Solid State Amplifiers That Have Soft-Clipping

[Dusk]

Hi All

Iam looking to find out about diy solid state guitar power amplifiers that when driven pass their SOA will clip soft instead of the nasty ss clipping. I know there are some HI-FI units that use limiter-circuits at the input of the power amplifiers , that are preset to reduce the gain, once it pass a set input voltage(anyone have a simple schematic for one?),but iam looking for something like the Randy Slone Optimos . This Optimos power amplifier has a vas stage that is dynamically bias,that is it has drivers for the output mosfets that once they are overloaded will operate in their knee, non-linear area(iam not a amplifier Guru......i just try to play guitar)
and send this output to the output mosfets, henc......soft-clipping.

[teemuk]

You could try i.e. LDR or FET based limiter (VCA) in the input and sense the output signal. All differential input power amps are essentially just high power opamps so you could try similar tricks than with them. For example, soft clipping in Vox Busker: http://www.schematicheaven.com/voxamps/busker.pdf
Usually it's more effective to deal with signal limiting in the input stage since it operates on lower currents. I never experimented with limiting the voltage amplifier stage.

[Dusk]

Hi Teemuk

Is this Vox circuit the whole power amplifier,or just the input stage of thepower amplifier? also if it is just input can it be use with any any amplifier? will this circuit add noise to the amplifier output

[teemuk]

The power amp is the part in the upper left. The IC2 is type TDA2030 so it is just an opamp that is capable for high output currents = "chip amp". As so, the circuit would work to higher loads as well but why bother? The gain of this circuit is about 100 and the diode resistor network works only by limiting the gain when the output voltage is getting near the rail voltages. It's pretty similar to typical "soft" limiter configuration that use two parallel diodes in the feedback loop - except that in this configuration the input can swing higher than the forward voltage of the diodes.

I have never tried this circuit - except from simulating it with pspice. Surely requires some tweaking to get it running from higher supply rails. Works as a preamp limiter as well but naturally the circuit has to be attenuated to limits of power amp's input sensitivity. If attenuated the SNR will increase - as is the case with every passive circuit.

Another form of limiting using feedback is shown in:
http://www.fender.com/support/amp_schematics/pdfs/Steel_King.pdf
It's a large schematic but see block B-1 on the first page. The circuit is labeled as "limiter".

Personally I try to avoid fiddling with impedances of the feedback loop since this might introduce DC offset. This is perfectly acceptable in the input stages but not in the power amp. Rectifying an attenuated portion of the output signal and using the outcome to drive either a led in LDR circuit or a gate of a FET acting as resistor seems less "risky".

I have designed some conceptual schmatics of LDR applications that detect the power amp output signal but they are quite crude. The problem is that all circuits like this have to be designed for a certain application: A circuit for amp with output swing of +-10V requires different component values than circuit with +-30 output swing. Basically, the detector is the heart of such circuits so better the "detector" is the better the circuit performs. Using a crude circuit is not very "HiFi".

Here is one example of more "HiFi" solution of power amplifier's detector/limiter circuit:
http://www.fender.com/support/amp_schematics/pdfs/KXR_60_Schematic.pdf (bottom right, first page).

I hope I was able to provide some ideas.

Do you happen to have a schematic for that Slone circuit? I'm always interested in solutions I'm unfamiliar with...

[Dusk]

Hi
Thanks for the info.....i have Slone's book " The Audiophole's Project Sourcebook"....but i don't think i can post it up here because it is copyrighted.....sorry

.....on second thought i think someone has put one  on the diy audio for fanatics......just do a search on Optimos........iam just trying not to break any ones rules.
I think the limiter-LDR type gain reducer would be best because it could be set for a gain level at the input of the power amplifier,that won't let overdrive happen.....iam no amplifier GURU.....but do you know where i can find a simple LDR/DIY circuit......THANKS SO MUCH

[teemuk]

First hit on google "ldr limiter": http://sound.westhost.com/project53.htm

Edit: Found the schem: Pretty nice idea... the VAS stage is not just a typical common emitter but a cascode (Slone calls it a "pseudo cascode") that varies it's transconductance according to difference between output and rails. In other words, the transistor's gain gets lower when the output is swinging close to rails. Works with BJT output devices too.

The only setback is that this requires symmetric VAS stages (thus a symmetric "cascoded" input stage as well) since one VAS can detect difference to only one of the rails. The designs tend to get pretty complex with that stuff included... Anyway, the "soft limit" effect is pretty subtle (did a pspice simulation), it's closer to soft clipping than lowered gain effect of a dedicated opto/FET limiters circuit. Interesting solution anyway.

[Dusk]

Ok
This one.....seem that you have to keep making ajustment for different output wattage andspeaker ohm impeadance.......is there one that just reduces the input gain.... without reguard to speaker impedance or output wattage......iam more of a guitar player than amplifier GURU.......THANKS for all the info

ps i have noticed that more Bass (ss amplifier) use limiter..why is that?..seem like guitar amplifier (ss guitar amplifier would need this limiter more?

againTHANKS

[teemuk]

Well, you can't get away simple here: You can move the detector circuit to input side but it still has to work according to a specific signal amplitude and the problem is that among different amplifier circuits the amplitudes are never completely identical. As I said before it depends on the power output/signal amplitude. If the detector is on the input side, it has to work according to input sensitivity of the power amp (max input signal before power amp clips). You have to tweak those circuits, that's the point.

I have noticed the same thing as you: Bass amplifiers have more of the stuff that should be in guitar amplifiers as well: Limiters, clipping indicators, graphic EQs, FX loops, line out sends etc. They are generally designed better than guitar amps too. I really don't know why though... maybe an average guitar player settles for a more mediocre amplifier than an average bass player.

One of my guesses for the reason is that many guitar amplifiers are stuffed with FX circuits like distortion, nowadays all that modeling stuff too. Sometimes it's good but I'd rather pay for those limiters and graphic EQs and leave FX stuff outboard. Usually higher priced amps tend to have all those things that bass amps do so it is a price issue as well - and people that buy cheaper amplifiers think they need those "bells and whistles" more than they need an output stage limiter.

I personally don't like amps that have too many effects because this usually means that they lack features that are more important. Anyway, spring reverbs are nice.

Guitarists also tend to be more narrow-minded: Why do you think that so much of the stuff still uses inventions from 50's and 60's. Why so many tonestack circuits are still based on the same passive topology that Fender used in 60's while active topologies utilizing series resonances (i.e. graphic equalizer) perform so much better: You can dial in the same response with the latter and it has better SNR and less interaction between controls. Why no biamping guitar amplifier circuits that would use a digital amplifier for low frequencies? Why using different cabinet structures as a method to shape guitar sound was "invented" around 90's when the theories involved have been known since 1940's and used widely in HiFi/PA/Bass applications? Why so many guitarists are still uneducated about speakers system vs. output power theory? Why not that stuff used in bass amplifiers? Narrow-mindness, that's why.

A simple amplifier might be good when speaking about tubes but that is not a case with SS stuff. One common factor to nearly all famous SS amplifiers is that their design is somewhat "unorthodox" and complex when compared to typical stuff out there. Check Acoustic Control Corps's old amplifiers or Norlin's Lab series for example. Lot's of innovative stuff is used in them.

[Dusk]

Yes
I see you havedone your homework on this ....i alway hear about how ss amplifier don't compare with tube amplifier as far as overdrive sound. I am always wondering way not eliminate the bad overdrive in ss units by using limiter and using tube pre-amplifiers to get the so-called tube overdrive...i see Marshall has an the Mode Four this is  hyrid amplifier tube/ss power amplifier 350watt i have heard it and too me it sound good imho,but people say it's not power tube overdrive.....beats me? by having a Chip amplifier  for an output you save on power tube replacment.
The Acoustic amplifiers you mention i know that the great Albert King used them in the 70's and80's and to he sounds great..a non-amplifier GURU

PS
i Know that you could use a variable pot to set the input level on limiter  on the input of the power amplifier....you would not have to worry about different ohm speaker or wattage levels......Thank

[teemuk]

I believe that a great part of that "tube power amplifier" tone relies in the interaction of high output impedance tranformer coupled output stages and the speaker. Solid state amplifiers generally have a low output impedance. If you want to play a bit with circuit theory you can think of the output stage as two series resistors fed by a power supply: The bottom one represents the speaker and the top one the output impedance. Speakers do not have a constant resistance throughout the audio band; the resistance rises (sometimes as high as 100 ohms) at their resonant frequency and towards the higher frequencies. Pick any value for output voltage and consider what happens to current, voltage and power over bottom resistor (Rspeaker) in following situations:
Routput = 0,1 and Rspeaker = 4 and 20 (latter in resonance and HF), this is for SS amps. Then Routput = 4 and Rspeaker = 4 and 20, this is for tube amps.
This is rough generalization of SS and tube amps and not all behave like this. Don't believe all tube hype. In most cases a great part of creating a decent overdrive tone is the shaping of frequency response - both post and pre clipping. Keep in mind that the tone is never created by the active device alone - it is always the product of the active device and the surrounding circuitry.

By the way, that ESP circuit is pretty simple: It's just controlling a transistor according to rectified sample of output signal. The transistor uses leds as it's collector load. Here is another example: http://www.picks.force9.co.uk/circuits.htm
and here:
http://www.silonex.com/s_images/audio/limit/fig1.gif
Here's example of a very crude circuit:
http://machines.hyperreal.org/categories/DIY/info/compressor
Here's another compressor idea based on varying FET resistance:
http://freespace.virgin.net/ljmayes.mal/comp/comp.htm
As you see, the detector circuit in this one is very similar to one in LDR circuits.

[Dusk]

Well....mmm

I noticed that in the http://www.picks.force9.co.uk that the max input is 3mvrms and max is 2.6rms,now this unit would be placed between the output of the preamplifier and theinput of the power amplifier....ok, with it's input andoutput rms voltage would the limiter componet have to be changed....remember ....iam not an Amplifier Guru


also i have seen adiscreet transistor rig to the output of a ss (guitar)amplifier. It was tired to the input differential somehow to mimic a high impedance output i think i saw a ESP web site on a 100watt (diy)guitar amplifier...don't really know how it worked

as alway THANKS

THANKS

[teemuk]

What do you mean by having limiter component changed? The circuit at http://www.picks.force9.co.uk/circuits.htm controls the gain of the opamp circuit by using a light dependent feedback resistor. This reacts to light emitted by led.

The variables you need to focus on these kind of circuits are the signal level required to start illuminating the led and the resistance of a particular LDR in dark and lighted conditions. For example, you can build the circuit but it will not work ideally unless you use LDR with same resistance figures. For different LDR models the gain set of the opamp stage would have to be altered.

You don't have to be an amplifier guru to deal with this stuff - I certainly am not - but it helps A LOT to know basic circuit theory.

Regarding the other topic, the ESP guitar amplifier mkII uses a variable output impedance circuit (known also as mixed-mode feedback) and so do many other guitar amplifiers - it's a quite common trick these days. The concept is also used in driver amplifiers of spring reverb circuits. The site (ESP) contains a very good article about the idea.

[Dusk]

......Oh

I just reread your reply on the Slone Optimos amplifier.I don't have any Pspice simulation...i would not know how to even use it......but you say it shows that the Optimos is closer to soft-clipping than the Opto/fet limiter circuit....mmmh maybe i should just get a pc board of the amplifier and make one.....you say it really round thepos and neg output....and you can see this with that simulation you have? I only use tube preamplifier with my power amplifiers,i have a Hafler pro-500watt power amplifier/250per side....i know this is overkill but the Hafler is the same type of circuit symmetrical differential input but not the  same VAS cascode that Slone use

[teemuk]

Yes it's closer to soft clipping since the circuit clips to rails. The decreased gain just backs off the worst effect of this a little bit causing the clipped signal to have a softer "knee". The LDR limiters usually stay cleaner since they attenuate the input signal and the amplifier never really has to push it up to the clipping point. The only noticeable distortion in them is due to excessive "overdrive", which compresses the wave shape so much that it gets audible. There are also some speed issues with those circuits that may cause a little distortion. If considering "softness" of - can't talk about clipping here - then a properly working LDR limiters are much softer. The difference: Slone's circuit clips but does it fairly softly, Limiter: prevents clipping by attenuating waveform peaks.

I have tried to acquire Slone's book but haven't yet bumped into it anywhere. I don't want to order it via internet - I'm a bit sceptic about that stuff. Anyway, I have D. Self's book about power amplifiers and it's pretty good. Recommended reading for anyone interested about the topic. At least some of the Slone's circuits are based on the topologies presented in that book.

[R.G.]

Just curious - since a SS amp is great at just doing what it's told at anywhere less than clipping, why wouldn't you do an entire soft clipping circuit ahead of the power amp and then just feed that to the accurate power amp?

The soft clipper only has to soft clip the input at a voltage less than would drive the power amp to clipping. As long as that happens, the power amp is completely unaware what's happening to it.

As long as a SS amp never clips, it never sounds ugly, just accurate. It's possible to follow a soft clipped signal accurately.