Interesting harmonic generation circuit from the 70's

[Mark Hammer]

Here is an interesting circuit from the esteemed Bernie Hutchins.  I was flipping through some binders in a box in the basement yesterday and came across my old pile of Electronotes "preferred circuits collection".  Bernie prefers that his IP be respected so I'm not going to start scanning and posting stuff, but I figured this little tidbit could be provocative and useful and is old enough and small enough that no harm is done by posting it.  The figure shown is a redraw.  In describing a "mallet-controlled" signal generator, he writes:

"There are several design goals we had in mind here.  First, there should be a small linear region about zero.  Secondly, we want only weak non-linearities when they do start.  Thirdly, we would like a non-symmetric non-linearity somewhere in the middle range in order that we may generate even harmonics.  Finally, we would like to end up with a somewhat balanced waveform at the maximum amplitude.  The circuit...realizes these goals.

For small signals, all diodes are off, and the circuit is linear and unity gain.  When the output exceeds +/-o.6 volts, diode pair A and the extra 100k resistor come into play, rducing the gain factor, rounding the top.  At two diode drops of input signal (+/-1.2V), diode pairs B come in with their 22k resistor, greatly increasing the gain, and peaking the top.  Beyond three diode drops (1.8V), diode pair C comes into play, substantially reducing the negative going gain.  This gives the desired non-symmetry.  Finally, at slightly over a 5V level, the section with diodes D comes in, effectively allowing no further gain on the positive side, allowing the negative side to catch up if it can.

Thus as the output of the filter (nearly a sine wave) increases in amplitude, the output of the non-linear circuit has harmonics, and the relative amplitudes of the harmonics changes with amplitude (and thus with time).  This is the type of response we want."



The circuit shown assumes signal levels more typical of synths, and consequently makes diode and resistor choices based on that.  Of course, there is no reason why the zener HAS to be 5v, and no reason why the other diodes have to be Si types or even why the reversed-biased combos have to be matched in terms of number or type of diode.  The idea of having the path that the input signal travels be dictated by its level is an intriguing one to be sure.  That alone strikes me as the basis for a really interesting and simple dynamic expander.

I leave the rest to your imagination.  In the meantime, all hail Prince Bernie.  The idea that at one time Cornell University possessed the skills of BOTH Robert Moog and Bernie Hutchins is sort of mind-blowing.

[Brett Clark]

Hmm. That's interesting prior art vis-a-vis some of the later distortion patents.

[brett]

Hi.
I'm slightly unclear about what is being described.
Is it like this?  We descussed such things on the forum about six months ago.
cheers

[brett]

I guess I should explain the ideas behind the schematic.
The left hand op-amp is just a booster so that the diodes on the right will work.
On the right we have two banks of diodes.  As the signal increases, more of the diodes become forward-conducting.  As more conduct, the resistors involved also get smaller, so gain is reduced (initially it is 10x).  

The final "trick" is that the two banks of diodes have different resistors, so the voltage-gain relationship is different for positive signals and negative signals.

I did some calculations in an excel spreadsheet and concluded that the "steps" in gain might be too strong and produce nasty sounds, but I've never built it to find out.

cheers

[petemoore]

All very interesting, I don't know that "D" would be needed, as 5v is alot of signal.
 I can't decide which one to recommend someone try building !!!
 Cool concept, that's for sure.
 getting all those diodes to 'do' something looks like a balancing act, *big kudo's to Bernie Hutchins and Brett, who've calculated these circuits...this second time seeing it, and the archived material reminds me that I can't wait to hear what that strange' of thing sounds like !!!

[gez]

There are a few schematics out there using variations on this theme.  I've used it myself to do sine and hyper-triangle shaping for LFOs.  It works, but if you want consistent results you need a regulated supply and the amplitude of the input signal (this method is usually used for triangle to sine conversion) needs to be rock solid (well, for oscillators it does).  

The resultant wave form can be a little angular - 'Frankensine' - but the more 'tiers' you use the smoother things get.

[Stevo]

I was looking you can get a zener diode at around 3 volts :?  Still not enough signal coming form a guitar though  :? ...would a 3 volt work I dont know??

[brett]

In my schematic, the first stage of clipping happens at 0.7V (just like a tubescreamer, etc).  Clipping (ie gain reduction) increases at 2 and three times that value. ie 1.4 and 2.1V, as more diodes conduct.  In a practical circuit, I think it would be good to use 2 x Si diodes for the "threshold", then several Ge diodes for the clipping "steps".  

The result *might* be like a softer overdrive version of a tubescreamer.  Which makes me wonder whether it has any advantages over your basic tubescreamer...  

cheers

[WGTP]

I tried this when it came up before and it reminded me more than anything else of using GE diodes in a TS type distortion, even though I only used SI's.  It had the same sort of ??grainey?? distortion character to it that I associate with GE diodes.  Not sure if that is a product of soft clipping or what.  

[jmusser]

Well Mark, you polled a couple days ago on what type of threads we like. This type right here is my favorite! I always hard copy these, and stick them in a file drawer, sort of like when R.G. did the "Technology of the Super Fuzz". Also, it had your name at the top. :wink:

[puretube]

another golden oldie... (US2890832)

[Mark Hammer]

Where do you find this stuff?

With a patent date of 1959, obviously that little triangle does NOT represent an op-amp.....well not a 741 in any event.  Which provokes the question:  "Just how long HAS the triangle been used as a sign of an amplifier stage in schematics?"

[Paul Marossy]

"Just how long HAS the triangle been used as a sign of an amplifier stage in schematics?"

For a long time! That's just a generic symbol for an amplifier and has probably been in use as long as the rest of those electronics symbols...  :wink:

[puretube]

Where do you find this stuff?...

partly here:



fun aside: I want to be sure, not to violate any patents or similar stuff...
so I need to know what`s been going on in the past,
and what`s happening currently.
(AND what has not been done yet!).

That`s why I`ve been doing patent research since ~ 10 years
in the relevant classification categories...

[Mark Hammer]

Well that makes absolute perfect sense.  The circle is complete.

Take note, kids, this is the sign of a professional.  Are you interested in becoming a designer?  Doing it professionally means that what you end up with can legitimately be called YOUR intellectual property, and that whatever revenue you make from your ideas does not have to be funnelled into litigation.  At the end of the day, your great ideas should be ideas that ARE worth pursuing, and not just somehing someone else thought of and stopped making.  That's not a criticism, veiled or otherwise, of anyone.  Rather it is a reality of what it means to make your living by thinking up things in the world of electronics.

Thanks, Ton.  You did many a service by showing those shelves.  I've only got one like that.  I'm jealous.

[rodriki1]

Hi all

Hey mark have u tried that???

http://www.freepatentsonline.com

very good site.

Pritchard patents have exaustive examples of diodes
and transistor curve shaping.

[puretube]

This One is the link that the patent-lawyers in europe use for in-depth research, if you pay them well...

While this here is nice for sharing one`s findings (once the number of the relevant file is known), since it enables the download of the complete .pdf-files in 1 rush...


BTW: never blindly trust everything written/drawn in a patent...

[EDIT]:
P.S.: may I add, that before having access to the "web" (since ~ 6 years now),
I had the luck to live only 1 hour by bike away from the Nuremberg
PO branch...  

P.P.S.: the most interesting part of a patent-file is the "Claims"-section!

[Stevo]

Please back to topic at hand....will Marks original circuit work with guitar or would you need a 3 volt zener diode since the signals dont get that high?   I think it would be a cool dual op-amp with one side set as gain and the other side set to harmonize like explained??

[soggybag]

I wonder if you could use transistors wired as diodes B to E. I think these can act like zener diodes. I seem to remember one of Arons designs using IRF50 (?) as clipping diodes.

[puretube]

IMHO, you can replace the Zener with a multiple of normal (Si-) diodes
(x-times ~0.6V);

also, me thinx that a transistor`s B/E will behave very, very similar to a normal diode; (both in Si & Ge, resp.);

sorry, if this one is being regarded as "[OT]" by the starter of this thread, or anyone else... :wink:

(if so, read on
here, to stay "OT"...)