How do you replace JFET buffer with an opamp buffer?

[tatou]

I'm experimenting with a RAT variant (my work-in-progress schematic below). So far, it just adds a clip switch and lists a few R/C value variants I've come across for other people's mods.

I want to try out opamps other than the LM308, and since I several DUAL opamps laying about, I'd like to reuse the second side to replace the output buffer. I'm an electronics n00b and still learning, though, so I'm not entirely sure how to do this...

I read the AMZ article on buffers but there doesn't seem to be a 1:1 correspondence between the simple JFET buffer used by the RAT (gate resistor bias? to ground) and any of the omamp examples.

Any help or guidance is appreciated!

[PRR]

[antonis]

You MAYBE have some problems using one part of a dual op-amp for gain stage..

Dual op-amps DON'T have compensation pins availiable (pins on which C7 is connected) so, depending on the specific op-amp in use, Slew Rate and Gain-Bandwith product may be out of limits..

IMHO, even if you use some "standard" stompboxes op-amp (like TL072-TL082) which are by far faster than LM308 (13V/μs compared to 0.3V/μS of LM308) with a wider Open Loop Frequency responce, you'll still "loose" RAT's original sound..
(part of which is due to op-amp "collapsing" and non-linear "treatment" of high frequencies..) 

[tatou]

Thanks @PRR! This configuration was my first guess, but I don't know enough theory yet to convince myself that it would work... I'll breadboard that up and give it a listen!

@antonis — That may be OK. I'm not necessarily trying to build a RAT (I've already made a couple of faithful LM308 RAT clones), so this project is a learning experiment. That said... do you have any suggestions for mods if subbing TL072, for example, in a circuit like this? Tweaks to add different character, etc. (not just make it more RAT-like).

Thanks!

[ElectricDruid]

There are various things that are significant about the op-amp choice in the Rat circuit. The low slew-rate has already been mentioned. I don't know what the Gain-Bandwidth is for the LM308, but I bet it's low = more roll off of highs. The final thing is that it copes fine with having the output slammed into the rails. Some op-amps don't like this and can behave weirdly. In this circuit, there's no limiting in that gain stage except for the rails, so the behaviour of the amp at its clipping limit is important. Ok, it *might* be important - it probably won't sound that different after those diodes have mangled it!

It's a good place for a few experiments, so try a few different chips and see what you get. Good luck and have fun with it!

Tom

[antonis]

I don't know what the Gain-Bandwidth is for the LM308, but I bet it's low

Maybe lower than you bet.. 
(less than 1M for unity gain..)

For max gain of about 67db, roll-off starts at frequencies lower than 500Hz..
(e.g. at 10kHz, amplification is less than 30db..)

That may be partially "recovered" for a higher GBW product amp by setting apropriate value of LPF capacitor on NFB loop but not in a "linear" responce..
(because of variable LPF cut-off point due to variable gain pot setting..)

[merlinb]

Slew rate limiting? In a 9V, 5kHz system? I'll believe that when I see it

[ElectricDruid]

Slew rate limiting? In a 9V, 5kHz system? I'll believe that when I see it

I wouldn't rule it out. Check out the datasheet:

http://www.mit.edu/~6.301/LM108.pdf

There's a graph on page 4 showing a -5 to +5V square wave after it goes through the amp (with the standard 30pF compensation cap, like the RAT). The up slope takes 40uSecs, which is the period of a 25KHz waveform.

Ok, so it won't affect much at 5KHz, but don't expect any upper treble to come through untouched. It's basically adding more roll-off at the top of audio.

Tom

[PRR]

> Slew rate limiting? In a 9V, 5kHz system?



Of course naked guitar will be severely falling before 2KHz. And even grossly distorted, by an amp that can't slam >500Hz well, it may not slew; certainly be hard to tell by the time you had it cranked that far.

[BubbaFet]

Friends don't let friends replace JFET buffers with opamps buffers. Just sayin', friend. WHY? Because you lose that nice second harmonic distortion that JFET buffers supply. Yah, yah ... buffers are supposed to be 'transparent', but IME, a jfet buffer is anything but that.

[tatou]

In case anyone was following along at home:

I've actually breadboarded the modified circuit (suggested by @PRR) using a 4558, and AB/tested it against a MadBean Slow Loris build I did a while back using a Jazzmaster. To my ear, I can barely tell any difference... I've continued to tweak the caps to remove some of the shrillness of the RAT sound and make it fatter and/or fuzzier, with some pretty good success... I like how it's shaping up...

[ElectricDruid]

The 4558 is a good choice, I'd have said. It's another old-school low-spec op-amp. It's basically a dual 741, and that was designed in the late sixties, as was the LM308. Plus it also doesn't mind being slammed into the rails, which helps here.

Tom

[reddesert]

> Slew rate limiting? In a 9V, 5kHz system?



Of course naked guitar will be severely falling before 2KHz. And even grossly distorted, by an amp that can't slam >500Hz well, it may not slew; certainly be hard to tell by the time you had it cranked that far.

Is the y-axis of that graph "Output Swing (+/- V)" intended to be read as peak-to-peak, or 0-to-peak, ie +8 V to -8 V ? I usually read it as 0-to-peak, if so, the graph indicates 8V peak-peak is at about 4 kHz. The slew rate of the LM308 is spec'ed at about 0.3 V/micro-second, so it would take about 30 micro-sec to slew 0 to 9V, so a sine wave with a period of 60 micro-sec would be reproduced poorly - that's 16 kHz.  I went through this once and wasn't able to convince myself that the famous LM308 slew rate was really limiting its behavior in the guitar audio range.

[ElectricDruid]

It needn't be necessarily within the guitar's typical frequency range. One side-effect of the low slew rate is that it limits the amount of higher harmonics the amp can produce when clipping. Roughly, the waveform produced is more of a trapezoid, clipped-triangle affair, rather than a neat square.

Tom

[tatou]

The 4558 is a good choice, I'd have said. It's another old-school low-spec op-amp. It's basically a dual 741, and that was designed in the late sixties, as was the LM308. Plus it also doesn't mind being slammed into the rails, which helps here.

Tom

That was part of why I chose it... I knew it was a little "dirtier" than the TL072, for example, and a dual 741 equivalent.

Is there a good reference for single/dual/quad equivalents and a rundown of their different characteristics? As in some sort of reference table? I've found this page by David Morrin which goes part of the way there... but I wonder, for example, if there's a dual with roughly the characteristics of a LM308 out there (the same way 4558/741 and TL072/TL071 are comparable)?

[bool]

4558 isn't a dual-741. A 747 dil-14 IS a dual 741; and a 1458 is very close.

But none of 4558 opamps is a true 741.

[ElectricDruid]

Well I dunno what a "True 741" is - does taking the phase compensation off rule it out? Anyway, here's what Texas Instruments think:

"The RC4558 device is a dual general-purpose operational amplifier, with each half electrically similar to the μA741"

from

http://www.ti.com/lit/ds/symlink/rc4558.pdf

So at least for *their* 4558, it's "electrically similar" to the 741 (I think that's marketing speak for "we stole the design").

There's an article about the JRC4558 over on electrosmash which suggests that the dual version 4558 had a few improvements over the earlier generation 741:

http://www.electrosmash.com/jrc4558-analysis

Either way, straight copy or tweaked improved version, it's in the right ballpark/era for RAT duty.

Tom

[PRR]

> marketing speak for "we stole the design"

As you say later, they changed what they stole. Mostly to do just-as-good for less money. 45 years later, the changes are hard to spot; not any "Wow!" change. But in extremes (which may include this crowd) there may be differences.

[tsili]

Hello,
I am fresh to this forum.
So after 3 years is there a change that there is the schematic?
I mean the modded with the dual op amp. I would like to try it out.

[antonis]

Hi and Welcome..

Just delete Q1 & R12 and connect upper leg of R11 to op-amp non-inverting input and lower leg to VB..