Hoffman 6G15 - does it need drive filter?

[boogieWoogie]

Hi All,
I'm about to build both the Hoffman 6G15 standalone reverb & Hamptone's solid state reverb driver.  The Hamptone design notes from Tapeop point out that Accubell (tank manufacturer) says you should drive the tanks w/ a signal which rises 6dB/octave (i.e. 20dB/decade) w/ increasing frequency (i.e. a simple low-pass filter w/ a very high -3dB point would accomplish this).

The Hamptone design uses some filters so that the voltage source which drives the tank has a freq-response like so:



Accutronic's words are at http://www.accutronicsreverb.com/ > Application > DRIVE CIRCUIT / OVERLOAD MARGIN CHART

I don't see similar filtering in the Hoffman design.  I wondered if I'd missed it, or perhaps the Hoffman design is a constant-current design, which IIUC does not need such freq-response treatment.

If the Hoffman design is indeed a voltage-driver, should I consider adding a filter right after the driver triode to create this frequency-response?

Thanks!
BW

[sluckey]

Hoffman's design is  a copy of the original Fender 6G15 with a couple power supply improvements that don't affect the operation. It works just fine as designed.

[2deaf]

I don't see similar filtering in the Hoffman design.  I wondered if I'd missed it, or perhaps the Hoffman design is a constant-current design, which IIUC does not need such freq-response treatment.

Constant current is what you always strive for (at least for the frequencies of interest) because the current is what makes the magnetic flux that twists the springs and you want all of the frequencies to activate the springs the same.

The input coil on the tank doubles its impedance with each increasing octave.  In order to have the same current when the impedance doubles, you would have to double the voltage which is +6dB. 

There is more than one way to achieve a voltage that increases at the rate of 6dB/octave.  Op Amp drivers typically place the input coil in the NFB loop so that when the coil impedance doubles, the NFB drops in half and the output voltage of the Op Amp doubles.

The 6G15 utilizes a driver transformer that reflects the tank input coil impedance to the primary.  The 6V6 puts out about the same current regardless of the plate voltage, so when the impedance of the primary of the driver transformer doubles, the voltage across the primary also doubles.  When the primary voltage doubles, the secondary voltage also doubles so that the input coil gets a voltage that increases at the rate of 6dB/octave.

[PRR]

If you drive a tank with a pentode it naturally has that response.

(A small triode is also do-able, as in the 12AT7 driver.)

[boogieWoogie]

Hi All,
Thanks for the replies.  2deaf, thanks for going into detail!  I followed all of it except that

The 6V6 puts out about the same current regardless of the plate voltage

.  My little bit of tube theory knowledge can't speak to that, and a google search of "power tube output current" didn't turn up much to clarify.

2deaf, if you can, can you describe how a newbie could convince themselves that the current output of a 6V6 will be constant despite load impedence?  My guess is that the small-signal model of an output 6V6 is a constant-current source, but if so, why isn't that also true for a solid-state driver?  (Or to ask it another way, why does a solid-state driver need this filtering to generate the rising +6dB/octave response, and the tube does not)?

Thanks!
Mike

[shooter]

and the tube does not

Look at it another way;
Fenders G15 design has worked well and well received since the '60s.  If you are totally bored, build it, get it working, then put a mic in a room while playing and record, go back with analysis software and analyze the waveforms you recorded

[PRR]

The Plate Resistance of 6V6 at normal operating point is about 50K. The transformer in 6G15 is 5K or 7K. The plate impedance is much higher than the load impedance. Nearly constant current drive.

(Tank impedance is rated at 1KHz and continues to rise. The mis-match runs out when reflected impedance rise to plate impedance. This is about 7X or 10X of 1KHz, so well beyond what we need or want.)

Naked transistors tend to be infinite impedance, yes. Also so much gain we can't keep them steady. The stock approach is heavy NFB. This can either increase or decrease output impedance; the usual forms make output impedance near zero, but there are other ways.

If you have 6G15 parts on bench, build it! Then decide if Fender was a fool, oblivious to the very dull sound of an uncompensated tank. History says he did OK. And decades of experience with very fancy EQ-ed tanks tells me that while it can be "different", there's not a lot of "better" possible.

[Willabe]

If you have 6G15 parts on bench, build it! Then decide if Fender was a fool, oblivious to the very dull sound of an uncompensated tank. History says he did OK. And decades of experience with very fancy EQ-ed tanks tells me that while it can be "different", there's not a lot of "better" possible.

 

I built the 6G15 and it's got WAY more reverb than anyone could ever need or want.

PLENTY of high end sparkle and chime. In fact I roll off the tone on it to about 6 or 7.

And I play a Les Paul with HumBucks for slide.   

I've never heard anyone complain about it being dull or lacking tone, ever. And we've had a lot of guys build them here and post about their build. 

Even the surf guitar guys only set the 3 controls ~6 each. 

Have you ever played through one?

[2deaf]

2deaf, if you can, can you describe how a newbie could convince themselves that the current output of a 6V6 will be constant despite load impedence?

Imagine an Average Plate Characteristics chart with grid voltage lines that are dead horizontal.  Now if you draw some load lines for various impedances, each one that intersects the same grid voltage line will be at the same plate current.  The 6V6 Average Plate Characteristics are not dead horizontal, but they are somewhat horizontal.  Just looking at the chart with no load lines and seeing that plate current doesn't change very much with different plate voltages and that changing the load impedance is one way to change the plate voltage might convince you.

why does a solid-state driver need this filtering to generate the rising +6dB/octave response, and the tube does not

I can configure SS drivers that don't need filters and I can configure tube drivers that do.  If you take the same Op Amp that I mentioned earlier (that didn't require filters) and put a fixed NFB on it, it will put out the same voltage over a staggering range of load impedances and it will require pre-filtering to achieve a +6dB/octave response. 

   

[boogieWoogie]

Thanks again for the details.  The flat slope of power tubes' characteristic plate curves was illustrative.  I did some more reading & in Bruce Rozenblit's Tube Audio Design he states the same thing that you've said:

the [plate] voltage can increase several hundred volts w/ scarcely any rise in plate current. This simply means that plate voltage has very little effect on plate current.