Wider output range from op-amp
I have been developing a pre-amp for an electric guitar using op-amps. I emulate valve sounds by using an array of diodes and for a soft "crunch" sound it is perfect. However when trying to provide a thicker distortion I am clipping right at the output of the op-amp and then it is too late to do anything to shape the output curve. Increasing the power supply of course helps but you can only go up to 36 volts in most op-amps.
Is there a way maybe I can use two op-amps like in a cascode way to allow an output swing over 36 volts?
simplest is if you have balanced reciever at other end - then bridging doubles the net Vswing
or even simpler - just change system gain structure - add 2x in the guitar amp (just turn it up?)
a line level audio transformer could do the step up too
I have cascaded op amps - not highly recommended for the complexity, possible stability problems
http://www.edn.com/contents/images/45890.pdf is one approach with discrete output
if you have extra sec windings you could make a floating supply amp - can double swing - again uncommon design criteria
adding a discrete transistor gain stage output but inside the op amp feedback loop is more common - but also has stability design requirements
If you are using diodes to get distortion then the output will never swing to the power rails , the diodes clip the output signal to the diode voltage (0.7v for silicon 0.3v for germanium and 1.5v for LED"s) .......
If you want a larger output voltage I would use one opamp stage with clipping diodes then use another opamp gain stage without diodes for makup gain ......
you can also use diodes in series to extend the output voltage before clipping , 2 silicon diodes in series will give you double the output voltage before clipping than a single diode .......
+/- 36V OUTPUT with +/- 40V supplies
I am thinking aloud as I am typing this :)
I am trying to create a solid state guitar pre-amp / amp that sounds like a valve amp, as much as possible. If I cannot reproduce the sound then I will probably go and buy a Mesa Boogie Lonestar and be done with it.
The above valve amp, has a pre-amp voltage gain of Av=50 on the "clean" channel and again, another Av=50 on the "distortion" channel, which is sequential from the clean channel but through a "drive" knob which at best divides by 2, ie Av = 50 * 50/25 = 1250.
We also know that valves have quite a lot of headroom before they clip hard, starting with a smooth rounding of the curve rather than an abrupt clipping.
A Gibson Les Paul Custom at full volume produces 0.7V RMS on a 1M load when strumming hard, and about 0.3V RMS when playing solo sections.
Having run some listening tests, I have arrived at a circuit that uses an array of diodes to compress the output swing in a soft manner, and after some tests we have achieved the best sounding "crunch" (soft overdrive where each note is distinct even when strumming chords) using 6 pairs of BAT48s.
The diode array comes after the op-amp gain stage. Obviously if the gain stage has already clipped, we have lost the game.
With 1V peak and an op-amp stage of Av=50 on the clean channel which has no "drive" knob, we need a voltage swing of +/- 50V which must not be clipped (otherwise it is not clean). This also implies a supply of over +/- 50V. We are approaching then valve-like voltages.
My initial circuit was tested with two 9V batteries, so the available voltage swing was puny. The diode array did its best to smooth the sound, and we did get a great sound, the "crunch" was the best ever my son said, but of course if we have clipped inside the op-amp we are not doing it right.
So just to recap, I need a clean voltage swing of +/-50V to be able to match the valve sound on the pre-amp stage.
Here is how I do it.
go to: DigiKey Electronics - Electronic Components Distributor
search for "amplifier"
gets you 45,169
select "Linear - Amplifiers - Instrumentation, OP Amps, Buffer Amps (32816 items)"
then check the box "in stock"
that gets you 13,067
select number of circuits 1
now go to the far right and select "bulk", "cut tape", and "tube"
now go to "Voltage - Supply, Single/Dual (±)" and scroll to the bottom of the column and select all that can operate at +/- 60V supply or greater.
That left me with 34 choices.
at the top of the column labeled "unit price" click on the arrow pointing up...then select "simple price..." in the popup window.
That is how I would find an opamp for your requirements.
Or get a design for a AF power amp that uses +/-70V rails and don't use a heatsink.
Or use a small power amp and a step-up transformer.
Or drive a diode pair with an AC current source to simulate a high voltage swing/large resistor combination.
Lots of other ways.
My first choice: Current source
A little C to help stability (if required)
Play with the values to get the sound you want.
Voltage swing has nothing to do with getting a valve type sound , it"s all about getting the right type of clipping then shaping the tone , even if you could get a SS preamp to put out a voltage swing of +/-50v the SS power amp in the next stage couldn't handle the signal as most SS power amps expect a 0.7v-1v line level input signal and 100v would clip the power amp like crazy or even fry it ......
There several preamps out there that have a tubeish sound that run off of as little as 9v and have less than +/-4.5v output swing though I have found the best ones aren"t opamp diode clipping circuits but use J-fets which have a much more tube like sound when clipped ......
I spent years trying to get a great tube like out of solid state circuits and while I had built a lot of great sounding pedals and preamps none of them emulated the sound of a tube preamp to my satisfaction ......
if you want a great tube sound then I would just build and actual tube preamp , its not really that hard and there are charge pump circuits that will give you 250v+ from 12v DC or you can use back to back 12v AC Xformers to give you voltages needed for a tube preamp ......
Thanks for all the suggestions above. Every day my son and I spend about an hour making or tweaking a circuit and performing listening tests. Unfortunately we do not have a tube amp at home to compare with.
I have ordered two OPA445AP chips (high voltage op-amps with FET inputs so you can plug a guitar on them and use high pots for the tone stacks) - they are arriving today to try things on. They are very expensive, hence I bought just two.
Currently experimenting with TL072/OPA2341 (both FET inputs), supplies at +/-17V with max Av=50 and Rf bypassed with 25 pairs of germanium/schottky diodes. Very similar to Doug's schematic above, which is also found in the Fender Princeton Chorus (at least). The difference is I use the right number of diodes to account for the total voltage of the rails, which is almost the voltage across "Rf", and they are germanium and/or schottky. For 17V supplies you need 120 diodes in total. I have used around 50 as I did not have any more.
I am trying to get the shape out of the array of diodes with no clipping in any stage/op-amp. With +/- 15V supplies it gets quite hard to achieve. Our first try also has quite a "fuzz" sound on the heavy distortion channel so we are experimenting there too.
I also just noticed on the Mesa Boogie Lonestar, a pair of back to back 15V zeners (4744) just before the post-amp stage... Hmmm...
Some of the tests are we are doing are:
1) strum the low strings, hear the distortion (and look at the scope) - how long does it last before it falls back to pure sinusoidal? Ideally it should not, ever revert to sinusoidal. In practice, with the soft-overdrive, probably 2-3 seconds.
2) strum the high strings, repeat test above. Is the time taken before it goes back to sinusoidal exactly the same? If not there is an issue. And we have an issue in that our array of diodes creates a fuzzy sound - caps need to be strategically placed to tame the fuzz unless you like the Jimi Hendrix sound (my son does not so we will probably have a switch "fuzz/deep".
3) do we detect hard clipping anywhere at the output of any op-amp using the scope? If yes then we are driving too hard and must cut down the drive.
4) examine shape of compressed curve. Is it nice and rounded or is it clipped anywhere? Typically that means that at the part of the circuit we have exceeded the supply rails.
Today I will try to compress two channels in sequence so as to stay within the +/-17V supplies, before the new, high volt chips arrive. Although this is not how the Lonestar does it. Looking at its schematic, the first stage multiplies by 50 and does not distort at all under any circumstances (it has no variable gain controls for the first stage so the sound must be totally clean by definition).
Dear akis, you are asking for help but you are not showing the schematic you are using ;)
Then any suggestion is only a wild guess.
It is too soon to publish any circuits as I am experimenting with almost every aspect at the moment and the sound is still not quite right.
I have transcribed yesterday's circuit from the breadboards and it looks like this.
The first stage was added in a hurry to give us more distortion late last night. Without the first stage we get a "clean" channel that overdrives when the gain is turned over 50%, but it never really distorts too hard.
The tone stack is the usual Fender adjusted to match 100K log pots which I have had handy as well as provide the contour of the Princeton Chorus. The pots should be all audio log.
|All times are GMT. The time now is 06:28 AM.|
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio