Discrete guitar preamp

[suzyj]

Bugga. The Mouser price looks a bit better, but the freight cost doesn’t.

I generally wait until I've got a bunch of things to order, then order in one go. Same sort of thing goes for Digikey.

The preamplifier I’m designing is a completely over the top design to go with the completely over the top hybrid 1kW rms class A stereo amplifier I’m currently designing/building for my lounge room

Blimey - I wonder how many tons of CO2 that'll be responsible for

I’ve had a read through your power amp page. Nice job! Your comments on keeping the input sensitivity low to take advantage of the preamps output swing has given me a good idea. I’ve decided to do the same thing and whack my power amps sensitivity right back to 5 or 6 Vrms. The lower closed loop gain will improve the power amplifiers phase margin, meaning the loop compensation can be increased in frequency, giving lower high frequency distortion. Lowering the closed loop gain will also increase the ratio of open-to-closed loop gain, thus lowering distortion throughout the entire audio band.

Thanks for the compliments. I'm a bit confused though - I wouldn't have thought that changing the feedback ratio would have any effect on phase margin.

Certainly the ratio of open loop gain to closed loop gain (effectively increasing NFB) does improve distortion though.

Cheers,

Suzy

 

[runebivrin]

Rune,
yeah, I wouldn't have need to explain this to you.
I suspected maybe you had the knowledge.

But no regrets for writing, once more, one mini-tutorial
There are more silent readers here around!
than we will ever know ....


Nice house ... Stockholm - neighbourhood ??

An externally hosted image should be here but it was not working when we last tested it.

We're veering off topic, but you should have an answer. I'm in Huddinge, south of Stockholm.

Rune

(Oh, and I wish the forum pages would specify encoding. I used windows, you apparently used UTF-8)

 

[GK]

I generally wait until I've got a bunch of things to order, then order in one go. Same sort of thing goes for Digikey.

Yeah, that sounds sensible, but I’m far too impatient. I want it now!

Blimey - I wonder how many tons of CO2 that'll be responsible for

But in winter I’ll be able to turn my heater off

Thanks for the compliments. I'm a bit confused though - I wouldn't have thought that changing the feedback ratio would have any effect on phase margin.

D’oh! Sorry, my brain wasn’t engaged - I got my power amp mixed up with my preamp. Trading off a lower closed loop gain in the power amplifier for a higher closed loop gain in the preamplifier will increase the phase margin of the preamplifier. The phase margin of the power amplifier will be decreased.
The phase margin increases when the closed loop gain is increased because the higher the closed loop gain is, the further it is from the point at which the open loop gain begins to roll off at more than –6dB per octave.
Having the closed loop gain set above the point at which the open loop gain begins to roll off at more than –6dB per octave (a single pole, or time constant) is the requirement for stability.
Suppose the open loop frequency response of a given amplifier has a single pole roll off of –6dB per octave from some point, with a second pole kicking in at a gain of 10, giving a combined roll-off from then on of –12db per octave. If this amplifier was operated with a closed loop gain of 10, the phase margin would be zero degrees and it would not be stable.
This amplifier can be stabilised in either of two ways –1) Sufficiently increase the closed loop gain to give a safe phase margin. 2) Increase the value of the dominant pole setting capacitor so that the –6dB roll-off extends sufficiently below an open loop gain of 10.
Option 2 decreases the high frequency response. Option 1 decreases the open-to-closed loop gain ratio.
The scenario outlined here is generally the case for high speed opamps not internally compensated for unity gain stability.
For instance, the data sheet for one such device may indicate that the device is not stable for closed loop gains lower than 20. This is because the open loop frequency response begins to roll off at more than -6dB per octave at not much lower than a gain of 20.

Looks like I’m going to have to do some simulations / experiments to find which pre/power amplifier closed loop gain ratio gives me the best overall compromise in performance.

Cheers,
Glen

 

[lineup]

(Oh, and I wish the forum pages would specify encoding. I used windows, you apparently used UTF-8)

I have one brother in storvreta and one sister in bromma.
I noticed, too ... that my characters: å ä ö Å Ä Ö does not display correctly
unless I change my character encoding in Firefox 'View' menu, and these pages of diyaudio.com while using vBulletin Forum commericial software DOES NOT specify any encoding. I am sure this wouldn't happen if use phpBB free & open source software for this board.
Or some other nice free forum php script .. like www.punbb.org made by a fellow from Sweden!
I have my browser set to use ISO-8859-1 as default if nothing specified.
cheers! /lineup

 

[suzyj]

The phase margin increases when the closed loop gain is increased because the higher the closed loop gain is, the further it is from the point at which the open loop gain begins to roll off at more than –6dB per octave.

I'd always thought that the phase lead cap ensures that the phase is essentially flat out until just before the roll-over point. Here's a typical pic (taken from a sim of my power amp):

An externally hosted image should be here but it was not working when we last tested it.

Here, unless the closed loop gain is less than ~20dB (in which case stability becomes an issue due to the rapidly changing phase), the phase margin stays pretty much the same, at ~80-90 degrees.

Of course without phase lead, it's a completely different story.

Cheers,

Suzy

 

[GK]

I'd always thought that the phase lead cap ensures that the phase is essentially flat out until just before the roll-over point. Here's a typical pic (taken from a sim of my power amp):

An externally hosted image should be here but it was not working when we last tested it.

Here, unless the closed loop gain is less than ~20dB (in which case stability becomes an issue due to the rapidly changing phase), the phase margin stays pretty much the same, at ~80-90 degrees.

Of course without phase lead, it's a completely different story.

Cheers,

Suzy

Okay, with a dominant pole of –6dB per octave the phase margin will always be 90 degrees by virtue of the fact that a single pole can only contribute 90 degrees of phase shift. At the critical point of a total loop gain of unity the total loop phase shift equals 270 degrees; comprised of the 180 degrees contributed by the inverting amplifier topology plus the 90 degrees of phase shift incurred from the single pole open loop roll-off.
When a second pole inevitably kicks in, below unity loop gain, the total phase shift doesn’t jump from 270 to 360 degrees though. Progressively lowering the closed loop gain beyond the point at which the second pole kicks in progressively reduces the phase margin from 90 degrees to zero degrees.

You are correct though – beyond 90 degrees phase margin, increasing the closed loop gain cannot increase the phase margin beyond 90 degrees. However, if you are running at a closed loop gain higher than that required to achieve a 90 degree phase margin, your amplifier is overcompensated and its slew rate is being unnecessarily sacrificed.


Cheers,
Glen

 

[GK]

At the critical point of a total loop gain of unity the total loop phase shift equals 270 degrees; comprised of the 180 degrees contributed by the inverting amplifier topology plus the 90 degrees of phase shift incurred from the single pole open loop roll-off. When a second pole inevitably kicks in, below unity loop gain, the total phase shift doesn’t jump from 270 to 360 degrees though. Progressively lowering the closed loop gain beyond the point at which the second pole kicks in progressively reduces the phase margin from 90 degrees to zero degrees.

D'oh, stupid mistake - that should read:


At the critical point of a total loop gain of unity the total loop phase shift equals 90 degrees; comprised of the 180 degrees contributed by the inverting amplifier topology minus the 90 degrees of phase shift incurred from the single pole open loop roll-off. When a second pole inevitably kicks in, below unity loop gain, the total phase shift doesn’t jump from 90 to 0 degrees though. Progressively lowering the closed loop gain beyond the point at which the second pole kicks in progressively reduces the phase margin from 90 degrees to zero degrees.


Same thing, just the other way 'round.

Cheers,
Glen

 

[suzyj]

Yup, that makes sense. I was thinking more for the case where the amp is working well away from the roll-off point, so over compensated.

I suspect a lot of the hoodoo that global negative feedback has in audiophile circles comes from poor understanding of the principles of compensation.

Cheers,

Suzy

 

[Dominique]

Hi,

I've just looked back into the thread, because I am thinking about building a guitar preamp with an op-amp.

(Well, I could try to build that same circuit, but I rather like to be able to overview what I build and that discrete circuit is too much for a low-level diyer like I am)

I see that after searching for the suiting op-amps, suzy's solution is way better, of course.
But how much noise does that discrete circuit produce? Does anyone know?

Cheers,
Dominique

 

[Radioman62]

Keep it simple?

I have just skimmed this thread and I found it when I searched something about FET's

Anyway, from the beginning there was a wish of a low output impedance from the guitar.

I have built a unit I have inside my guitar from a very old article in the Magazine Practical Electronics. It has worked perfectly well for me the last 28 years. I have just switched the OP to a better one
If you want it really simple, go for the circuit at the end of this article. A couple of resistors and one FET, thats all !
My cousin did the bare bone one to fit in his Gibson SG ... 28 years ago. Works perfect as well.

The link to the article is a page on my own homepage-site:
Switched tone treble boost

 

[nigelwright7557]

Hi Suzy

It may be a matter of taste, but personally I'd go for something simpler.LV

I usually use an op-amp cct for this type of application.
It gets rid of a lot of discrete transistors.

 

[lineup]

Very good option, G.Kleinschmidt.
In fact, such a buffer is seen more than often! in good Op-Amp internal schematic. Used as the output buffer for very good performance (chip-)amplifiers. It is a design element (part of amplifier) that is proven to work very well. So that is why it is so commonly used.

lineup

Hey, that's a pretty cool circuit. Looks like a great replacement for the follower on the back end of the diffamp. I just knocked one together with LTspice. It drives to within 0.7V of the rail, has a nice low output impedance and high input impedance, and 0.7% THD...

Eight bipolars and a pair of matched JFETs... Where will it end!

Cheers, Suzy

G'day, Suzy Jackson!!!


I was from start thinking of using some lownoise BC550C / BC560C.
At my second thought, I would lean towards suzyj way:
... using lownoise JFET.

GKlein (G.Kleinschmidt) has posted some really good circuit!
But The thing is, we want rather low current consumption, for our batteries.

I be back with my Circuit of The Day
Doing some last minute Spice Sim testdata verifications, first.
Stay tuned!

Lineup - Master Of Lineup Audio Labs

 

[guoming]

look like a toy.