hello there:
i´ve seen the schem of the mxr envelope filter and saw that there was a 4066 used as variable resistors..but the detector circuit looks really complicated. has anyone tried a 4066 as a variable resistor with a simpler circuit? what are the conditions?
i think there´s a good possibility of fet-trimpot-free phasers there.
I don't know about "simpler," but yeah, see nelson's recent thread (http://www.diystompboxes.com/smfforum/index.php?topic=76747.0) and R.G.'s ASMOP (http://www.geofex.com/Article_Folders/ASMOP/ASMOP.htm) article for examples. I have failed to get a PWM phaser working in the past, but I came at it from a sufficiently ignorant position that it wasn't surprising :) I concluded that I would need a scope and some more quality time with textbooks to get it working.
Edit: for posterity, my post 555, and in a PWM thread!
Hi
the 4066 is suitable, but you'll need self-biasing or hand tuning to get it operating with a usable "slope" rather than as a switch.
Why not use a 4049 or 4069? I think they'll be a lot less trouble.
A while ago I designed a compressor that used a discrete MOSFET as a variable resistor. It was connected from the input to ground and the gate was simply the signal (rectified, amplified). The greater the input signal, the greater the voltage on the gate, and the greater the signal loss to ground. It was dead simple and worked fine.
cheers
The use of a 4066 to simulate a variable resistor is different than the use of invertor sections as variable resistors. Essentially, the 4066 (or 4016) switch section is switched on and off, either at a different rate, or for different durations, to produce either an average capacitance or an average resistance.
inverter as variable resistor? please tell me how!?
Well, for starters, consider the limiter and phaser projects that were published in Electronics Today so many years ago. Both used the mosfet transistors in a 4049 as variable resistors to ground. In fact, so did the first-issue Bad Stone, except it used a 4009 hex invertor.
These are not, however, the most noise-free possibilities, which is why one tends not to see them too often. In the case of phasers using this approach, the input signal has to be significantly attenuated (much like it would with a 3080, I suppose) to avoid clipping the invertor secton/s, and then recovery-gain is added, which simply amplifies whatever hiss is acquired through the circuit. Some time back, Mike Irwin built the ETI phaser and wrote to me that adding regeneration to it made the noise level intolerable.
I designed a phaser using a CD4069 as resistors to ground. It worked OK, but as Mark says, it can't handle most signals without distorting, so I had to reduce first and amplify later = hiss. A compander would have been a better solution. The good part is that you get 6 more or less matched variable resistors.
Regards!
Miguel
hmm.. how exactly you vary the resistance of the inverter? varying the supply voltage?
i mean, the idea is that the resistance between the in and out of the inverter varies with V+?
i can see it clearly in a 4007 ic, but not in a 4069..it´s the V+ the control voltage?
Here's the ETI Phaser: http://www.synthdiy.com/files/2004/eti447.pdf
The LFO output is fed to the input pins of the 6 invertors, and their output pins are treated as source, with the drains going to ground.
It is a little tricky. You use the output of the inverter as the input of the resistance (in a phaser, this would go to the + input of the opamp in place of the usual FET drain). The input of the inverter receives the control signal. The GND pin of the chip goes to your Vref (in a phaser, where the FET sources go), that's why this only works as resistor to ground (O Vref), and each inverter shares the "ground". The Vcc of the chip should not be connected to anything. Basically you are using the lower MOSFET of each inverter.
At least that's how I used it, having looked at several circuits using this kind of arrangement.
Good luck!
Miguel
My variable resistor works pretty well. It has a bunch of extra options (linear/exo FM, sync)
http://www.electro-music.com/forum/viewtopic.php?p=214528&highlight=#214528 (http://www.electro-music.com/forum/viewtopic.php?p=214528&highlight=#214528)
You can use any type of electronic switch. I prefer 4016 over 4066 as the range is greater. The 4053 and similar are fun to use because they have an inhibit control. In case you were going to use it with a single 9V refer to my other Lunetta VCO that uses just a current sink to control the vco.
You can also use 4007 gates as voltage controlled variable resistors. Look at my resonant VCF schematic and you can get an idea of how I use them.
http://www.electro-music.com/forum/viewtopic.php?p=214528&highlight=#214528 (http://www.electro-music.com/forum/viewtopic.php?p=214528&highlight=#214528)
I tried using 4049 and 4069 as variable resistors and they were pretty limited. Lots of CV bleed through and they need voltage protection.
Nice to see you over here synthmonger :)
I once used a 4066 as a variable resistor (PWM controlled) in a compressor design.
Some things I remember about it:
- I used two switches in a series-shunt ("L") arrangement: when the series switch was on, the shunt switch was off, and vice-versa. You need two control signals (out of phase) for that.
- If the power supply was not *very* good, weird noises would show up in the audio.
- The middle of the control range was linear, but not the extremes of the range (close to 0% and 100%).
Quote from: Mark Hammer on June 03, 2009, 09:01:39 AM
The use of a 4066 to simulate a variable resistor is different than the use of invertor sections as variable resistors. Essentially, the 4066 (or 4016) switch section is switched on and off, either at a different rate, or for different durations, to produce either an average capacitance or an average resistance.
My take on a Voltage Controlled Resistor (Switched Resistor) using PulseWidthModulation goes s.th. like this:
PWM01 (http://img.photobucket.com/albums/v437/latronax/2009_6/PWM01.jpg)
PWM02 (http://img.photobucket.com/albums/v437/latronax/2009_6/PWM02.jpg)
PWM03 (http://img.photobucket.com/albums/v437/latronax/2009_6/PWM03.jpg)
(to be explained in detail, later...)
hmm interesting..
the resulting resistance is a function on the relation between on and off pulses duration
so the pwm has to be at a really high freq?
ok, i read some more..looks like the pwm with 4066 is a little bit complicated..
in your opinion, what´s better? 4069 as variable resistors or pwm driven 4066? i mean in parts count, noise, and range?
It seems we're getting a few PWM posts recently. Here's one started by Nelson:
http://www.diystompboxes.com/smfforum/index.php?topic=76747.0 ...as mentioned earlier.
And another about the ETI phaser just got bumped recently by Puretube.
I came up with a discrete-component PWM modulator to drive the 4016 Control gates:
If using MOSFET IC's, you might as well use something like the 4016, that's designed to be used as a chopper modulator.
If just using the FET chips with linear control, then I suppose the inverter IC's are better hacks.
http://www.geocities.com/transmogrifox/PWM.jpg
(http://www.geocities.com/transmogrifox/PWM.jpg)
Quote from: Transmogrifox on June 11, 2009, 01:23:10 AM
It seems we're getting a few PWM posts recently...
Another simple (http://www.diystompboxes.com/smfforum/index.php?topic=77045.0) one... :icon_wink:
Let's one thing straight. All these above circuits including switched 4066 can only replace a pot to ground (or virtual groundZ) configuration, right? They cannot be used as rheostats.
Quote from: igor12 on June 11, 2009, 10:42:19 AM
Let's one thing straight. All these above circuits including switched 4066 can only replace a pot to ground (or virtual groundZ) configuration, right? They cannot be used as rheostats.
The simple PWM generator/oscillator (http://www.diystompboxes.com/smfforum/index.php?topic=77045.0) above (reply#17)
will turn an analogue switch of the 4066 type into the behaviour of a variable resistor, that can be "floating" indeed,
as depicted in the links of above post (http://www.diystompboxes.com/smfforum/index.php?topic=76880.msg631089#msg631089)
(reply#13).
When both outputs of the oscillators are used to control 2 switches, one resistor will increase,
while the other will decrease proportionally.
Those 2 switches can be combined to behave like a pot...
(to be explained later).
Hi
to control the resistance of a CD4049 or CD4069, you simply need to vary the gate over a small range around the "ON" voltage, which is usually close to Vdd/2. For a 9.X volt supply (fresh battery), a 10k/(10k+/- a few ohms) voltage divider should work well.
A while ago I designed a compressor/limiter (sounds similar to the one described by Mark Hammer, above). I was a 386 power op-amp that has a MOSFET to ground in front of it (drain to op-amp input, source to ground). The gate was connected to the output via an envelope detector which turned the MOSFET on slightly. The envelope detector was (in order from the output) a rectifier (1 diode), and a cap (22uF) and bleed resistor (?470k) between the control line and ground. It worked quite well. It biases it self so that the MOSFET is acting as a variable resistor. Too much input and output, and the resistance of the MOSFET goes down, bleeding the signal to ground. When the signal gets bled beyond a certain amount, the output falls and the MOSFET resistance rises and the signal is bled less.
Here's an old thread that discusses it. Unfortunately the link is dead. It'll see if I can find it somewhere over the next few days.
http://www.diystompboxes.com/smfforum/index.php?topic=41768.0 (http://www.diystompboxes.com/smfforum/index.php?topic=41768.0)
cheers
Hi again
here's the original MOSFET-as-a-resistor compressor schematic (thanks to puretube for sending me this copy :)):
(http://www.aronnelson.com/gallery/main.php?g2_view=core.DownloadItem&g2_itemId=39931&g2_serialNumber=1)
Note that there are many improvements that you could make to this. It is intended as a starting point.
In the context of this discussion, I'm fairly sure that you could use 1 MOSFET or inverter (e.g. from a CD4049) as an input buffer, another as the amp (in place of the LM386), another as the variable resistor (in place of the MOSFET), and another as an output buffer.
cheers
I spent a good 10 minutes tonight using some 4066 switches as my VCOVR's in a phaser circuit. I gotta say, I wasn't that big of a fan of phasers until now! It sounds very smooth and worked splendidly well. It's still in it's early stages but I do plan on making a modular version and maybe a simple one for a stomp box. Schematic to follow when I get my other work done :icon_frown:
I'll get around to testing the PWM method versus mine. I do use a PWM circuit for my aliaser pedal with some fun results. Adds a lil' extra timbre! :)
Thanks all!!
i think i now understand a lot more. In fact, i´m designing my own PWM (as simple as simple can), and planning to design a 4049 phaser
Quote from: synthmonger on June 15, 2009, 05:59:56 AM
I spent a good 10 minutes tonight using some 4066 switches as my VCOVR's in a phaser circuit. I gotta say, I wasn't that big of a fan of phasers until now! It sounds very smooth and worked splendidly well. It's still in it's early stages but I do plan on making a modular version and maybe a simple one for a stomp box. Schematic to follow when I get my other work done :icon_frown:
I'll get around to testing the PWM method versus mine. I do use a PWM circuit for my aliaser pedal with some fun results. Adds a lil' extra timbre! :)
I look forward to it. IMHO, one of the properties that makes a phaser pleasing is a "graceful turnaround". That is, how it negotiates the extremes of the sweep. That factor is the principle reason one strives to match JFETs in a phaser: so that all control elements are changing their resistance at the same rate/time. With switched-resistor arrangements, the matching is much easier to achieve.
that 386 compressor looks nice.. why did you use a 386 and not an opamp like TL072?
Why a 386?
Because someone was raving about them at the time (??Charlie aka Moosa??), and they are a single supply device. Today, I would go for a TL071 set up for high input impedance (probably non-inverting, with a 1 M bias resistor, and gain of around 100 (?). Then 0.05Vp-p = potentially 5Vp-p, but half that (2.5Vp-p) would be just starting to turn on the MOSFET and divert the input. So any signal over 0.05V would be compressed). Compressing when the MOSFET resistance falls below 1 Mohm is a much more respectible figure than the 50k set by the input resistance of the 386.
Actually, with a couple of (or three) feedback diodes, you could have a combined distortion unit and compressor in one. Maybe I'll think about that a bit more......
Thanks for the interest in this simple concept.