Author |
Message |
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
|
Back to top
|
|
|
toppobrillo
Joined: Dec 10, 2005 Posts: 766 Location: oakland, ca
G2 patch files: 1
|
Posted: Fri Feb 17, 2006 9:30 pm Post subject:
|
|
|
Quote: | The other one would be an enormous diode/DIL switch matrix which lets you force the wave shape. |
or a ROM? |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Sat Feb 18, 2006 1:31 am Post subject:
|
|
|
Yep,
maybe even a ROM. I do want to keep the basic circuit easily put together with minimum knowledge/money. But yes, the potential for expansion is definately there. _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Sat Feb 18, 2006 3:51 am Post subject:
|
|
|
This is what I'm trying to work out now. In the attached pic the blue lines represent the times that the signal is moving towards one of the voltage levels on the divider (ie, one or more switch in the 4066 is closed) Where the blue line stops, the signal starts to move towards 0V.
My question is why are there gaps between those times? If there weren't, the wave would look a lot more like the red line, which is what I'm really chasing.
First I thought that the 4066 takes longer to connect a switch than to open it. But according to the data sheet the propagation delay for 9V operation is about 65ns for both low-high and high-low control signals, which is way to fast to be seen at audio frequencies (~15Mhz?)
So the next suspect (should have been the first) is the speed of the diodes. (And as you said Jan, their associated capacitance) I've looked at the 1N4148s data sheet but it only quotes the reverse recovery time, which is down around 4ns anyway.
I'll keep experimenting, and let you know if I find an answer. I'm going to start a stripboard layout of the DIL switch version so I can play with other waveshapes too.
Description: |
|
Filesize: |
25.11 KB |
Viewed: |
353 Time(s) |
This image has been reduced to fit the page. Click on it to enlarge. |
|
_________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004 Posts: 24075 Location: The Netherlands, Enschede
Audio files: 277
G2 patch files: 320
|
Posted: Sat Feb 18, 2006 5:54 am Post subject:
|
|
|
The scoop shots have been puzzling me as well . ... it's as if the DAC is only driven half the time. From what I've seen in the schematics sofar I've no explanation for it.
How do you control the switches in the diode matrix, could there be something wrong with that circuit ?
What speed are you running the main oscillator on ? _________________ Jan
also .. could someone please turn down the thermostat a bit.
|
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Sat Feb 18, 2006 8:01 am Post subject:
|
|
|
I may be way off base here, but....
Is you Winscope performed with your sound card? In that case, I imagine it may be AC coupled, and it would not 'hold' a DC level, but move back towards 0V once the level became stationary.
The best way to trouble shoot it would be to pulse the input slowly, step by step, observing what comes out of the DAC.
Is it possbile to post a schem of how you have things right now?
Cheerio,
Scott |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Sat Feb 18, 2006 8:25 am Post subject:
|
|
|
I'm still pretty useless trying to use Ray's schem publisher so I'll try to describe it. It's basically the same as the first drawing in this thread, except I've added a 220nF de-coupling cap followed by a 10K volume pot on the output, and there are 5K pull down resistors on each 4066 control line. The control lines are fed by 10*1N4148 diodes on various 4017 outputs to combinations of 4066 control lines.
Yes the winscope goes into the soundcard, but there's definately a point where the signal stops heading towards the expected output, and starts heading towards 0V. I'm sure that means the 4066 switches aren't staying on long enough.
If I crank up the main frequency the trace shows much straighter lines from one voltage level to the next.
Thanks for the ideas anyway, I'll keep on the case! _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Sat Feb 18, 2006 8:41 am Post subject:
|
|
|
Taking the de-coupling cap out flattens the tops a lot, but at low frequencies, it still just looks like a lot of different sized spikes. It's like the 4066 switches only stay on for a brief moment.
I'm starting to think you're right Scott. Some kind of capacitance effect is allowing the signal to bleed back to 0V. It doesn't get enough time to at higher frequencies. The question is, how do I "hold" it, at least down to say 200Hz? _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Sat Feb 18, 2006 8:53 am Post subject:
|
|
|
In this case, you might try avoiding any capacitance and probing on the DAC output itself without buffering. Isolate the DAC with a buffer and do all of that stuff downstream from the buffer. A TL07X set up as a voltage follower will work fine for that. Connect the output of the dac to that, and just do all of your coupling and probing on the output of the buffer.
Cheers,
Scott |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Sat Feb 18, 2006 3:40 pm Post subject:
|
|
|
The words "voltage follower" were actually rolling around and around in my head as I tried to get off to sleep last night!
Now how do I make a TL071 etc into a voltage follower with a single +9v supply again?
I know some things well, but there's other stuff which I don't.
How about I go in to the non-inverting input, and feed the output back into the inverting input? (maybe via a resistor?) Is that right? _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Macaba
Joined: Jul 13, 2005 Posts: 160 Location: UK
|
Posted: Sat Feb 18, 2006 4:32 pm Post subject:
|
|
|
Just a heads up, might not have any relevance, but hey...
Why do you have the diodes in there, out of interest? |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Sat Feb 18, 2006 4:36 pm Post subject:
|
|
|
So that each step of the 4017 only closes certain switches in the 4066. Without the diodes most patterns would wire back to close every switch on every step. _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Macaba
Joined: Jul 13, 2005 Posts: 160 Location: UK
|
Posted: Sat Feb 18, 2006 4:50 pm Post subject:
|
|
|
Ahhh, you diagram didn't show the same counter driving all 3 switches.
I may just have to breadboard this then. |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Sat Feb 18, 2006 9:05 pm Post subject:
|
|
|
I think, in the case of a non-inverting voltage follower, even if it's a single supply circuit you can still wire it up the same. Signal to non-inverting input, output connected to inverting input. Note: I am spoiled by dual supply designs - the Dim C is the only single supply thing I've done.
The 1N4148's should not be a concern here - this is the kind of circuit they love
Cheers,
Scott |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Wed Feb 22, 2006 6:27 pm Post subject:
|
|
|
I've been away for a couple of days, but I've been thinking lots!!
I reckon the bits where the signal moves towards 0V are points between steps where the last steps switches haven't dropped out yet. Almost every time this happens, there will be a small moment when too many switches are closed. (closer to 0V)
So I'm thinking that the capacitance of the diode matrix is holding the 4066 switches up just a bit too long. Now, could that be helped by putting pull down resistors on the 4017 side of the matrix (ie: 8 of them) instead of, or as well as on the 4066 side (ie: 4 of them)?
I'll get back home tonight and try it out. _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004 Posts: 24075 Location: The Netherlands, Enschede
Audio files: 277
G2 patch files: 320
|
Posted: Fri Feb 24, 2006 6:14 am Post subject:
|
|
|
That looks better indeed !
Glad you found it. _________________ Jan
also .. could someone please turn down the thermostat a bit.
|
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Fri Feb 24, 2006 11:16 am Post subject:
|
|
|
Yes, velly goood. |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Sat Feb 25, 2006 2:57 pm Post subject:
|
|
|
A couple of things,
First, it now seems to make no difference whether I listen to the signal pre or post TL071 voltage follower. Is there a good reason to leave it there?
Secondly, if I connected the outputs of 4 AND gates would they work as if they had been OR'd together or would the other three "sink" the high on the fourth?
And thirdly,.... umm, I forget what the third question was!! _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Sat Feb 25, 2006 3:43 pm Post subject:
|
|
|
Quote: | First, it now seems to make no difference whether I listen to the signal pre or post TL071 voltage follower. Is there a good reason to leave it there? |
If the output of the circuit is arranged so that plugging it into inputs of varying impedance won't affect the performance, I don't see why you would need the buffer if it sounds the same with or without it. If the buffer is providing that bit of isolation from outside world influence, might not hurt to leave it in.
Quote: | Secondly, if I connected the outputs of 4 AND gates would they work as if they had been OR'd together or would the other three "sink" the high on the fourth? |
If you're talking about connecting the outputs of logic gates directly together, that's generally verboten (and the reason tri-state logic was invented). Methinks you can OR them together with diodes.
Quote: | And thirdly,.... umm, I forget what the third question was!! |
42
Cheerios,
Scott |
|
Back to top
|
|
|
Macaba
Joined: Jul 13, 2005 Posts: 160 Location: UK
|
Posted: Sat Feb 25, 2006 4:58 pm Post subject:
|
|
|
Uncle Krunkus wrote: |
First, it now seems to make no difference whether I listen to the signal pre or post TL071 voltage follower. Is there a good reason to leave it there? |
The main reason that springs to mind is that if you connect to something that has capacitance, without the buffer, it could affect the circuit.
My discovery of this was when I built a wien bridge oscillator. When connect to oscilloscope (extremely high resistance) I got my nice sine wave. When i then connected to a frequency counter directly, it distorted waveform bad, hence the need then for a buffer.
EDIT: Just occured to me- just check the slew rates of the buffer op amp, it needs to be really good. I used an LM358 once, and actually was disgusted at the poor performance due to the sloppy slew rate. |
|
Back to top
|
|
|
Uncle Krunkus
Moderator
Joined: Jul 11, 2005 Posts: 4761 Location: Sydney, Australia
Audio files: 52
G2 patch files: 1
|
Posted: Sat Feb 25, 2006 5:52 pm Post subject:
|
|
|
Hey Scott,
Yeah, I had a feeling that might be the case. (with the gates)
I suppose I was hoping (beyond hope) that you might say "Oh yeah! that's fine, no worries!!"
Anyway, the reasoning behind it was that the final circuit I'm looking at, will now be a complete condominium of CMOS cockroaches!!
8 * sqr wave clocks - NE556 * 4
driving
8 * 4 bit binary counters - 4520 * 4
switching
32 * dual input ANDs - 4081 * 8
matrixed against
an 8 step sequence - 4017 & 555 * 1
and OR'd - 32 * 1N4148s
into a 4066
I'm thinking of calling it Frankenstein's Oscillator! _________________ What makes a space ours, is what we put there, and what we do there. |
|
Back to top
|
|
|
blue hell
Site Admin
Joined: Apr 03, 2004 Posts: 24075 Location: The Netherlands, Enschede
Audio files: 277
G2 patch files: 320
|
Posted: Sat Feb 25, 2006 6:02 pm Post subject:
|
|
|
Or simply : the Monster, We'll know what you mean
You could consider ... well see : http://electro-music.com/forum/post-68326.html _________________ Jan
also .. could someone please turn down the thermostat a bit.
|
|
Back to top
|
|
|
Scott Stites
Janitor
Joined: Dec 23, 2005 Posts: 4127 Location: Mount Hope, KS USA
Audio files: 96
|
Posted: Sat Feb 25, 2006 6:43 pm Post subject:
|
|
|
Hey Uncle Krunkenstein,
Fortunately diodes aren't too much of a hassle to put in there - just as easy as putting in a jumper, usually.
If you have any concern about it, you could try 7556's instead of the 556's - I love the 755X IC's - they operate on a mere fraction of the 55X IC's power requirements, and are sooo much friendlier to the power rails. Rene Schmitz turned me on to them with his VCO3 design.
Cheers,
Scott |
|
Back to top
|
|
|
|