All transistor Rat

[Fancy Lime]

As long as we are talking about impedance bootstrapping, sure. But there is also gain bootstrapping. The reason why I called the 250 biturbo "biturbo" is, that it uses both: C4 impedance bootstraps Q1 and C6 gain bootstraps Q2 (numbers according to the first schematic in the thread). Unless you object to calling that bootstrapping, but if the nomenclature is good enough for Tim Escobedo, it's good enough for me. The technique is not exactly new, though. In fact, Q2+Q3 in the biturbo are essentially a Jordan Bosstone with DC decoupling in the bootstrapping path, so that the gain control (which actually controls the amount of bootstrapping) does not mess with the bias of Q2.

Andy

[Vivek]

https://www.diystompboxes.com/smfforum/index.php?topic=95356.0

Bootstrapping for gain

Maybe we can call it BOOSTrapping

https://web.archive.org/web/20100330091933/http://folkurban.com/Site/BootstrappingforGain-692.html

[PRR]

For more gain, you can increase R5 value (or even delete it - which isn't recomended for stability issues..)

If C15 is "large", R5 has very little effect on audio gain. As you say, R15 is important for DC stability. Even then, this is not a plan for beginners. JFET variability makes trouble.

Any of the externally-compensated chips (now rare) can be over-compensated to "slow" or "dull". '301 can beat 70V/uS one way, struggle for 0.1V/uS another way. It is hard to know if this matters for e-guitar. If you decide to experiment, keep compensation cap leads very short (no switching).

Bootstrapping for gain is an old-old sin, distantly related to the original radio sin of regeneration.

[antonis]

C6 gain bootstraps Q2 (numbers according to the first schematic in the thread). Unless you object to calling that bootstrapping

Not at all..
As you said, "boostrapping" is a designation for raising apparent value..

If I recall it right, Collector resistor bootstrapping (at least for BJTs) was firstly mentioned (albeit erronously*) in T.D.Towers https://www.keith-snook.info/wireless-world-magazine/Wireless-World-1968/High%20Input-Impedance%20Amplifier%20Circuits.pdf, figure 3 b. & d. , figure 4 a. etc

(*) Can't be obtained via a single trasistor..
P.A. Johnson https://worldradiohistory.com/hd2/IDX-UK/Technology/Technology-All-Eras/Archive-Wireless-World-IDX/60s/Wireless-World-1968-09-OCR-Page-0019.pdf

P.S.
Didn't realise you refer on Q2/Q3 CFP..

[antonis]

Even then, this is not a plan for beginners. JFET variability makes trouble.

That's why I recommened to replace Q1 with a BJT (n-p-n)..

I've the feeling that Mitchell is prone to mixing/modifying building blocks troubles..

[GFR]

I've found this graph with a rule of thumb about distortion in opamps:

https://eng.libretexts.org/Bookshelves/Electrical_Engineering/Electronics/Operational_Amplifiers_and_Linear_Integrated_Circuits_-_Theory_and_Application_(Fiore)/05%3A_Practical_Limitations_of_Op_Amp_Circuits/5.04%3A_Slew_Rate_and_Power_Bandwidth

The graph above makes some assumptions that are mentioned elsewhere in the article

A) The power supply is +/- 15V

B) Therefore the maximum unclipped Vp that we can expect is 13V (assuming 2V dropout from rail)

C) The slew rate is 0.5V/uS  and expected output is 12Vp <<<<<<<<< These are the important specs that give the 12v/6631Hz data point


If any of above change, the graph will change.


Hence that graph is not a rule of thumb for all Opamps and all circuit configurations.

Hence that graph is not a rule of thumb for all Opamps and all circuit configurations. => Of course not. Given a specific gain / frequency response / slew rate / supply voltage etc. you can sketch a similar graph and estimate when that circuit would be clipping or distorting due to SR limiting.

[soggybag]

Here are a couple evolutions of the original idea. The first uses a MOSFET. I figure this has plenty of gain by itself to drive the rest of the circuit. Not sue if I got the R8/C4 and R9/C5 filter gain control correct looks close.




I looked up the diode compression op-amp and shoehorned it into the circuit here.

[Fancy Lime]

Yet another option: Joe Davissons CCS Drive:

I like this one a lot. Works well with the diodes to ground instead of in the feedback loop. Capable of impressive gains if the lower transistir is high gain or a MOSFET. Put a buffer right after the gain stage for even more gain.

Andy

[Rob Strand]

The gain resistors on the RAT are set for high gain.   In order to actually get that gain the amplifier part is going to need a lot of open-loop gain.

iainpunk, recently posted this one.
https://www.diystompboxes.com/smfforum/index.php?topic=129311.0

Just after that I was looking for stuff on the Danelectro FAB D-1 (see Lounge) and I found this mindjogger.
Ibanez used all transistor discrete opamps in a few pedals; from C6 to C10
(a cheaper version of Boss's JFET + BJT hybrid):



All these circuits can be 'slowed down' by putting a cap across the C and B of the final transistor.
That might help emulate the slow LM308A.   The Ibanez circuit has C7 but it might need to be
increased.   The gain control and feedback resistors on the Ibanez circuit and be re-jigged to be
like the RAT.

These types of circuit might just have enough gain.  (That's not to say the bootstrap circuits won't.   Someone needs
to check if the gain can be met.)

The bottom line is work is required to get it to sound the same as a RAT.

[antonis]

The first uses a MOSFET. I figure this has plenty of gain by itself to drive the rest of the circuit.

Don't know GAIN pot value but make R5 = 1M & R4 = 1M8 - 2M2 as a start..
(max gain should be slightly under X30)

[soggybag]

Thanks again for the feedback Atonis. I wasn't sure about the gain pot value but figure it should be C5k.

The gain pot makes sense to me. My understanding of the R8, C5 and R9, C5 network is a little vague. Let me try and explain what I think is going on here.

R8 and C4 attenuates frequencies below 1.5KHz. R9 and C5 attenuates frequencies below 60Hz. This would be affected by the gain pot setting but I'm ignoring that. The affect of R8 and C4 is to decrease gain for signals below 1.5kHz. R9 and C5 cut the gain again for signals below 60Hz.

Can I ask about the values you chose?

I had planned to use 10M for both R4 and R5. This should put the gate at 4.5v. You are suggesting 1M for R5 and about 2M for R4. This would put the gate around 6v. What's the difference here? I commonly see this circuit block biassed at half the power supply.

Second question, the change in these two resistors must affect impedance! Creating a voltage divider here to set bias voltage could generate the same voltage with larger or smaller resistors, what's the difference?

[Vivek]

In the end,

would an all transistor Rat sound like an all opamp Rat sound like an all FET rat ?

Because sound depends upon transfer characteristics and interstage frequency response

and if each above has same transfer characteristics and interstage frequency response, they should all sound same, no ?

[antonis]

I wasn't sure about the gain pot value but figure it should be C5k.

No problem but you'll result into about 10% (or more) useless rotation..
Q1 gain is roughly set by R3/GAIN pot ratio so for pot set FCCW gain should be less than unity..

The gain pot makes sense to me. My understanding of the R8, C5 and R9, C5 network is a little vague. Let me try and explain what I think is going on here.
R8 and C4 attenuates frequencies below 1.5KHz. R9 and C5 attenuates frequencies below 60Hz. This would be affected by the gain pot setting but I'm ignoring that. The affect of R8 and C4 is to decrease gain for signals below 1.5kHz. R9 and C5 cut the gain again for signals below 60Hz.
Can I ask about the values you chose?

For R8 & R9 values 1% & 10% of GAIN pot respectively, you realize that there isn't "correct" R/C values combination..
(they are strongly dependent on GAIN pot setting..)

I had planned to use 10M for both R4 and R5. This should put the gate at 4.5v. You are suggesting 1M for R5 and about 2M for R4. This would put the gate around 6v. What's the difference here? I commonly see this circuit block biassed at half the power supply.

It depends on Q1 Source resistor value (GAIN pot)..
BS170 has a threashold voltage between 0.8 and 3V (for I_D=1mA)

Worst case scenario is for V_TH = 800mV so we have to add max signal amplitude (like we do for BJTs) or we'll result into negative waveform clipping..
For a grounded Source Mosfet, 1.5V to 2V V_GS should be fine but here we have to add Source voltage due to channel current..
...................(deleted)
To make long story short, Mosfets are biased likewise BJTs but with their particular V_TH instead of V_BE..

Second question, the change in these two resistors must affect impedance! Creating a voltage divider here to set bias voltage could generate the same voltage with larger or smaller resistors, what's the difference?

In case you're talking about voltage divider (trimpot) with Gate big biasing resistor, it's OK..
In case of voltage divider like R4 & R5, equivalent impedance is the parallel combination of R4 and R5/(stage gain +1)..
(for a gain of 30 say, R5 is seen at the Gate as about 323k..)

P.S.
An antutored person could hold that Q1 stage gain is set by R5/R1 (80dB!!) which clearly doesn't stand 'cause open-loop gain is lower than closed-loop one.. 

edit: I'd start with below Q1 configuration:

[iainpunk]

i know im late to the party, but i think the ''discrete opamp'' described in this thread would be a nice option to replace the LM308 in a RAT. the gain stated in the first post isn't accurate, as i originally made a mistake while testing it, had both PNP in reverse.
https://www.diystompboxes.com/smfforum/index.php?topic=129311.0

slewrate

at the voltages were working with, the slew rate of the 308 only affects frequency's a guitar amp can't meaningfully reproduce. frequencies above 16kHz if the clipping diodes are discarded, and over 100kHz if they are taken in to consideration. that a wave is slewed doesn't mean we hear a difference. ive been experimenting with slew rate limiters a lot lately, and it starts to be noticeable in the top end around 0,1v/us. asymmetric slew rate also has a really cool sound to it!

cheers

[soggybag]

Thanks again Atonis! I spent yesterday considering your suggestions. I’m trying to wrap my head around the math!

Looks like the gain pot, not considering R8/9, C4/5, is R3/Gain or 4k7/1K-0. At minimum gain that would x4.7. At max gain we run into 9v limit. I’m sure how to call that, let’s just ballpark it at x50.

With R8/9, C4/5 these are in series with the gain pot. Since R8/9 are parallel I can look at these as something less than the smallest value, call it 43r.

I’m trying to come up with the numbers you have. With the gain pot at min I get a gain of 4700/1043 or x4.5 at max gain I get 4700/43 or x109.

That doesn’t consider C4/5.

I’m guessing your max gain calculation is plotting the gain by frequency for the filter created by C4/5.

[mac]

Almost, it's Joe... Davidson's? Diode compression discrete op amp.

My notes on Joe's DCDO

https://www.diystompboxes.com/smfforum/index.php?topic=44255

mac

[antonis]

I’m trying to come up with the numbers you have. With the gain pot at min I get a gain of 4700/1043 or x4.5 at max gain I get 4700/43 or x109.
That doesn’t consider C4/5.

But you should..!!

Just replace C4 & C5 with their capactive reactance (X_C) at specified frequencies, calculate  1K // (R8+X_C4) // (R9+X_C5) equivalent parallel resistance and add in series 1/g_m (about 5mS at I_D = 1mA but don't entrust it)..
That's by what you have to divide R3 to find the Gain..

[soggybag]

Thanks again! That’s a new concept for me to study: capacitive reactance!

[antonis]

JUst another mode for restless Mitchell..



A band-pass zone variable from 340Hz - 15kHz..(you can make it as narrow as you like..)
A 22k stopper resistor in series with 1M pot should make its full rotation more usefull..

(a couple of μV DC across 1M pot shouldn't make it crackle..)

[Vivek]

Can we put big cap on wiper so no DC on wiper ?