A Stronger One tube Reverb with a SF ?

[kagliostro]

Few days ago I readed a thered where was posted a piece of schematic (to me seemed drawn by Tubenit, may be it wasn't)

I don't remember the thread, what was exactly on the schematic and which was the argoument of the thread

however the faint image I've on my brain led me to think about a strongher one tube reverb

(someone say they like the one tube reverb but consider it a bit weak)

so i've think to add a Source Follower (may be the image I don't remember was exactly for this purpose)



What do you think about ? Is it feasible ?

Franco

{EDIT-- title typo --PRR}

[John]

The one-tube reverb still uses a reverb transformer, at least the one that I use does, copied from Tubenit. I can't imagine it not giving enough reverb for anything less than full on surf, but that's maybe just my taste?

[kagliostro]

About output transformer ..... I think that also this way it can be used, only must be find the required impedance

Hower there are also many reverb circuit that are capacitor coupled

Franco

[John]

Hower there are also many reverb circuit that are capacitor coupled

I know nothing of those.   

[kagliostro]

Here in Italy there was a brand (FBT) taht used the capacitor coupled reverb





but there are also other brand that did the same

Franco

[2deaf]

What do you think about ? Is it feasible ?

I tried it with a LND150 instead of a 12AX7 and it worked. 

[kagliostro]

Well, so seems it is feasible

A way to use only a tube and have a stronger reverb, also because if you use a pair of double triodes, remain one triode unused

Franco

[2deaf]

Well, so seems it is feasible

Yes, but not with those component values.  Try changing the 120K source resistor to 20K and insert a 10K resistor between the 1uf capacitor and the tank (presumably an 8F or 4F type).  The 120K resistor is way too large when driving a low impedance load such as a reverb tank and you will only get a very small signal at the point of clip.  If you go lower than 20K, the current consumption will increase above my 7mA limit.

The inserted 10K makes a load that can be driven when the source resistor is 20K.  The 10K also makes this a constant current driver because the series LRC impedance is about the same across the frequencies of concern.  The 10K also makes the resonant frequency irrelevant.  With only a 1uf cap. and a 235mH coil, the resonant frequency is 328Hz which is a frequency of concern.  The total impedance at and around 328Hz becomes very low and the driver signal will clip at an extremely low level.

A way to use only a tube and have a stronger reverb . . .

This driver will only do about 1.6 times the published nominal input current.  Plenty to drive the tank, but no where near what a 12AX7 can do with a transformer.     

[2deaf]

The 120K resistor and the 500pf cap. on the output coil act sort of like a parametric equalizer with a high value of Q.  The output coil puts out its maximum signal for a circuit at the resonant frequency and then the signal drops off on both sides of the peak just like a parametric equalizer. 

The value of the resistor determines the height of the maximum signal peak like the gain control on the equalizer.  The value of the cap. determines the frequency of the peak like the frequency control on the equalizer. 

The output coil has an inductance of 358mH.  The resonant frequency with a .0022uf cap. is 5.7KHz  A 220K resistor allows the maximum signal that the coil can produce and going larger than 220K will not increase the peak.  However, going lower than 220K can reduce the height of the peak.  The 220K/.0022uf combination is like having a parametric equalizer set to 5.7KHz with the gain set to maximum.  The end result is that the reverb sounds brighter.

The resonant frequency with a 500pf cap. is 11.9KHz which is much too high to be of any use to us.  You will be hard pressed to tell the difference with the cap. installed or not.

With no cap. at all, the value of the resistor determines the high frequency roll-off.  A 27K resistor will cause a cut-off frequency of 12KHz which will diminish the magnitude of our frequencies of concern a little.  Lower values of resistor will cause an undesirable cut to the higher frequencies of the reverb.       

[tubenit]

As an FYI,  anytime I find anything about a "one tube reverb" ............... I will typically place it in ARCHIVES if I think it might be useful.

http://el34world.com/Forum/index.php?topic=7957.50

With respect, Tubenit

[2deaf]

I drew some graphs and did a little breadboarding over Thanksgiving to see what a hybrid source-follower reverb driver could do.

Tube/MOSFET Driver #1

The DC Plate voltage on the 12AX7 sets the DC Source voltage on the IRF820 and, therefore, the DC Source current at idle.

The total impedance of the series 1.0uf capacitor, 4.7K resistor, and 0.235H coil doesn't change significantly from 100Hz to 2KHz.  The AC Source voltage is set by the AC Plate voltage, so the AC current through the LRC series circuit is about the same from 100Hz to 2KHz for a flat signal.

The 500pf/.001uf bandpass filter in front of the Grid provides just about all of the low frequency roll-off that we want.  The series LRC impedance increases significantly above 2KHz because the coil begins to dominate the equation.  The bandpass filter combined with the increased LRC impedance results in a sharp roll-off at higher frequencies.

The IRF820 goes into cut-off when the negative AC signal from the Plate drops low enough.  When in cut-off, the Gate to Source voltage can exceed the maximum rating because the Source voltage no longer follows the Gate voltage.  The vertical zener diode prevents the Gate to Source voltage from exceeding the maximum, but when it breaks over, it causes the negative side of the plate signal to clip.

The IRF820 draws about 10mA DC and the 12AX7 draws about 1mA  DC for a total of 11mA for the complete driver at idle.  It will deliver 5.5mA_rms to the input coil at the point of negative clip.

The IRF820 requires a small heatsink.

 

[2deaf]

Tube/MOSFET Driver #2

The series LRC circuit changed from 1.0uf to 0.47uf and from 4.7K to 2.7K as compared to Driver #1.  The 2.7K resistor doesn't even out the impedance across the frequencies as much as the 4.7K resistor.  The total LRC series impedance is also smaller with a 2.7K resistor, so more AC current flows through the LRC circuit.

The 0.47uf cap. increases the impedance of the LRC series circuit at lower frequencies as compared to a 1.0uf cap. resulting in some low frequency roll-off.  The 2.2uf Cathode cap. also rolls off some low frequencies and the two effects combined causes a sharp low frequency roll-off at the tank.

The 500pf cap. at the Plate rolls off high frequencies and the increased LRC series impedance at higher frequencies also rolls off high frequencies.  The two effects combined results in a sharp high-frequency roll-off.

This driver gets it band-pass filtering from the combination of effects from the circuit components instead of a bandpass filter in front of the Grid.

This driver also has a total DC consumption of 11mA like driver #1, but it can deliver 6.3mA_rms to the input coil.

[2deaf]

Tube/MOSFET Driver #4

The 12AX7 is running way over at the cold side in order to get the Drain to Source voltage at 52V DC.  If I run the 12AX7 any colder, the signal becomes too rounded and asymmetric when giving the IRF820 the necessary swing.

The IRF820 can swing more current and voltage at 52V Drain to Source than at 100V DC.

This driver consumes 13.4mA DC and can deliver 7.4mA_rms to the input coil.  This is 3.7 times the published nominal current necessary.

[kagliostro]

Many Thanks 2deaf for sharing your emulations

Franco