String pluck detector??

[RGP]

Thanks Tom. Your article has me understanding the wave forms better. I'm new to this, but not to electronics.   

[PRR]

Where is the full Marston article?

PAIA-guy John had a 50 cent trick for making a ramp (the easiest wave) into a triangle.....

https://modularsynthesis.com/paia/4700/4720A/4720a.htm
https://modularsynthesis.com/paia/4700/4720A/PAiA%204720A%20Voltage%20Controlled%20Oscillator%20Manual.pdf
https://modularsynthesis.com/paia/4700/4720A/PAiA%204720A%20Voltage%20Controlled%20Oscillator%20Schematic.pdf

The manual explains the ramp-to-triangle gizmo on printed page 19.
The manual has a cut-off schematic.
I remember staring at a 'scope, AND listening, for an hour, and realizing it doesn't get better than this. (Actually it did, on our ARP, but that was >1000X more expensive. And IIRC, it made BOTH the ramp and the tri in two separate integrators with a trick /2 current source.)

Note the diode shaper to make sine from triangle. The 2-transistor scheme would have been knowable by John but bulk-buying diodes (and outsourcing labor to kit builder), the diode plan may have been cheaper.

[ElectricDruid]

Hey Paul,

You mentioned the tri-to-sine waveshaping that never removes the "point". This is often known as the "cusp" or "pip", and there's been a lot of effort spent trying to get rid of it. Tim Stinchcombe has a lot of good references and a nice overview:

https://www.timstinchcombe.co.uk/index.php?pge=trisin

The OpenMusicLabs paper about it isn't bad either:

http://www.openmusiclabs.com/files/otasine.pdf

Hope you find something interesting amongst it,
Tom

[Rob Strand]

Ah, good the image is there. Maybe this simple approach may workm but I have my doubts that it will work for multiple frequencies. What about dupicating it for each octae I'm using?

The main advantage of sine-shapers is there is no filtering and they operate over an enormous range of frequencies.   Function generators with square-tri-sine outputs quite often use sine-shapers.   The purity of the sinewave isn't that great, say 1% THD.   For applications with limited range or fixed frequencies you can improve the purity with a fixed filter.  For your application I doubt anyone would know if you filtered off everything above 2kHz.

You mentioned the tri-to-sine waveshaping that never removes the "point". This is often known as the "cusp" or "pip", and there's been a lot of effort spent trying to get rid of it. Tim Stinchcombe has a lot of good references and a nice overview:

Yes, the sine shapers which use the "differential pair" (like the example on your site) and the diode type used by Marston have pointy tips.   To some degree this is a sign the high order harmonics don't roll-off very quickly.

The types of shapers that use a diode limiters produce aesthetically pleasing sine-waves without tips.  They also roll-off the high spectra better.   The 80's era function generators often used this type.   The THD is still about 0.5% to 1% and with some effort you can sometimes null one of the lower order harmonics.    These sine-shapers look more complicated because each break-point requires four diodes instead of two (like the Marston version).

Sine-shaper is top left section:


At some point I found this one which is a little more economical.  Quite good output.  Input needs to be 4.0V to 4.1V peak but you could tweak the input divider for different input levels.



Fairly impressive performance, especially for a simple circuit:





I'm still trying to find the origin of this circuit.  I could not find it previously.
One reference is from Wavetek's 1983 patent.  Except here it is used in reverse.
(Wavetek are famous for there function generators.)

https://patentimages.storage.googleapis.com/f5/8f/6a/adbe36c8d96aef/US4415860.pdf

EDIT:
- I have confirmed this form of the circuit was used on Wavetek Function generators after the mid 70's.
- The example circuit above connects the second set of diodes to ground whereas the Wavetek
   circuits connect this point to the input terminal of the first pair of diodes.

I've some stashes of info on sine-shapers. Here's another one which came to mind,
https://till.com/articles/sineshaper/

[RGP]

PRR. It's in Marston's book. Here it is.

Rob & Tom Wow! What a wealth of info. Thanks for shairng it. 

[RGP]

PRR. If you don't have a digital copy of the Marston book I'd be more than happy to e-mail it to you. Since the circuit needs a triangle wave input that makes it easiser for me to choose a square to triangle wave converter and gives me a 4th waveform to boot. I found this one earlier today.   
The others than Rob & Tom posted arer also great choices too.

[RGP]

I've redrafted Marcus' schematics and added notes about setting the trim pots. I drated it all onto a single page, but this is likely too large to post. I'll break it down into sections & post them soon. The big question that still remains is where does the input signal come from to feed the EMM harmony generator clone? I suspect it's from the AMS100's compressor output but can't confirm that as a fact. Marcus' schematic also states some sort of filtering was used, but only he knows what it was.     

[PRR]

PRR. It's in Marston's book. Here it is. .... a digital copy

https://archive.org/details/110OpAmpProjects/page/n81/mode/2up?view=theater
like print page 77

The ARP 2500 had over-engineered VCOs and sine-filters; Pearlman was a heavy geek.
http://www.guitarfool.com/ARP2500/1004.html

[PRR]

....diode limiters produce aesthetically pleasing sine-waves without tips....

Typically by over-driving them, widening the wave, raising the 3rd and 5th. Nice phatt sound, but something I was doing didn't want that. Forget what it was.

Altho 3rd -90dB 5th -50dB aint bad.

And yes, there may be much else from the last 40 years I have missed.
__________

> the origin of this circuit.  I could not find it previously.  One reference is from Wavetek's 1983 patent

As said, the technique goes way back to hollow-state days.

Electronic Analog Computers, 1956,  by Granino Korn and Theresa Korn
https://archive.org/details/in.ernet.dli.2015.84794/page/n309/mode/2up?view=theater
Print pages 290-299 show vacuum diode wave-benders of every sort except sine, "many combinations are possible".

Maybe Theresa never published a good derivation of a sine shaper? (Maybe she held it as a commercial secret?)

[RGP]

PRR. I think you'd need a pilot's license for the ARP

[PRR]

A copy of Operational Amplifiers: Design and Applications by Jerald G. Graeme at the bedside is mandatory:
http://www.introni.it/pdf/Graeme%20-%20Applications%20of%20Operational%20Amplifiers%203rd%20generation%20techniques%201973.pdf

I see there is a later "3rd generation" edition.

[Rob Strand]

The ARP 2500 had over-engineered VCOs and sine-filters; Pearlman was a heavy geek.

It's not over engineered. It's got deliberate motives.  The first two diodes are the bare minimum "diode clipper" triangle to sine, which bends lower voltages.  Then the 4xdiode rounds the peaks to remove the pointy tip issue.  For the diodes it uses transistors for better matching so the peaks are accurately shaped.  I kind of like it.  I got complex shaper circuits which are very ineffective and probably not much better!

> the origin of this circuit.  I could not find it previously.  One reference is from Wavetek's 1983 patent

As said, the technique goes way back to hollow-state days.

The Wavetek one is a unique connection which seems to work well.  Tricky stuff like this is where people like Wavetek and HP (Agilent etc.) stayed ahead of the competition.   (If you plot the derivative of the output it is very sine like. where on some poorer ones the derivative looks like a square wave.)

The classic connection is the Marston type where the diodes tap off a divider.   That's the form most books.
IIRC the shapers in the ICL8038 chip sort of follow the classic structure and replace diode with transistors; XR2206 apparently uses a diff amp.

Then there's circuits which only use the 4xdiode structures like the Goodwill function generator I posted earlier.
(There's some twists on this one.)

There's a whole heap of patents with differential amp based circuits.   HP have a patent on at least one.  I don't know what product it is used in.

And yes, there may be much else from the last 40 years I have missed.

It's impossible to keep up with this stuff.  There must be hundreds of function generators and synths out there with shapers in them.   Today it's almost obsolete technology.

[RGP]

I finally have the schematic redrafted. It is in 3 parts. The 1st is the modified AMS100 input processor. The 2nd is Marcus' pluck detector. The last one is The E&MM Harmony Generator with Marcus' mods and mine, so far. 

[RGP]

HMM, somtthing went wrong with posting the images. I'll try again. I'm gertting a message no images can be uploaded. Is this a common problem?

[RGP]

Trying again to upload. Sorry, still not working.

[PRR]

Trying again to upload. Sorry, still not working.

Not relevant, but posting. How are you doing it?

[RGP]

PRR. I'm trying to add the image to the post. It didn't give me any problems before. Is there a better way?

[RGP]

Trying again. Could it be the file extension or size causing the problem?

[RGP]

You might like this one. 

[RGP]

trying again. No luck