Op Amp Questions, diode rectification and Bufferring...

[jBam]

Hi all…

I'm saving up for a DIY kit, and in the meantime I've been reading and reading about electronics.  I'm actually erring toward now inventing a gadget and building it - possibly as my first studio gear DIY build!

I've drawn up a little work-in-progress flow chart of processes my gadget needs, and I'm slowly ticking off understanding how to construct each step.

That said, I've got a couple of questions that I'm wondering about - perhaps someone here can assist in my understanding?!

1) Bufferring
I understand the basics of Unity gain buffering using an opamp.  What I'm not 100% sure of is:

For perfect (within practical limits) signal integrity in splitting signals, would I need to have each split stage buffered IF I was sending the signal to another Op Amp for a process?  I'm currently planning on splitting, separately processing and then recombining a rectified signal, and for my needs accuracy is key…  I'm also planning to pop a unity gain buffer stage on each split and THEN sending it to another OpAmp for a particular process, but now I'm thinking "wait - if the next process is via another Opamp, does it effectively buffer / isolate the system?" …  Hopefully that makes sense?!  I'm a little stuck on understanding this part…

i.e. do I need to put a buffered unity stage before an opamp when splitting signals in a gadgets signal chain to avoid a minor voltage drop?

2) Rectifying…

Would a typical diode based rectification of a signal be relatively low damage to the signal… I've been reading into different diode arrangements, but not sure if I'm reading too much into certain types of diodes and how they affect voltage signals… basically confused again! So:

Can a diode based rectifier be low loss (or theoretically no loss), or will it introduce anomalies into a signal.  I've briefly read through a few sites that manufacture IC style chips for rectifying… but I'd be happy with a handful of diodes if it works fine

3) General Op Amp reading…

I'm covering a lot of good reading via the net regarding the awesome different things that various Op Amps can do, but at times I'm a little short on detail.  Is there a really good go-to resource anyone could recommend for understanding basics (the range of things they can do) and then technical stuff (equations etc.)…  I'm happy with the web; but I'd also be happy with a book if there's an Op Ap Bible out there

---------

Thanks for anyone's assistance

Jonny

 

[Samuel Groner]

I think it might be best if you post a block diagram of what you have in mind. Without context it is difficult (for me at least) to give an informed response.

Samuel

 

[moamps]

3) General Op Amp reading…

I'm covering a lot of good reading via the net regarding the awesome different things that various Op Amps can do, but at times I'm a little short on detail.  Is there a really good go-to resource anyone could recommend for understanding basics (the range of things they can do) and then technical stuff (equations etc.)…  I'm happy with the web; but I'd also be happy with a book if there's an Op Ap Bible out there Jonny

I would recommend Op Amps For Everyone  from Ron Mancini:

http://web.mit.edu/6.101/www/reference/op_amps_everyone.pdf

 

[JohnRoberts]

Hi all…

I'm saving up for a DIY kit, and in the meantime I've been reading and reading about electronics.  I'm actually erring toward now inventing a gadget and building it - possibly as my first studio gear DIY build!

I've drawn up a little work-in-progress flow chart of processes my gadget needs, and I'm slowly ticking off understanding how to construct each step.

That said, I've got a couple of questions that I'm wondering about - perhaps someone here can assist in my understanding?!

1) Bufferring
I understand the basics of Unity gain buffering using an opamp.  What I'm not 100% sure of is:

For perfect (within practical limits) signal integrity in splitting signals, would I need to have each split stage buffered IF I was sending the signal to another Op Amp for a process?  I'm currently planning on splitting, separately processing and then recombining a rectified signal, and for my needs accuracy is key…  I'm also planning to pop a unity gain buffer stage on each split and THEN sending it to another OpAmp for a particular process, but now I'm thinking "wait - if the next process is via another Opamp, does it effectively buffer / isolate the system?" …  Hopefully that makes sense?!  I'm a little stuck on understanding this part…

i.e. do I need to put a buffered unity stage before an opamp when splitting signals in a gadgets signal chain to avoid a minor voltage drop?

The very input of a signal chain needs to be buffered. One op amp buffer can drive several unbuffered circuit inputs.

2) Rectifying…

Would a typical diode based rectification of a signal be relatively low damage to the signal… I've been reading into different diode arrangements, but not sure if I'm reading too much into certain types of diodes and how they affect voltage signals… basically confused again! So:

Can a diode based rectifier be low loss (or theoretically no loss), or will it introduce anomalies into a signal.  I've briefly read through a few sites that manufacture IC style chips for rectifying… but I'd be happy with a handful of diodes if it works fine

A simple diode rectifier (half wave) will immediately drop the forward voltage of the diode (typically 0.4V to 0.6V). So the small level parts generally get ignored.

More complicated ("precision rectifiers") can convert AC to DC without the diode loss.

3) General Op Amp reading…

I'm covering a lot of good reading via the net regarding the awesome different things that various Op Amps can do, but at times I'm a little short on detail.  Is there a really good go-to resource anyone could recommend for understanding basics (the range of things they can do) and then technical stuff (equations etc.)…  I'm happy with the web; but I'd also be happy with a book if there's an Op Ap Bible out there

---------

Thanks for anyone's assistance

Jonny

There are many good books out there.. You can also look at as many schematics as you can find and try to figure out why they used the parts they did and how they connect everything together to get a good result.  Learning how to design begins by studying other designs first. You must first acquire a vocabulary of circuit blocks, before you can form full sentences with them (string circuit blocks together to do something useful.)

JR

 

[Bomper]

This link has a lot of info...

http://www.analog.com/library/analogDialogue/archives/39-05/op_amp_applications_handbook.html

Regards, Thomas

 

[jBam]

To all who've so quickly responded ---> Thanks!!  Nice to wake up up wondering, and then find a bunch of helpful responses…

Thomas and Moamps ---> I'll check all of the those links at lunchtime…

JohnRoberts ---> Thanks a lot for your advice there - the buffered topic is what I was starting to realize - that once buffered the signal after is kind of self correcting (or is that it?… still plenty of reading to do!)…

I'll have good look into precision rectifiers - I'm quite certain this would be what I'd be wanting.

As for general schematic dismantling - yes!  that's what I've been doing; and I've actually starting to really understand it all.  Electronics has always been a huge knowledge gap - seemed like hocus pocus wizardry until a couple of months ago!  i'm really pleased that I'm starting to actually understand what is going on, and reading through schematic after schematic is how I'm learning…  And Block Diagrams --- > totally!  I'm sitting here with a block diagram slowly ticking off parts that I can understand and build!

Samuel ---> I might get around to posting up a total block diagram (once I refine it to something that makes sense!).  The basics of my buffering question are though:

Audio In ---> Buffer (split to 2):

1 (audio signal) ---> VCA
2 (sidechain) --->Rectify ----> Split to 2 "a" and "b" [buffer??]

a ---> subtract a chosen voltage (user controlled) via another Opamp (effectively the threshold) ----> send to sidechain summing amp "SC AMP"
b ----> send to sidechain summing amp "SC AMP"

SC AMP:  sum non-inverted "a" and inverted "b"

This will effectively extract any part of a signal above the "threshold" (the final sum will actually be [initial peak above threshold] x 2, so after this step I'd like to divide exactly by 2 to restore the initial peak information...  The faster and more accurate each step is, the better  the outcome (i.e. low distortion; super high slew rates).

^^^  From here, there are a huge range of things that I can do (this is the very start of how my circuit is intended to work).

To be honest ---> I have no idea how a solid state compressor work, or how the threshold works, but this is the approach I want to take… most of the ahems I've looked at are smaller scale, or completely unrelated (e.g. preamps / eq).  I feel like I invented this approach, but chances are that this is exactly how it's done in some or all solid state compressors? hahah… who knows?!!! I'd rather invent this thing, than fuss about how it "should" be done though !! 

----------------

While I'm on here ----> are there any passive components (or other) that have a non-linear response - perhaps in resistance…?  I'm trying to work out a soft-knee idea, where perhaps low voltage has low resistance, and high voltage has high resistance (or the other way around - either is fine).

Cheers again all!

 

[PRR]

> I have no idea how a solid state compressor work

Agree.

Walt Jung Op-Amp Cookbook- used copies turn up at prices far below its true worth.

Research, study, and digest ALL audio schematics you can find. There is much basic knowledge and convention still ahead of you on your path. The ones that don't "directly" relate to your thinking often turn out to have clever ideas which will be useful.

Nearly ALL audio design is re-hashes of old-old stuff. We stuff it into smaller bottles, or 3-pin pills, or 50-transistor chips, dress it up with a CPU and a tablet-app instead of proper meters, but there is little which has not been done before. We all stand on the shoulders of Giants, and tall stacks of fairly ordinary men.

AFAICT, what you have invented is a fairly expensive clipper. If you have a user-set control, precision is often not essential (user turns to desired effect, not 0.775,000 V).

What will your invention do which conventional audio boxes don't do now? In these days of arbitrary headroom, precision peak control is not the problem it was for disk-cutting or AM radio broadcasting. What real-time limiting always needed was anticipation (turn-down before the peak happens); this is trivial in computer audio post-processing.

> low voltage has low resistance, and high voltage has high resistance

Incandescent lamps (if there are any left) have that property; however there is a time-lag. Lamp filaments *have* been used for automatic level control. H and P adapted an old idea into their R-C audio oscillator 200AB. Many juke-boxes used a lamp bridge for more-boomy bass. Knowing of these done-it-before trends will guide you on your path, help tell which side-paths are worn dead-ends and which need further brush-beatings.

 

[jBam]

> I have no idea how a solid state compressor work

Agree.

Walt Jung Op-Amp Cookbook- used copies turn up at prices far below its true worth.

Research, study, and digest ALL audio schematics you can find. There is much basic knowledge and convention still ahead of you on your path. The ones that don't "directly" relate to your thinking often turn out to have clever ideas which will be useful.

Nearly ALL audio design is re-hashes of old-old stuff. We stuff it into smaller bottles, or 3-pin pills, or 50-transistor chips, dress it up with a CPU and a tablet-app instead of proper meters, but there is little which has not been done before. We all stand on the shoulders of Giants, and tall stacks of fairly ordinary men.

AFAICT, what you have invented is a fairly expensive clipper. If you have a user-set control, precision is often not essential (user turns to desired effect, not 0.775,000 V).

What will your invention do which conventional audio boxes don't do now? In these days of arbitrary headroom, precision peak control is not the problem it was for disk-cutting or AM radio broadcasting. What real-time limiting always needed was anticipation (turn-down before the peak happens); this is trivial in computer audio post-processing.

> low voltage has low resistance, and high voltage has high resistance

Incandescent lamps (if there are any left) have that property; however there is a time-lag. Lamp filaments *have* been used for automatic level control. H and P adapted an old idea into their R-C audio oscillator 200AB. Many juke-boxes used a lamp bridge for more-boomy bass. Knowing of these done-it-before trends will guide you on your path, help tell which side-paths are worn dead-ends and which need further brush-beatings.

Hey PRR... great - thanks for Op Amp Cookbook suggestion ---> Sound perfect...  Been busy working through lunch, so haven't had a chance to check other links yet...

Just a thought on my approach to this topic PRR ----> 1st up ---> Yes - absolutely will continue looking through schematics and theory... I assume I'll continue this "forever" in the pursuit of knowledge and understanding

At the same time, I guess a hobby is there for a couple of reasons, but one in particular:  to bring additional satisfaction into your life.  Under this guise, I must admit that spending time thinking about theoretical ideas and mathematical conundrums (or coincidences) is really fun for me... So theorising possibly stupid ideas is actually time well spent for me, because it adds value to my life ---> I get engaged in a moment, and solve random puzzles that I've made up myself...  Basically, I find it fun to think: "ok - so how the hell would I achieve that..."  Then I'll come up with theories, and think, and develop something  that should work.  9 times out of 10 it's not the same as someone else's way of doing it, and even in that I find fun and satisfaction "creating something new" (even if it's kind of stupid haha)...

I've had a little far off dream of a little company called "Oops Audio"... making stuff that is kinda stupid, but works.

e.g. ---> Electro-Mechanical Compression 

Basically ---> Imagine a typical VU meter, but rather than a needle, it has metal bar.  As the control voltage drives the bar up, gravity and general physics delays and restricts the way the bar moves... this restricted movement gets fed back to an amp as another control voltage.  So the "compression" occurs as the result of the mechanical process...  I have a feeling that it'd be a really nice compression curve......

Like I said - stupid ideas... but fun

So I guess on one hand, I'm "on my own personal trip" with some of this stuff...

On the other ---> and as a separate part of my hobby interest, YES:  totally right mate!  I need to read and understand and build and trouble shoot etc.

As a quick snap shot of one of things I was theorising that relate to my initial questions, I'm trying to come up with something that perfectly extracts a peak above a threshold (within practical limits)...  See the attachment for a fairly loose flow of the processes I think may work...

This, in theory, could extract a perfect version of the signal above the threshold.  In reality, nothing's perfect, and losses etc. may occur, and integrity somewhat compromised, but for now - I'll work in theoreticals!

There's a HUGE range of things I could then use this peak information for (in my crazy / stupid gadget I'm dreaming up).  Let's not worry about those for now though hehe...

 

[PRR]

> Imagine a typical VU meter..............

A HIGH power meter. Tie the needle to a volume knob. As the needle hits half-scale, it turns-down the input signal.

It can be (almost) as simple as that.

Which reminds me.... add to your studies, Watt's engine Governor, and Maxwell's essay thereon. This IS the audio limiter, except mechanical rotation instead of electric potential. This theory expands into Servo and Feedback theory, stability criteria, closeness of control, and other relevant topics.

 

[jBam]

> Imagine a typical VU meter..............

A HIGH power meter. Tie the needle to a volume knob. As the needle hits half-scale, it turns-down the input signal.

It can be (almost) as simple as that.

Which reminds me.... add to your studies, Watt's engine Governor, and Maxwell's essay thereon. This IS the audio limiter, except mechanical rotation instead of electric potential. This theory expands into Servo and Feedback theory, stability criteria, closeness of control, and other relevant topics.

Legend!  Thanks mate... I'll look into all of this.....

The more resources the better

 

[jBam]

Actually... I guess to experiment with the electro mechanical idea, I could just start with a cheap little electronic motor... Pop a heavy enough bar on it (like a propellor) and send it RMS signals...

I doubt a whirring motor is exactly what a studio needs haha... But as a cheap experiment, it seems practical ...

Be well all!...

 

[gyraf]

Thanks guys. It's a pleasure to see threads like this.

Jakob E.

 

[jBam]

Thanks guys. It's a pleasure to see threads like this.

Jakob E.

Ahh good - thanks Jakob ---> and now my hobby oriented satisfaction has spread to making others happy too haha…

Back on the initial discussion before the "electro-mechanical" compressor (which I'm still totally keen on)…  maybe it'd be fun to open this concept project to the forum as a genuine "group" DIY project!  i.e. I'll probably need help from all of you!

At the least, I might finalize my flow chart of processes, and people here could perhaps assist in me transposing some of it to a board one day…  There's a huge amount of experience here, and I'd certainly appreciate any ideas / trouble shooting at the theory level, and then I'd LOVE to build the thing I'm thinking of (hopefully it doesn't end up being a piece of crap haha)…  In honesty though, the end product I have in mind should be really cool and powerful as an overall dynamics system - it may even be something that others will want to build and have in their studios

I'd have only one proviso on opening it up to the forum in full ---> That it's intended for use by DIY people only, and that it's colorfully referred to as the "Bam Box Dynamics Processor", or BamBox for short  hehe...

(Haha… a disclaimer for something that doesn't exist and may not work haha).

Like I said, the "peak extractor" is only one small building block in the gadget… I'll get cracking on finishing up the flow chart, maybe sketch up some more CADD examples like the attachment above and then I'll post it up and see what others think on here… (I use CADD for my day job, and with this audio stuff it really helps me understand what I'm trying to do in theory, and assists me in "proving" the theoretical stuff because I can measure things and realize "ahhh… so that's twice that; or a third of that"; or "ahhh ok so they will cancel out, but result in a constant voltage of X" etc etc)

Cheers 

 

[CJ]

"Mopeds in the United Kingdom have had to have a 30 mph (48 km/h) speed limiter since 1977"

then how did that guy in Quadrophenia get so much air at the end of the movie?

here is a precision rectifier, probably want to modernize the OP90,

input diodes are for wave shaping,

 

[jBam]

here is a precision rectifier, probably want to modernize the OP90,

input diodes are for wave shaping,

Great - thanks for your input cj...! I don't think I quite understand the schem just yet, but I'll need to at some point!

Discussing my attachment above, and what I've been thinking of trying to achieve ---->

I'd originally just dreamed up ideas for basic dynamics principals...over the last year or so... But then I joined this forum, and wrote a joke comment in a Donald Trump thread about a "Donald Trump edition Mexican brick wall limiter"... And it got me thinking about absolute peak limiting... And the old " OK, so how would I do that"... It got me thinking about this peak extraction thing, and peak control ---> this then feeds into ideas of an overall dynamic control system I'm dreaming up...

BUT:

I've gotten a little carried away, and PRR is kinda right about a fancy clipper... I've been thinking of feeding the extracted peak back into the original to immediately reduce the level of the peak... This is flawed because my idea for peak extraction has been at a cyclic level of audio, and without looking back in time to a zero crossing, I'm effectively just instantly changing the shape of a wave... So Ive basically come up with a fancy wave shaper haha... For synthesis, this is cool, and I've already gone through a huge range of "whoah... So then I could do this, or create this" moments (from extreme shaping to more subtle things like creating a sine wave from a triangle, or rounding off the top etc). But for dynamics, it's flawed because cyclic control above a threshold is destructive control.

So limiter... Meh... I'll worry about the seemingly impossible look-back-in-time solid state conundrum another day haha

BUT (another but)...

If I consider this same technique at a broader level - applied to the signal of a wave follower - then it becomes hugely useful again!!  ;D

Yay

I'm then able to extract the peak info over a threshold, and then do a whole range of things with it... Invert to the original (flat line cutting the top off); change its level, and then stitch it back on (summed back together). Not sure if that makes sense, and I'm down at the beach house so no cute CADD pictures just yet... But there's a bunch of nifty things that I can now control regarding peak information in the wave follower signal.  And control at this level is not going to screw up like the wave shaping issue. I'd prefer to sketch it up in CADD, but one option is to use the extracted peak of a follower to amplify itself... This could then create a control signal that increases compression the more the input level is over the threshold (sort of like an endless soft knee).

Even cooler is that, with the approach in my attachment above, I can theoretically extract not only a perfect "peak" replication, but also a "middle chunk" (or any number of middle chunks). Not quite sure just how useful that is yet, but I suspect it could be really useful... At a cyclic level, more funky wave shaping... But at a wave follower level, I could target power/ level ranges in an audio signal... Sounds cool to me ...

Hopefully the above makes sense... I'll add some diagrams when I'm back in town next week...

Thanks again to all contributing here... At the least I'm finding this little exploration really interesting!!!

 

[JohnRoberts]

"Mopeds in the United Kingdom have had to have a 30 mph (48 km/h) speed limiter since 1977"

then how did that guy in Quadrophenia get so much air at the end of the movie?

here is a precision rectifier, probably want to modernize the OP90,

input diodes are for wave shaping,

 

This is what I meant by precision rectifier.

JR

 

[PRR]

> what I meant by precision rectifier.

And you don't need this "precision" if you are only caring about HIGH level signals. Boost your peaks to 10V, you don't care about 0.6V diode-drops (they can be calibrated away).

 

[Rob Flinn]

"Mopeds in the United Kingdom have had to have a 30 mph (48 km/h) speed limiter since 1977"

then how did that guy in Quadrophenia get so much air at the end of the movie?

Because Scooters (which are what you see in Quadrophenia Lambrettas etc) aren't considered mopeds in the UK.  Mopeds are under 50cc.

 

[jBam]

OK cool... So: thanks again everyone, and thanks for popping that precision rectifier setup John - I think I can just scrape my brain around understanding it ...

Thanks also again PRR - I might touch base with you regarding what your talking about, but I think I understand - the signal integrity the initial push away from zero is possibly compromised, but peak info is fine (sort if like friction - hard a first, but then skids consistently). Perhaps you can point me to some reading if I'm wrong there? Because I'm wondering about stitching a chopped top back on to the lopped bottom, I might aim for something super precise like the precision rectifier idea, but hey ---> I'm still in concept stage, so the conversation can continue! Certainly appreciate your input and advice - that's for sure!

Now for the fun bit: I've been sketching away, and wondering about "so I can separate the high level; and now I can separate a middle portion of level... And therefore leave the remaining low level... But what is the middle part good for??"...

And I've worked it out - and I think I think it's gonna be great!!... In fact - I think this gadget might actually work and do something useful!!

I'll get the flow chart happening, and sketch some stuff out and post it up... For now - well I'm at the beach and should probably try and stop thinking about this stuff haha... Maybe be even go fishing to free my mind if not for me, perhaps others --> my girlfriend's certainly getting a little bored of me showing her hand drawn graphs and rectifying maths theory ...

 

[abbey road d enfer]

e.g. ---> Electro-Mechanical Compression 

Basically ---> Imagine a typical VU meter, but rather than a needle, it has metal bar.  As the control voltage drives the bar up, gravity and general physics delays and restricts the way the bar moves... this restricted movement gets fed back to an amp as another control voltage.  So the "compression" occurs as the result of the mechanical process...  I have a feeling that it'd be a really nice compression curve......

Like I said - stupid ideas... but fun

Not so stupid. PRR suggested motorized faders for limiting.
There is a fundamental notion there, it's solid friction. The absence of solid friction in limiters results in LF distortion, which has been the subject of numerous attempts to get rid of. Hysteresis and dynamic time-constants have been used, but don't solve the problem entirely.
Electromechanical gain control solves the problem beautifully, but cannot achieve the promptness of reaction needed for true brickwall limiting.
Only recently digital processing of the side-chain has solved the issue, but once digital processing is involved, there is a small gap to go fully digital.