Wider output range from op-amp - Page 2 - diyAudio
Go Back   Home > Forums > Amplifiers > Chip Amps

Chip Amps Amplifiers based on integrated circuits

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 13th March 2013, 02:11 PM   #11
JMFahey is offline JMFahey  Argentina
Sin Bin
 
Join Date: Mar 2009
Location: Buenos Aires - Argentina
Thanks for posting, otherwise it would have been only a guessing game.

1) it looks good

2) you do *not* need +/-60V Op Amps, 50V RMS, etc.
You are doing fine with what you are showing.

3) with those rails, 20X gain in the first stage is about max you can risk with "normal" pickups.
It will clip with an EMG or a Super Distortion pickup, so I suggest you lower U2a to 10X, to play it safer.

4) you don't need higher rails for U1a
Since the diodes are in the feedback loop (I was not sure about your configuration before), and you have "only diodes" there, no resistors, peak voltage will be peak diode array saturation voltage.
*You* are deciding this, so pull some diodes until you are sure you don't reach rails.

5) you are using a non inverting configuration for U1a, so to that "own" generated signal, you also add signal present at its input (pin 2), so if you put, say, 5V (clean signal by the way) there, those same 5V will present at the output, mixed with the distorted signal.
And I think there is where you have a problem.
A little clean signal mixed with distortion helps clarity, but here it may be overpowering.
Another reason to lower U2a gain.

6) the rest looks fine, with the only small detail you forgot to scale the treble cap, C2, which should be 470pF

7) be realistic as what you are trying to achieve.
You can get a very good sounding SS preamp, which is excellent, but tubes clip in a very different way, totally unlike diode clipping.
You can simulate both and look at the very different waveforms.

In fact, when clipping diodes are introduced in tube amps, such as infamous Marshall's JCM900 (and a few JCM800) sound is later described by Musicians as harsh/brittle/fuzzy/you name it.

Good luck.
  Reply With Quote
Old 13th March 2013, 02:57 PM   #12
akis is offline akis  United Kingdom
diyAudio Member
 
Join Date: Jul 2008
Argentina? It must be in the small hours there?

Let me reply :

3) Yes first stage was a quick bolt on very late at night when my son said "what would the distortion channel sound like?". I will need to make sure there is no clipping under normal circumstances. We have 3 electric guitars, two with EMGs , but the old Gibson is still the loudest on its bridge pickup.

However there is not much I can do if someone plugs in various stomp boxes and pedals at the input with the volume knobs to max. This is a major and often overlooked part of a design. You have to draw the line somewhere. You cannot account for everything. Obviously using a tube that can swing to +/-100V at the input certainly helps. We can also add back to back zeners at some arbitrary voltage but that is not soft clipping. Still no substitute for a 200V tube.

This is one of the parts I want to get right - having a +/- 50V chip at the input will certainly help. Shame my bench PSU only goes up to +/-25V.

4) The diode array as you say determines the peak votage, which is more or less the rail voltage (less the capability of the op-amp, less the 2K resistor). Adding more diodes increases this peak voltage and removing diodes decreases it. Using a lot of small voltage drop schotttkies or germaniums helps shape the curve gently, rather than using say a zener which would be abrupt and almost as bad as the op-amp clipping.

In terms of sound quality, removing a few diodes results in clipping earlier, thus producing more distortion. Adding more diodes suppresses the signal more gradually.

My ideal would be to use just enough diodes to suppress 20-30% below max rails level - some peaks would still escape and reach the rails but with the extra headroom it would still not clip.

In the simulation I used a lot of diodes to drop 15 Volts in either direction. In the actual breadboard I ran out of diodes and used a combination of germanium and schottkies, 26 in each direction, keeping the signal from clipping just under 8-9V with max rails at 17V.


5) I am not sure I understand that. Are you saying that stray capacitances around U1 will feedback distorted singal from pin 1 (out) and 2 (Vin-) back to pin 3 (Vin+) and thus create some sort of unwanted positive feedback? I will try to make U1 an invering amp to see if it makes a difference.

6) for the treble cap I checked with the scope at 50Hz and 5000Hz - I tried to equalise the treble with the bass when the respective pots are turned to max. Around 220pf gave me same treble as bass. I used the ToneStackCalculator for initial design trying to copy the Fender Princeton's curves. I decided that the Mesa Boogie's Lonestar stack is not good, and looked at the Princeton Chorus stack, with uses 50K pots and achieves much more bass, so on the tonestackcalculator I tried to match the curve of that. It still is not good, too little bass, treble is OK. Very subjective matter BTW.

However, my son's friend buight an Egnater valve amp head, and while he was playing I was messing with the clean channel's tone controls, and believe me I could not hear a difference with the treble past 12 o'clock.

7) the whole idea behind it is this - and I may be very wrong.

a) there is a difference in the tone with amplifying the guitar singal by 50 with a tube without clipping, then tapping 1/10, resulting in a mathematical "Av=5" and with simply amplifying by 5 in an op-amp.

b) if I amplify by 50 and get a 50V pk swing, then use 400 germanium diodes to "shape" this signal (saturation at 100V), I will emulate the tube which also clips at 100V and plays with maybe 1% distortion at 50V.

But I may be very wrong.
  Reply With Quote
Old 13th March 2013, 09:19 PM   #13
JMFahey is offline JMFahey  Argentina
Sin Bin
 
Join Date: Mar 2009
Location: Buenos Aires - Argentina
Quote:
Originally Posted by akis View Post
5) I am not sure I understand that. Are you saying that stray capacitances around U1 will feedback distorted singal from pin 1 (out) and 2 (Vin-) back to pin 3 (Vin+) and thus create some sort of unwanted positive feedback? I will try to make U1 an invering amp to see if it makes a difference.
No,I do not speak about signal feedthrough or parasitics, I speak about having at the output (and in phase) *any and all* signals present at the +input.
Even if you short the diodes , you still have the full input signal at the output.
That does not happen in an inverting amp

Quote:
6) for the treble cap I checked with the scope at 50Hz and 5000Hz - I tried to equalise the treble with the bass when the respective pots are turned to max. Around 220pf gave me same treble as bass. I used the ToneStackCalculator for initial design trying to copy the Fender Princeton's curves. I decided that the Mesa Boogie's Lonestar stack is not good, and looked at the Princeton Chorus stack, with uses 50K pots and achieves much more bass, so on the tonestackcalculator I tried to match the curve of that. It still is not good, too little bass, treble is OK. Very subjective matter BTW.
Yes indeed. For guitar preamps, "flat" is "bassy".

Quote:
However, my son's friend buight an Egnater valve amp head, and while he was playing I was messing with the clean channel's tone controls, and believe me I could not hear a difference with the treble past 12 o'clock.
Marshall type EQ uses a linear pot there, Fender type a special Log with 30% centerpoint, instead of the usual 10% , so unless you use the special pot, it's either "too slow" or "too fast".

Quote:
7) the whole idea behind it is this - and I may be very wrong.

a) there is a difference in the tone with amplifying the guitar singal by 50 with a tube without clipping, then tapping 1/10, resulting in a mathematical "Av=5" and with simply amplifying by 5 in an op-amp.
Yes, there is, but not *that* large, the op amp signal will be very clean, the tube will have, say, 1 or 2% distortion, quite soft.

Quote:
b) if I amplify by 50 and get a 50V pk swing, then use 400 germanium diodes to "shape" this signal (saturation at 100V), I will emulate the tube which also clips at 100V and plays with maybe 1% distortion at 50V.
Maybe they start clipping at the same level, if you choose the proper amount of diodes, but beyond clipping waveform will be very different.
  Reply With Quote
Old 13th March 2013, 10:55 PM   #14
akis is offline akis  United Kingdom
diyAudio Member
 
Join Date: Jul 2008
So the first change is to make the amps inverting so that the input signal does not appear at the output, especially if the output is "modified". Had never thought about that before, thank you.

The second problem is to find a bench PSU for +/-45V... Not an easy task and am a bit cautious to breadboard straight out of a +/-45V supply without any current limiting. And the two OPA445 chips are beckoning...

The third item would be to find a way to pack 100 or 400 diodes on a breadboard and run more tests.

I will post comments/schematics/results as soon as possible.
  Reply With Quote
Old 14th March 2013, 12:35 AM   #15
diyAudio Member
 
Loudthud's Avatar
 
Join Date: Nov 2007
Location: Near Dallas Texas USA
You might find these two threads on a different forum of interest.

Triode Emulation X-Y plots
Look Ma, NO POTS!

I think you have to join the forum to see the attachments.
  Reply With Quote
Old 14th March 2013, 12:39 AM   #16
benb is offline benb  United States
diyAudio Member
 
Join Date: Apr 2010
I see your diode block is general and not detailed, but it looks familiar. Someone has already mentioned the inverting idea, and I've done that too. I posted an LTSpice schematic with a bunch of diodes here in post #18 (remove the .txt file extension, and load it in LTSpice, a free simulator program):
Modifying in-built distortion: transistors in an op-amp feedback loop?

Also, it may be cheaper and faster to simulate this in LTSpice than to keep making circuits. LTspice also has a neat feature I've never used (haven't looked into exactly how it;s done), you can send a .wav file into a circuit and generate a .wav file from a circuit's output. I've made this circuit and the output has a big dynamic range, both in distortion type and in volume, and in that respect it's "better" than tube distortion. I use 1n914/1n4148's with their usual 0.6V drop, but I've considered using germanium or Shottky diodes for the same reasons as you, to get more diode drops per volt of signal.

One of the things I've wanted to do with that circuit is make it with the diodes in a single direction instead of both, and make two of the circuits cascaded with an attenuator in between so the distortion given to the positive and negative parts of the wave can be varied independently.
  Reply With Quote
Old 14th March 2013, 12:40 AM   #17
diyAudio Member
 
nigelwright7557's Avatar
 
Join Date: Apr 2008
Location: Carlisle, England
There was a circuit for a soft limiter in the Wireless World magazine in the 1980's.
It worked very well for emulating valve type distortion.
It used two transistors in the feedback loop of an op amp.
It had a pot to alter the depth of distortion.
  Reply With Quote
Old 14th March 2013, 10:42 AM   #18
akis is offline akis  United Kingdom
diyAudio Member
 
Join Date: Jul 2008
Quote:
Originally Posted by Loudthud View Post
You might find these two threads on a different forum of interest.

Triode Emulation X-Y plots
Look Ma, NO POTS!

I think you have to join the forum to see the attachments.
Just from the first post it seems this guy was reading my mind! I had the idea of using dozens of diodes to chip away little pieces of the curve at a time way back, honest!

I am also of the opinion that a "passive" diode array would be better than clever tricks with "active" op-amps or transistors in the feedback loop.

Having said that, I recently spent a lot of time on a wien bridge oscillator using diodes and then FETs in the feedback loop (FETs being less temperature sensitive). I was testing with a 3KW fan heater heating up the circuit to 70-80C and trying to get a stable waveform (amplitude). The original design (HP comes to mind) used a light bulb.

So I acknowledge there is a lot one can do in the feedback loop, so am now registering and ready to read (and learn) more.
  Reply With Quote
Old 14th March 2013, 11:11 AM   #19
akis is offline akis  United Kingdom
diyAudio Member
 
Join Date: Jul 2008
Quote:
Originally Posted by benb View Post
I see your diode block is general and not detailed, but it looks familiar. Someone has already mentioned the inverting idea, and I've done that too. I posted an LTSpice schematic with a bunch of diodes here in post #18 (remove the .txt file extension, and load it in LTSpice, a free simulator program):
Modifying in-built distortion: transistors in an op-amp feedback loop?

Also, it may be cheaper and faster to simulate this in LTSpice than to keep making circuits. LTspice also has a neat feature I've never used (haven't looked into exactly how it;s done), you can send a .wav file into a circuit and generate a .wav file from a circuit's output. I've made this circuit and the output has a big dynamic range, both in distortion type and in volume, and in that respect it's "better" than tube distortion. I use 1n914/1n4148's with their usual 0.6V drop, but I've considered using germanium or Shottky diodes for the same reasons as you, to get more diode drops per volt of signal.

One of the things I've wanted to do with that circuit is make it with the diodes in a single direction instead of both, and make two of the circuits cascaded with an attenuator in between so the distortion given to the positive and negative parts of the wave can be varied independently.
Diodes are an array of between 52 and 80 in the feedback loop depending on what I had in my drawer and how bored I was trying to connect so many together.

For simulation I use NI Multisim version 12. Listening tests are a must, and the curves on the scope after distortion and filtering are nothing to do with the simulation.

I believe that to get a smooth distortion you need to "bend" the curve slowly and incrementally. Otherwise you might as well let the op-amp clip due to its output protection circuitry or because it has reached close to the rails.

This means that the feedback circuit has to be such that it will not activate the op-amp's various protections or limiters in anyway. Some op-amps are much better than others in driving higher loads at full range. I think we did detect an audible difference between the TL072 and the OPA2134 but I cannot say categorically. I am however sticking with the OPA2134.

I attach some pictures to liven up this thread
Attached Images
File Type: jpg PHOTO_20130314_095757s.jpg (373.6 KB, 56 views)
File Type: jpg PHOTO_20130314_095747s.jpg (322.7 KB, 52 views)
File Type: jpg PHOTO_20130314_095738s.jpg (429.2 KB, 35 views)
  Reply With Quote
Old 14th March 2013, 02:32 PM   #20
diyAudio Member
 
Join Date: Feb 2006
Location: On a hill, in a wooden shack, next to the woods, in Somerset.
Though I have not read every word of the thread, it seems to me there is one major mis-requirement and that is to assume that you need high voltages to reproduce clipping or even your "chipping away" at a waveform. You can always amplify whatever shape it is you reproduce and it is not necessary for you to push an op amp into clipping to get that part of it. In fact, an op amp is about the last thing you should think of using to reproduce clipping since they have huge open loop gain and will try to correct the errors in a way completely apart from how a standard PA BJT amp would manage with a gain of perhaps 20,000. To do it this way perhaps use an old op amp like the uA741 which has a gain of perhaps 100,000 (still greater than most power amps). Or perhaps set up a 'dummy' push pull stage with low rail voltages and throw in some capacitors to slow it down.

Which brings me to the next thing, that whatever shape you create can always be amplified, meaning that you can have the sound you are after at any level. Although, of course, this is not what actually happens (or happened) with those old guitar valve amps, so some authenticity will be lost without the actual output level, it does mean that you don't need a 200m "no go" area around you to experience the effect. To get this right your actual power amp must be unimpeachable up to these levels.

Personally I would look at this theoretically to see what what the harmonic spectrum is for a clipping guitar amp. This can be done and people like Yamaha will have detailed data of what goes on - you will, after all - only be reproducing one potential instrument out of all those they have researched and incorporated into their synthesisers or other instruments. Where I would start is looking at the content of harmonics that arise with straightforward clipping - ie. taking a straight slice off the top of a sine wave. I don't know whether this stays in identical proportions (though not quantity) the more of a slice you take, but my first suspicion is that's likely. Then I would look at how differently valve amps do this (though why you want a valve sound rather than something new - and don't want to use valves - I don't know). These will be directly mimicable with capacitors and resistors in various places because we know they can be modelled. (Incidentally, use LTspice, get some valve models and you can probably match the output waveform to a pretty close degree of accuracy.)

Also I would concern myself with how the power amp drives your speaker(s). This will have a very important effect on the sound. Valve amps have much higher output impedances, sometimes several ohms, and they don't do it quite the same way. You can do it with just a high powered resistor in series but personally I would do it with some degree of current drive (which will still require at least a wirewound R). I wouldn't be in the least bit surprised if this is the part that you feel is missing in what you have got and that you are much closer than you think. Series resistors do all sorts of nice things like extending the frequency response of a driver upwards in a very natural way and reducing distortion due to changes in parameters in the driver. Current drive does much the same thing, and more so, though will change the frequency response quite a lot since, when done fully, your response will then be changed by a function of the impedance response of the speaker. Generally this is not a great idea at low frequencies where the LF resonance can show up in its full glory, which may be 5x the normal output. The answer to how much you can bear in your application is to have a variable resistor in that bit of the feedback circuit. (It may not be easy to find a sub 10 ohm variable resistor if you are doing it on the output but the same thing can be done with larger values earlier on, but with added complexity. At least I think so.)

Finally, may I mention that the course you are on is in fact much the same as the course that most of us audiophiles have been on for a while, even though you may think you just want the solution to a particular problem. Finding 'that sound' that satisfies us is an almost universal quest; whether as a musician, an instrument maker (you) or a hi fi designer. It is a noble quest and you'd be surprised how much the solutions all end up in the same places.

Last edited by ChristianThomas; 14th March 2013 at 02:38 PM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Connecting Multiple op amp inputs from a single op amp output ZLyzen Analog Line Level 7 8th August 2012 01:17 PM
About Op-Amp input range xiaonan Parts 8 24th October 2006 07:49 AM
OP-AMP without an output stage? bigparsnip Solid State 19 14th February 2006 07:21 AM
OP-AMP output impedance hugeli60 Solid State 6 20th February 2004 06:56 PM
op amp supply range Bernhard Solid State 4 21st January 2004 01:35 PM


New To Site? Need Help?

All times are GMT. The time now is 08:46 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2