Joined: Jan 29, 2009 Posts: 120 Location: Fremantle, Australia
Posted: Wed Jul 28, 2010 3:01 am Post subject:
A couple more notes of possible interest.
Put in the Fairchild 4046 and the signals are much closer to +-5V than the NXP. Still a little high at +- 5.2V but really pretty close. I don't think that the Triangle is as clean for the Fairchild but the Sawtooth was better than the NXP !?! Will try and take some photographs of the output to describe what I am burbling on about...
While reading the Birthofasynth web page for X4046 again and more carefully, I noticed the link to the Scott Gravenhost harmonic multiplier. After a bit of a read through that and some further hacking away at the board I now have my second sync mode doing some pretty interesting things! I ended up changing the filter a fair bit and also the way I divided the output from the phase discriminator was a bit odd. I need to do some more head scratching before I am happy with it as a useful option...
Joined: Jan 29, 2009 Posts: 120 Location: Fremantle, Australia
Posted: Wed Aug 04, 2010 9:42 pm Post subject:
Done some testing of the tracking of some of the 4046 devices. I only have a Fairchild, NXP and TI parts. I used a Ruby script driving a MIDI interface and collecting the serial feed from a frequency counter. The results were then plugged into Excel. I took the natural log of the frequency and then did a linear fit over a reasonable useful range (the green bit) of about 50Hz to 1.5kHz. I then found the deviation from the best fit to give me some idea of the tracking... In an ideal world I would try and adjust for the best fit and then measure but I was to lazy to adjust each individually... Remember that the graph is logarithmic!
The TI part really doesn't cut the mustard... I did use this exact TI part in a very similar circuit that Tom designed for the VC Noise in the Synthesizer for the 21st Century and in that design it seemed to track fine? I will have to drag the board back out and check it again now...
see: http://forbinthesynthesizer.blogspot.com/2010/01/more-noise.html
The NXP and Fairchild Parts are very similar but I have found that the voltage that the NXP develops across the capacitor is quite a bit higher than the Fairchild so I have stuck to that part.
One of the datasheets that I found for the TI 4046 said that the design was from Harris (and so I suppose originally RCA). I am not sure how much individual variation there is some of the internal designs between manufacturers?
4046_vco_comparison.gif
Description:
Filesize:
63.55 KB
Viewed:
1133 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.
Great work Forbin!
That is definitely above and beyond the call of duty. _________________ What makes a space ours, is what we put there, and what we do there.
Joined: Jan 29, 2009 Posts: 120 Location: Fremantle, Australia
Posted: Sun Aug 08, 2010 2:46 am Post subject:
Thanks Guys -- i ran a test overnight of the Fairchild 4046 without a tempco and am currently running another one WITH a tempco -- i have always been interested in how bit a difference it is... will post the results tomorrow
Next I will run the other 8 or 9 VCO's through my measuring system, including the Thomas Henry XR, VCO1 and 21st Century... I have sped up the averaging a bit and brought back the steps to 3 semitones so it should be quite interesting...
Remember that a tempco adjusts the whole V->I converter depending on ambient temperature changes so it would be more meaningful to check the tracking of a particular note for various temperatures, rather than the linearity across a selection of notes. _________________ What makes a space ours, is what we put there, and what we do there.
Joined: Jan 29, 2009 Posts: 120 Location: Fremantle, Australia
Posted: Sun Aug 08, 2010 6:26 am Post subject:
The "result" that I am measuring against, is the slope of the best fit through the readings 50Hz -> 1.5kHz -- if it is perfectly in pitch the value would be 0.057762 in my configuration... It is probably better than watching just one value as it is more a representation of the "tuning" of the oscillator
Joined: Jan 29, 2009 Posts: 120 Location: Fremantle, Australia
Posted: Mon Aug 09, 2010 3:05 am Post subject:
This is probably getting a bit out of hand now but I am seeing ripple in the result when I increased the number of steps 0->10V. This was then a lot of testing of the MIDI interface and changing of DAC's -- this was improved things a bit...
vco_ripple.gif
Description:
Filesize:
27.09 KB
Viewed:
934 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.
The switching glitches on the peaks of the triangle waveform (and the derived sawtooth and sine waves) look a bit ugly but, as Thomas Henry points out, sound acceptable musically. They introduce a lot of rich high harmonics and could change the sound considerably compared with a pure triangle wave. Not a bad thing though; we all want harmonics, don’t we?
Unfortunately, the spikes do limit the VCO’s uses if it is used as an LFO to voltage control another device.
The good news is that there is a simple fix. The spikes are on the ‘outside’ of the waveform. As long as the waveform amplitude (and DC offset) remains constant* over the frequency range, a simple clipping or clamping buffer could be used to slice them off, improving the waveforms.
Even simple diode clamping to the power supply rails might** work, if they are clean enough and the output of IC3c waveform swings close enough to the power rails. The resistor values around IC3c and IC3d may need revising to achieve this, and the input levels into the sine converter may also need adjusting as a result as well.
*As I have yet to build this VCO, I do not know if the DC level and amplitude are in fact constant over the frequency range on this design. I would presume so.
** a few decoupling capacitors on the power supplies physically near the clamping diodes might be a good idea if this is done, just to keep the supplies clean.
A minor point: on the circuit schematic, R37 and R39 are labelled ‘Triangle Connect’ . Should this be ‘Triangle Correct’, or even better, ‘Triangle Adjust’?
thanks for the ideas on the clamping, don. maybe i will try that out.
i assembled one of the two 4046 VCOs modules for my next cabinet today. i used my prototype PCB (the same i sent to scott) and an additional waveshaper PCB i designed - and everything worked from the start.
i tried an ON Semiconductor MC14046BCP first and could not get it scaled correctly. it fell short of about 1/4 tone per octave.
i then tried the NXP part and scaled the VCO within 5 minutes using an LFO and a quantizer.
i will post a picture when i finished both VCOs (they share one frontpanel). _________________
cheers,
matthias
____________
Big Boss at fonitronik
Tech Buddy at Random*Source
The switching glitches on the peaks of the triangle waveform (and the derived sawtooth and sine waves) look a bit ugly but, as Thomas Henry points out, sound acceptable musically.
I was looking at the various waveform pics of the X-4046 VCO that Scott Stites has posted on his website. The pic labeled X-4046 VCO Sawtooth Wave looks a lot like the sawtooth waveform produced by the Yamaha CS-80 as well as other CS synths that use the same VCO IC. (The pic can be seen at http://www.birthofasynth.com/Thomas_Henry/Images/X-4046/sawtoothl.JPG)
If my memory serves me correctly, the glitches found in the Yamaha Sawtooth Waves are even a bit wider than the glitches shown in Scott's X-4046 VCO Sawtooth Pic. Nevertheless, the sawtooth waves produced by the Yamaha CS Series synths sound incredible and if you don't believe me listen to the fat brass horn sounds that were produced on the CS-80 in Vangelis' Chariots of Fire. (I should point out that much of this fatness is due to the fact that the CS-80 uses linear tuned VCOs.)
Don Erskine wrote:
They introduce a lot of rich high harmonics and could change the sound considerably compared with a pure triangle wave. Not a bad thing though; we all want harmonics, don’t we?
That depends upon what type of sound you need. A pure sine wave is 100% fundamental with no harmonics. I generally like my sine waves as pure as possible and use them for whistle-type sounds, soft leads, and to boost the fundamental on low bass sounds (the kind of bass that can be felt more than heard).
If you add enough odd harmonics to a sine wave you basically have a triangle wave, which sounds a lot like a heavily filtered square wave. However, I wouldn't mind if the X-4046 VCO was capable of adding some even harmonics to the sine wave. (In fact, the shape pot for VCO 1 on the EMS VCS3 allows the user to add even harmonics to the sine wave.)
I am a bit curious about what kind of effect the peaks will have on the triangle waves. Boosting some of the triangle wave's odd harmonics will probably have little effect on the sound other than making it sound a bit more square wavish, but if the glitch adds "even harmonics" the sound of the triangle could change quite a bit.
Since I already have several VCO's that produce accurate waveforms, I'm glad that there are some less perfect alternatives. Otherwise, many VCO's would tend to sound the same and I like having the variety available to me.
Joined: Jan 06, 2011 Posts: 18 Location: Lithuania
Posted: Wed Jan 19, 2011 2:06 pm Post subject:
Guys, I'm confused. I made PCB by myself, panel, everything as always. But I'm stuck on wiring. I see two J1, two J2, in schematic is writen to connect volt/octave to J1, but there's two of them on PCB, and also 1v/octave is writen there... I'm posting some pictures how my works look like. I hope you will help me. I don't want burn everything after such a nice working.
Posted: Tue Jan 25, 2011 11:41 pm Post subject:
Noob question
Hi there
I am newbie (both in terms of this thread and electronics/electronic music).
I have a dumb question (but if you don't ask you don't learn!) so please forgive this.
I am interested in this design to add some more wave forms onto stomp box producing a square wave output using the hard sync (I know I know it's sacrilege to ignore the wonderful 1 volt per octave bit ) Could that part of the cicuit work on different voltages (+- 12, +- 9) rather than +-15? or would I have to significantly redesign?
I made up the board following the Press 'n' Peel layout supplied in this forum - I assumed the orientation based on the mirrored text, however creating this board "as-is" resulted in all the power connections for the op-amps being reversed.
I'm not sure if reversing the design will fix it - but I got frustrated enough to transcribe the circuit in Eagle and layout my own board.
Pleased to say that using this, some TL082/84s, and cheap (unmatched) 3904 transistors from Radio Shack I got all the waves coming out as specified.
Nice.
Will have to post on my blog about this...
Steve
www.stephenhobley.com
For all your Laser Harp, Tesla Coil, and Killer Robot from Outer Space needs...
Joined: Jan 06, 2011 Posts: 18 Location: Lithuania
Posted: Fri Mar 04, 2011 2:08 am Post subject:
I spend so much money on this project... yes, there's something wrong with sine section.. no signals from tl074. And something wrong with course, i get from very high frequency to lower, and then it starts getting up again when turning course pot.
Can somebody help with this vco. A lot of people have troubles buuilding it
Joined: Jan 06, 2011 Posts: 18 Location: Lithuania
Posted: Fri Mar 04, 2011 2:08 am Post subject:
I spend so much money on this project... yes, there's something wrong with sine section.. no signals from tl074. And something wrong with course, i get from very high frequency to lower, and then it starts getting up again when turning course pot.
Can somebody help with this vco. A lot of people have troubles buuilding it
i built two units using my own PCBs. don't have any issues, the circuit is proven to be fine.
so someone who built it with ryk's layout should chime in, if the common debug routines don't help you. _________________
cheers,
matthias
____________
Big Boss at fonitronik
Tech Buddy at Random*Source
Ppl. my mate who made several of these noticed an error in the layout that ryk did made its around the sine circuit,can't remember what exactly.
TekniK wrote:
Its not a error that influence the working of the module,its a design error.
so it is an error in the layout, not the circuit, correct?
i built this VCO according thomas' schematics without any issues... or are you talking about the sine shape only? it might not be a perfect sine, but still quite nice (i actually like the "colored" sine). _________________
cheers,
matthias
____________
Big Boss at fonitronik
Tech Buddy at Random*Source
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You cannot attach files in this forum You can download files in this forum
Forum index » DIY Hardware and Software » Thomas Henry designs
FAQ
Calendar
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
Chat Room
) Could that part of the cicuit work on different voltages (+- 12, +- 9) rather than +-15? or would I have to significantly redesign?