SOLDANO SLO preamp analysis

[ltournell]

Hello,

Following our previous topic here , this is a more complete and synthesized analysis (or an attempt to ) of the SOLDANO SLO 100-50 and 30 preamp . Remember that the point was to discover the part of that strange 3rd stage in the schematics of the OD channel here below :

The LTSpice simulation was therefore made from this schematic in which the red arrows show where the loaded output of the previous stage was measured :

The various computations made from the load lines and the correct formulas give this table :

Remember the low cut-off frequency due to the cathode RkCk network is given by :

The purpose is to characterize the sounds produced and more particularly the nature of the harmonics of the overtones resulting from overdriving the stages. We can make out the rate of ODD harmonics and EVEN harmonics.

So to start with we simulate with very common guitar PUs : the 490R and 498T Gibson humbuckers for which an approaching value of the ATTACK level can be found in the next table :

For example we see that the 498T outputs the maximum level with the 6th string at 1 Vpp , so the first stage with its 3 Vpp headroom will in no way be overdriven , which is a good thing since this stage is shared with the clean channel !

Now the complete channel is controlled with only one pot and it will be interesting to make it vary on the full range to see where we go : the lower values will be usefull to see what happens in the decreasing tails of the sustained overdriven notes.

Now the "log" pot . Of course with 11 graduations it is more precise than with 10 but for the computations it will be more simple to take 10 . Of course real logarithmic pots only exist in laboratories and not in our amps. What we call log pots are simply 2 linear pots in series , the second one having a far steeper slope than the first one . For each graduation the percentage of the input signal is this one :

So now I suggest we take a breath ( ) before we :
1. measure the various outputs of the stages with various inputs
2. identify for the various outputs the harmonics generated looking into the spectrum of each signal
3. measure the rate of ODD and EVEN harmonics in the signals by dividing the quadratic sum of the first 10 harmonics by the amplitude of the fundamental frequency.

Feel free to comment ....

 

[Ten Over]

Remember the low cut-off frequency due to the cathode RkCk network is given by :

View attachment 1112101

That equation will give you the frequency at which the gain is 3dB less than the maximum gain. However, the cut-off frequency is not necessarily at -3dB. What if there is only 4dB between maximum and minimum gain like the example on pg 484 in RDH4? -3dB would be at an even lower frequency than 1/(2*pi*Rk*Ck). This is going to happen anytime the square root quantity is less than one.

 

[ltournell]

That equation will give you the frequency at which the gain is 3dB less than the maximum gain. However, the cut-off frequency is not necessarily at -3dB. What if there is only 4dB between maximum and minimum gain like the example on pg 484 in RDH4? -3dB would be at an even lower frequency than 1/(2*pi*Rk*Ck). This is going to happen anytime the square root quantity is less than one.

View attachment 1112263

View attachment 1112264

What we call the cut-off frequency at -3 dB is a convention : it has been selected following a Gaussian curve of the population showing as a crest value -3 dBSPL as the threshold from which the majority hears the lowest variation of sound volume. Below -3 dB a minority hear a variation of level and a majority only a variation of color to the sound but not of volume. Of course if the gap between maximum and minimum volume does not reach 3 dB there is no cut-off frequency ! But in your example you give a 4 dB gap, so THERE IS a cut-off frequency that is close to the lower level ( 1dB) . The formula I give here is consistent with the RDH4 though the -3dB convention had not yet been settled in those days.

The condition for this cut-off frequency to exist is simply (µ+1) x Rk > (RA + rak) x 0.414 ( with RA = total plate load)

Besides we are speaking here of the SOLDANO preamp and the formula applies to the 3 bypass stages since the gaps are respectively 4.7 dB , 6.6 dB and 4.6 dB so I do not really understand your point. And the documents you show are not consistent with themselves : fig 1 does not show curves of a 12AX7 but for a tube which µ =20 , Ri = 10 k , RA = 100k , RL = 500k and top of all Rk = 2700 ohms, so it has absolutely no relationship with the following computations around a 12AX7. Without mentionning that the author makes the classical mistake of taking the resonant frequency of Rk//Ck for the cut-off frequency of the stage itself !

Here below the derivation of the formula

 

[Ten Over]

What we call the cut-off frequency at -3 dB is a convention :

It is a convention for simple filters like an RC filter. It is not necessarily the convention in other areas.

RDH4 definition:

 

[Ten Over]

The formula I give here is consistent with the RDH4 though the -3dB convention had not yet been settled in those days.

Nonsense. The -3dB convention had been used for the "theoretical cut-off frequency" long before RDH4 was published.

 

[Ten Over]

And the documents you show are not consistent with themselves : fig 1 does not show curves of a 12AX7 but for a tube which µ =20 , Ri = 10 k , RA = 100k , RL = 500k and top of all Rk = 2700 ohms, so it has absolutely no relationship with the following computations around a 12AX7.

So your formula only applies to a 12AX7?

 

[Ten Over]

Without mentionning that the author makes the classical mistake of taking the resonant frequency of Rk//Ck for the cut-off frequency of the stage itself !

Who? Me? I did nothing of the sort.

 

[ltournell]

It is a convention for simple filters like an RC filter. It is not necessarily the convention in other areas.

RDH4 definition:
View attachment 1112631

Now it is a universal convention : the level beyond which we hear a difference of volume , that is the same thing that the frequency beyond which the attenuation increases rapidly.

 

[ltournell]

Nonsense. The -3dB convention had been used for the "theoretical cut-off frequency" long before RDH4 was published.

View attachment 1112634

OK . You are right . I admit I haven't read the RDH4 in its entirety

 

[ltournell]

So your formula only applies to a 12AX7?

Of course the formula applies to all tubes. Read again the two documents you gave : the first network of curves is not about a 12AX7 ( p.484 of the RDH4 it is the particular network of a tube with a µ=20, Ri= 10k and Rk = 2700 ohms the gap is 4.1 dB) and the second IS about a 12AX7 with a whole different configuration (RK = 1500 ohms and the gap is 7.5 dB) .

Who? Me? I did nothing of the sort.

No, not you, the author of the document. He gives nowhere the true -3 dB cut-off frequency of the stage he shows. That demo is useless , and still more in my topic for which it is a digression . Please come back to the right topic of the SOLDANO.

 

[Ten Over]

I think that the crux of the biscuit here is that you think that the -3dB convention is applicable to a partially bypassed common cathode gain stage and I think that it is not. I have Merlin and Dave on my side.

 

[ltournell]

I think that the crux of the biscuit here is that you think that the -3dB convention is applicable to a partially bypassed common cathode gain stage and I think that it is not. I have Merlin and Dave on my side.

View attachment 1112671

Ah , at last I understand your point!

I had the opportunity to exchange a lot with Merlin some fifteen years ago (at the time he had begun writing books with huge mistakes in them ) and I know he has since written excellent things and he's a reliable author to whom I frequently resort to , with a few points left on which I disagree.

Now I do not know if the extract you show is by him or by Dave Hunter , but it does not matter , I strongly disagree with them.

-3 dB as I said up here is the convention to define , as far as SOUND is concerned, not PHYSICS, the threshold from which we hear a difference in volume .
To be more precise , suppose you compare the sensibilities of loudspeakers . You send 1 W into a 98 dBSPL LP and into a 100 dB LP : most of the listeners will not hear a difference in volume ( but a few will) .
Now you take a 98 dB and a 101 dB LPs : most of the listeners will hear the difference ( but not all)
Now you take a 98 dB and a 102 dB LPs : almost all of them will hear the difference.

So on a common partially bypassed cathode triode stage , for me, there is a cut-off frequency if the gap reaches 3 dB and there is none if it does not.

This so-called "half boost point" has absolutely no other practical usefulness than drawing pretty curves for the satisfaction of physicists , but not really for musicians. Musically speaking it is meaningless.

Of course everybody has a right to follow Merlin and Dave's point of view : we are not in a sect here
but I have the right to contradict them.

So I started with a table of the SOLDANO preamp stages , showing a -3 dB cut-off frequency with an explanation of what it meant to me . Now if anyone disagrees with this value, then what does he suggest as a MEANINGFULL value ?

 

[Pete Farrington]

showing a -3 dB cut-off frequency

To me, the ‘cut-off’ term above implies that the frequency response continues on a downward trajectory, as per eg a regular RC low or high pass filter.

And that would tend to reinforce what seems to be a widely held misunderstanding, ie that partial cathode bypass can be used to roll off low end in the same way as a high pass filter.

Hence it seems better to use terminology that better describes the action of the different filter types.
And so the ‘half boost’ frequency of a partial cathode bypass type filter seems a good term, as per this handy calculator, which uses the RDH4 methodology http://bmamps.com/CapCal.html

 

[Ten Over]

Hence it seems better to use terminology that better describes the action of the different filter types.
And so the ‘half boost’ frequency of a partial cathode bypass type filter seems a good term, as per this handy calculator, which uses the RDH4 methodology http://bmamps.com/CapCal.html

I can't get that calculator to spit out a correct half boost frequency in either dB's or voltage gain. Any idea where they got the formula in RDH4?

 

[Ten Over]

Now I do not know if the extract you show is by him or by Dave Hunter , but it does not matter , I strongly disagree with them.

Merlin Blencowe and David James

 

[ltournell]

To me, the ‘cut-off’ term above implies that the frequency response continues on a downward trajectory, as per eg a regular RC low or high pass filter.

And that would tend to reinforce what seems to be a widely held misunderstanding, ie that partial cathode bypass can be used to roll off low end in the same way as a high pass filter.

Hence it seems better to use terminology that better describes the action of the different filter types.
And so the ‘half boost’ frequency of a partial cathode bypass type filter seems a good term, as per this handy calculator, which uses the RDH4 methodology http://bmamps.com/CapCal.html

Yes, Pete , I agree. You are right to point out that this kind of filter is different from the pure RC filter but below the cut-off frequency only . Be sure there is no misunderstanding about that for me. If the gap is superior to 3 dB there is for me a change in volume at the -3 dB frequency . And depending on this frequency , we remember that at the input of a guitar amp we are not interested in what happens below 82 Hz , the 6 sixth open string being supposed to output the highest level . It mitigates the difference between a true low pass filter and the response of the stage.

Let us take the example of the 2204 Marshall preamp ( which I built a few years ago) :

It is obvious that the choice of the 680 nF capacitor at the cathode of the high input aims at limiting low frequencies which are undesirable on a supposed high gain preamp : which ones ? I admit the problem is delicate since the stage is loaded by the other in parallel with the 1 Meg pot . To simplify let us take the 330 pf as an open circuit (> 2Meg at 200Hz) the total plate load will be around
101 * 820 + 63000 (Ri) // 100 K + 200 K = 260 k (let us drop the pot)
then 260 k // 100k (RA) = 72200 ohms.

With my tools ( Excel ! ) I find a gap of 9.6 dB and a cut-off frequency of 231 Hz at -3 dB . The half boost frequency is around 150 Hz . I persist in thinking that the -3 dB cut-off freq means more than this half boost frequency as far as sound volume is concerned.

In fact I tried to use your tool but there is a problem . I moved the slider for the half-boost freq to adjust the capacitor value ( 0.695 µF here ) I find the same gap but the real half-boost freq I find at - 4.8 dB then is 146.8 Hz (?) instead of the 109.6 Hz necessary to have the nearest capacitor value of 0.68 µF. Are you familiar with this tool ?

 

[ping-ping-clicka]

I admit this is an impressive view
of Mr. S's work , and
I as so lost as to appear to be drugged .
Thank you for posting this

 

[Ten Over]

In fact I tried to use your tool but there is a problem . I moved the slider for the half-boost freq to adjust the capacitor value ( 0.695 µF here ) I find the same gap but the real half-boost freq I find at - 4.8 dB then is 146.8 Hz (?) instead of the 109.6 Hz necessary to have the nearest capacitor value of 0.68 µF. Are you familiar with this tool ?

It seems that BMAMPS is using Merlin and Dave's formula instead of something from RDH4. I have my own formula for the half-boost frequency, but I'm sure that you are not interested.

 

[ltournell]

It seems that BMAMPS is using Merlin and Dave's formula instead of something from RDH4. I have my own formula for the half-boost frequency, but I'm sure that you are not interested.

Yes, I am interested in the discussion of that half-boost frequency but I suggest I start another thread about that topic , not to digress too much in this thread which is about the SOLDANO preamp. This thread should be called "DESIGNING A PARTIALLY BYPASSED COMMON CATHODE STAGE : the problem of the "cut-off" , "half boost" or "roll-off" frequency ? "
And Ten Over , you will be welcome to expose your point of view with your own formula and everyone can discuss ...

 

[Ten Over]

Yes, I am interested in the discussion of that half-boost frequency but I suggest I start another thread about that topic , not to digress too much in this thread which is about the SOLDANO preamp. This thread should be called "DESIGNING A PARTIALLY BYPASSED COMMON CATHODE STAGE : the problem of the "cut-off" , "half boost" or "roll-off" frequency ? "
And Ten Over , you will be welcome to expose your point of view with your own formula and everyone can discuss ...

Here is my chicken scratch. I haven't typed it up yet.