
|
|
UofMEShop
Joined: Jan 06, 2009 Posts: 5 Location: MI USA
|
Posted: Tue Jan 06, 2009 2:53 pm Post subject:
|
|
|
For the VC Quadrature Oscillator, I am observing clipping of the U2a "x" output, the one that connects to the Zener diodes. It is frequency dependent. Below about 10 Hz there is no clipping. Between 10 and 250 Hz the positive peaks clips. Then up to the limit of 4 kHz both positive and negative peaks clip. The clipped peak's amplitude (11.0 V pk-pk at 4 kHz) is less than the U2b "y" output (11.6 V). Actually, as frequency decreases, the unclipped amplitude decreases until at the low (< 10 Hz) frequencies it becomes equal to the clipped. This is using an OPA2277 op amp. An OPA627 or LT1462 clips even more severely. I do not have access at the moment to the OPA2234. Might that help ?
Connecting one or both of the diode bias (pins 2 and 15) to +V through various value resistors affected this clipping and could induce clipping on the U2b "y" output. |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Tue Jan 06, 2009 3:13 pm Post subject:
|
|
|
Hmmm ... The back-to-back Zeners limit the amplitude to ~5V. I can't understand what you are seeing. Sorry.
Ian |
|
Back to top
|
|
|
UofMEShop
Joined: Jan 06, 2009 Posts: 5 Location: MI USA
|
Posted: Tue Jan 06, 2009 8:51 pm Post subject:
|
|
|
Should the Zeners be limiting the output to 5 V pk-pk or to +/- 5 V (= 10 V pk-pk which would be close to the 11.6 V pk-pk I'm observing) ? |
|
Back to top
|
|
|
frijitz
Joined: May 04, 2007 Posts: 1734 Location: NM USA
Audio files: 54
|
Posted: Tue Jan 06, 2009 10:03 pm Post subject:
|
|
|
UofMEShop wrote: | Should the Zeners be limiting the output to 5 V pk-pk or to +/- 5 V (= 10 V pk-pk which would be close to the 11.6 V pk-pk I'm observing) ? |
Ooops -- misread that. Yes, 10V pp or a bit more is what you should see. The 2277 has a fairly low slew rate, but it should be OK at low frequency. With a TL072 the circuit worked OK except at the very low end.
What kind of caps are you using? You need a low leakage type, either polystyrene or C0G/NP0 ceramic. Other ceramic or mylar (PE) will be too leaky. Since you are loosing amplitude at lo freq, that would be my first guess. (Also, if you use polystyrene you have to be careful not to damage it with heat.)
Ian |
|
Back to top
|
|
|
UofMEShop
Joined: Jan 06, 2009 Posts: 5 Location: MI USA
|
Posted: Wed Jan 07, 2009 7:05 am Post subject:
|
|
|
Changing from polypropylene to either polystyrene or silver mica still produces clipping. Z5R or Y5P ceramic stopped it from oscillating at all.
Can you comment on this LM13700 quadrature sine/cosine VCO that swaps its phases for negative control voltages ("positive and negative frequencies") ? It has a wide range (0 to 10 kHz for 0 to 10 control V) and -40dB distortion.
www.edn.com/filtered/pdfs/contents/images/341451.pdf |
|
Back to top
|
|
|
|