Group Design: PWM Limiter Project

[SSLtech]

Switched capacitor filtering...

http://www.filter-solutions.com/switched.html

Keith

 

[Learner]

Hi Keith,
Just want to say thank you for sharing your idea with us, as much as I would like to take on the task of developing/building the compressor I simply don't have the technical experience and knowledge to do so. However, I am definitely very interested in participating in the development of the piece and would love to see it built one day! :thumb:

Having been looking at the schematic for the last couple of days and having read your circuit description I think I got the idea of how this thing work. just got a couple of question about the 555.........

In the schematic you have the 555 in parallel with the rectified DC signal going into the LM319 comparator, basically modulating the pules width of the 555. I was just wondering is it possible to skip the LM319 and just use the 555 as monostable?

The reason being in the monostable mode as I understand it,

Pin 5 (Control Voltage)"By applying a voltage to this pin, it is possible to vary the timing of the device independently of the RC network." As shown in the data sheet Fig.8 PWM.

Pin 2 (Trigger): "This pin is the input to the lower comparator and is used to set the latch, which in turn causes the output to go high. " Can we use this pin as a control for Attack of the transient, just by adding a potentiometer to set the input voltage?

Feeding the rectified DC voltage into Pin. 2 and Pin. 5

Pin 6 (Threshold): "Pin 6 is one input to the upper comparator (the other being pin 5) and is used to reset the latch, which causes the output to go low. "

Again, can we use this pin as Release by attatching an potentiometer from Vcc to gnd for adjusting the timing of the pulse to go low.

Eventhough Pin 4 (Reset) does the same but may be we can leave that connected to Vcc, that's all for now will be back with some more braoin diggin :green: :thumb:

oh, for the pin description/explaination I got it from here. Have yet to find a more shorter/simpler pin function description/explaination.....


http://www.uoguelph.ca/~antoon/gadgets/555/555.html

 

[Learner]

Ok, back with some more voodoooo ideas.......

Looking at the first stage lowpass decade EQ, do you think we could replace the 4 op amps with just an input differential amplifier followed by an unity gain low pass filter 3dB roll off at 20Khz with a buffer?

Also replacing the 4 pole filter with a state variable filter after the FET,
for individual Hi pass, bandpass and lowpass output plus a summing amp stage for all the filter outputs to be rectified.

Therefore, allowing us to select each filter output to be rectified or just rectify the summing signal for the feedback DC to control the FET. Not to mention, implementing adjustments for hi pass and lowpass filter roll off frequency(therefore the bandpass Q and center frequency) for rectification(before the gain stage "compression").

Just wanted to add that, ideas mentioned above are purely presented to explore the possibilities and options that might be available to us or what could be implemented. It is NOT in any way to suggest that they are an improvement to the circuit or the outcome, I hope you might find the ideas useful again may be not..........

I think this is a kickass project because it is yet to be defined, at the same time we(I mean I) can learn so much not only from the exploration of what we can do but from people that are a lot knowledgeble than myself!
:green:

Looking forward to hear some thoughts from all of y'all! :grin:

And let get start learning, building & listening!

Thanks again for the inspiration keith! I know how lonely it is to be working on a project all by yourself........:green:

:thumb:

:guinness:

 

[Svart]

learner, you are definately in the right place! I don't think the others are behind it but i'd rather start working on a design using a stand alone PWM, which would take the place of the comparator and the 555. this would simplify things down to a pure square output, not the half hearted triangle-ish output of the 555 circuit, leaving us to focus on controlling the beast. I do think that we could control it using a detector circuit (rms detector?) that feeds directly into either the INV or NI error amp inputs, attack and release being altered by RC networks worked into it.. what you think?

I do like the sg3525a since it has features like 500khz operation, softstart, deadtime control and 0-100% duty cycle(really only around 88%).

 

[SSLtech]

Do it man... do it!!! :green:

Learner, I'm not altering the output of the 555... I used the 555 as a (reasonably) constant frequency (but certainly easily built and cheap) triangle wave generator. You can substitute any triangle source at all.

The constant triangle wave is used as a comparison against a varying control voltage, to obtain a varying pulse width in proportion to the control voltage. The frequency doesn't change... the duty cycle varies completely from 0% (no attenuation) to 100% (full attenuation)

If you vary a control voltage (on pin 5? -from memory) on the 555, the duty cycle does indeed shift, but the frequency also tends to move aroound... That's what I remember when I built my little toy "stylophones" using the 555 as the monophonic oscillator... I think that I fed a triangle wave into pin 5 to provide a vibrato that also had an interesting duty cycle alteration/... soiunded like a cross between vibrato and chorusing in a funny kind of way... and easy & cheap to do.. but I digress...)

The 20kHz 3dB/8ve filter idea probably won't get the out-of-band harmonics down far enough. remember that aliasing starts to happen at half the chopping frequency, and the level of the combined artifact will be in proportion to the level of the offending signal component... that means tha tif you're working at a chopper freq. of 400kHz, you need the harmonics to be down to about -90dB at 200kHz... 3dB/8ve at 20kHz only attenuates the harmonics by around 10dB at 200kHz...

Keith

 

[Svart]

yeah the freq shift is another thing i hate about the 555 setups, of which i have tried many. the sg3525a is .70 cents from mouser.com... and it does away with the 555 AND the comparator since both a osc and the comp are built in! it stays rock solid frequency wise, unless of course you set it up to change.

 

[Learner]

[quote author="Svart"]learner, you are definately in the right place! I don't think the others are behind it but i'd rather start working on a design using a stand alone PWM, which would take the place of the comparator and the 555. this would simplify things down to a pure square output, not the half hearted triangle-ish output of the 555 circuit,

snip/............

[/quote]

hey Svart,
This is where I am getting confused, I thought the 555 puts out a square pulse in either monostable or astable mode?

http://www.tel.uva.es/~tri/tutorial.htm

http://www.doctronics.co.uk/pdf_files/ne555.pdf

 

[Learner]

[quote author="Svart"] I do think that we could control it using a detector circuit (rms detector?) that feeds directly into either the INV or NI error amp inputs,

[/quote]

We could probably get that by adding a logrithmatic amplifier after the rectifiying stage?



[quote author="Svart"] , attack and release being altered by RC networks worked into it.. what you think?

[/quote]



Yea, that's what i am trying to figure out......as the study continues on the understanding of the concept in envelope generators. Here is what I have came accross

http://www.uni-bonn.de/~uzs159/adsr2.png

http://www.midwest-analog.com/adsr1.html

http://www-timc.imag.fr/Yves.Usson/personnel/SDIY/archives/j-jacky-1980.pdf

http://www-timc.imag.fr/Yves.Usson/personnel/SDIY/archives/O-Gabor-1980.pdf

http://www.itn.liu.se/~nikro/diy/littleman/littleman.pdf

 

[Learner]

[quote author="SSLtech"]

Learner, I'm not altering the output of the 555... I used the 555 as a (reasonably) constant frequency (but certainly easily built and cheap) triangle wave generator. You can substitute any triangle source at all.

Keith[/quote]

right, may be something like this...

http://www.interq.or.jp/japan/se-inoue/e_ckt16.htm

In fact, it is being implemented here

http://www.cpemma.co.uk/pwm.html

However, I am really curious about implementation of the 555 as and envelope generator control. Need to wire it up when I get time, in the middle my HV regulated PSU at the moment.......need time.....

 

[Learner]

[quote author="SSLtech"]

If you vary a control voltage (on pin 5? -from memory) on the 555, the duty cycle does indeed shift, but the frequency also tends to move aroound... That's what I remember when I built my little toy "stylophones" using the 555 as the monophonic oscillator... I think that I fed a triangle wave into pin 5 to provide a vibrato that also had an interesting duty cycle alteration/... soiunded like a cross between vibrato and chorusing in a funny kind of way... and easy & cheap to do.. but I digress...)

Keith[/quote]

Yes, I was aware of the drifting factor when used in the astable mode. However, when use in the monostable mode it will only put out a pulse when you trigger it so I don't see how frequency drifting might be applied here?

Therefore the idea of implementing attack using Pin 2 (Trigger) to control the timing of the output pulse to go high and Pin 6 (Threshold) to control the timing of the output pulse to go low, and use Pin 5 (Control Voltage) to vary the pulse width with the rectified DC.

Thoughts?

 

[Mark Burnley]

Learner,

Check out the area on the schematic around the 555. You'll see that the TL072 buffer isn't connected to the output pin of the 555- it's a follower connected across the timing-capacitor. The 555 works by charging a cap through a series resistor, and monitoring the voltage across the cap. When it reaches 2/3 of Vs, the 555 then switches the cap from charge to discharge. The cycle then starts again.

This charging voltage across the cap is kept in the approximately linear part of the cap charge curve, so by picking off the voltage from across the cap with a high-impedance buffer, you get a good-enough triangle wave.

And yea, once we've got The Lab "Time Stretcher (TM)" project under way, with PCB's, we'll all be able to get on with a lot more DIY :wink:

:thumb:

Mark

 

[Svart]

learner, the output of the 555 is supposed to be a square, but in many testing rigs i've built(it's been years so i may be fuzzy..) it tends to be a triangle with it's top flattened(sorry for layman's terms..) the ramp up/down isn't very sharp.. this may be fine for audio as i am used to working with FETs and IGBTs that need "instant" on(full saturation)or off, with no signifigant ramp times(or they tend to burn out from the heat).

http://www.st.com/stonline/books/pdf/docs/4286.pdf

give that a once over, the test circuit works ok, however in that config the output is alternating 50% duty, if you use Vc as your output and leave the true outputs NC then you will get up to 100% duty. this works well unless you have to drive something inductive/capacitive then you need a driver that can sink more current(FET switch?). for that I use the tc4427 mosfet driver, which will sink/source 1A with perfect square output even at full loads.
It also has Osc output and sync input, so you can chain a few together.

ah only an idea.

 

[Svart]

everybody lost interest? damn.. oh well.

 

[Learner]

Nop! Still on the yet to do list, just have other priority at the moment eg. HV PSU and I think likewise with keith and everybody else(was there anyone else svart?)

Anyway, I think for this project to be going anywhere we need someone experienced in designing field that's willing to offer their guidance. Not to provide straight answers to our problem BUT dropping hints and coordinating workload and developing steps, in order to define the overall vision and objective of this piece.

Which leads us to our first problem, who has the experience and willing to offer the time for a no finacial return project? :?:

 

[Svart]

true.. maybe we can become a charity org? :green:
then we can give out tax writeoffs!

I'm still thinking about this thing though, I don't have much time other than checking the thread every few days but i should have some coming up where i can work on this thing.

how about suggestions for the different sections:
input preamp design?
output amp design(makeup?)?
detection circuitry(RMS detector like the DBX?) or something better like the ssl clone?
PWM design: to be determined after other sections.

I think simple 5532/5534 I/O like the ssl clone should suffice for the prototype.

at least that is my thought here.

 

[SSLtech]

[quote author="Learner"][quote author="SSLtech"]when use in the monostable mode it will only put out a pulse when you trigger it so I don't see how frequency drifting might be applied here?[/quote]
Well, since the pulse in monostable mode is a fixed length, the only way to vary the mark-to-space ratio would be to slow down or speed up the trigger rate... -or am I missing somthing obvious? (wouldn't be the first time!)

Keith

 

[Learner]

[quote author="SSLtech"]
Well, since the pulse in monostable mode is a fixed length, the only way to vary the mark-to-space ratio would be to slow down or speed up the trigger rate... -or am I missing somthing obvious? (wouldn't be the first time!)

Keith[/quote]

yes, I have been loosing sleep thinking about that one trying to understand the application of 555 as a envelope generator.

Remember we could use Pin 5 (Control Voltage)"By applying a voltage to this pin, it is possible to vary the timing of the device independently of the RC network." As shown in the data sheet Fig.8 PWM.

As the rectified voltage going into IC7a +ve terminal, we could vary the voltage of the -ve in order to vary the DC output voltage and feeding that into Pin 5 to change the PW.

While using Pin 2 (Trigger): "This pin is the input to the lower comparator and is used to set the latch, which in turn causes the output to go high. " as a control for Attack of the transient, and use Pin 6 (Threshold): "Pin 6 is one input to the upper comparator (the other being pin 5) and is used to reset the latch, which causes the output to go low. " as the Release control for the envelope.

Not quite sure if this will work as proposed, Any thoughts on this?






Still learning......................
:green:
:thumb:
:guinness:

 

[PRR]

> application of 555 as a envelope generator.

It is NOT an envelope generator!!!!!

While the 555 makes square waves, it times them with an internal ramp wave. Not so internal: the ramp is visible on the timing capacitor. Keith uses that as his $0.10 ramp-wave generator.

No, the 555 is not a precision device but we really don't much care what frequency. The 555 ramp isn't perfectly linear, but close, and in a feedback limiter that's good enough. (For FF we would want to replace the 555 timing resistor with a current source.) And since this is not a power push-pull affair, we really don't need off-time control or other frills added by "real" PWM chips. (If we did find we had to add a current-source to linearize the 555 bent-ramp, then other chips look attractive.)

Keith's complete plan makes it hard to see the basic operation. Set up a series resistor and shunt FET between a line out and a line in. When the FET is off, signal passes, when the FET is on the signal almost vanishes. We don't care about in-between, because we know the FET's part-on resistance is non-linear: this will be an entirely on-off operation.

To slice audio cleanly, the on/off rate must be faster than the audio. Use 100KHz for example.

Find or build a variable duty cycle rectangle-wave generator. You may have one in your bench signal generator. The 555 ramp wave plus a comparator and a pot will do the job too. Set it for about 100KHz, wired to the FET to turn it on and off at 100KHz.

When the FET is 99% off and 1% On, the output signal is nearly the same as the input(*). When the FET is 1% off and 99% On, the output signal will be about 100:1 or 40dB weaker. The duty-cycle knob on the rect-wave generator is now a Volume Control. And FET resistance linearity is insignificant.

(*)Nearly the same except there will be a LOT of 100KHz rectangle-wave on the output, slicing the audio very thin. That would be inaudible, except audio amplifiers overload on 100KHz trash and the sound will be full of whistles and beats. That's why we -need- a 20KHz low-pass on the output (for breadboarding you could use a guitar-amp, which low-passes fairly cleanly) and in most cases we need to cut any supersonics goiung into the R-FET network so we want an input low-pass. But it will "work" (test-bench demo quality) with no input low-pass and minimal output low-pass.

All the rest of Keith's drawing is side-chain (fairly routine stuff) to turn the VCA into a Limiter, and in/out interfacing for studio work.

 

[Learner]

[quote author="PRR"]> application of 555 as a envelope generator.

It is NOT an envelope generator!!!!!

[/quote]

Hi PRR,
These are what gave me the idea of using 555 as an envelope generator, without thoroughly understand their ideas I am just trying to create my own version of it :shock: :green:

http://www.uni-bonn.de/~uzs159/adsr2.png

http://www.midwest-analog.com/adsr1.html

http://www-timc.imag.fr/Yves.Usson/personnel/SDIY/archives/j-jacky-1980.pdf

http://www.itn.liu.se/~nikro/diy/littleman/littleman.pdf

 

[ToshiroTamigi]

In the 555 datasheet i've seen it used as a pulse position modulator without using an external comparator.
The only drawback is that the frequency is not fixed, but this shouldn't be a problem if the lower range is well far from audio range.

take a look at page 14 of
http://focus.ti.com/lit/ds/symlink/ne555.pdf

Alessio