The allure of the EF86 grabs most amp-heads at one point or another. Vox used 'em, and so did Matchless and Dr.Z. They are cool tubes, and they do have a different personality than the more common pair of triodes pre-amp arrangements. The problem with having an itch for a EF86 is that, in most cases, a EF86 based amplifier is well over $1000. That is a lot of scratch just to see if the EF86 matches your style.
With a desire to see how an EF86 would sound in a particular amp, this small add on EF86 box pictured above in the recess of a combo amplifier let me try out an EF86 and "A/B" test it against the twin triode 12AX7 pre-amp circuit already in the amplifier. This addition can be done without having to gut the circuit and punch holes in the chassis. If I like it, I can leave it there, If I don't, I take it out in five minutes (and try it in another amp).
The amp happened to be a nice example of a rare (but not particularly valuable) Gregory Apollo 800. From a desire to avoid massively rewiring a 12ax7 preamp circuit, this idea was hatched: a small box that could fit inside an amp, tap into B+ and ground sources and provide an input into the circuit for the EF86. It wouldn't require removing components from the circuit, nor would it be irreversible.
Anyone that puts an EF86 in a guitar amp for an initial gain stage most likely does so with a Vox sound in mind. The following is a EF86 schematic that Vox has used in various amplifiers. The components listed are from a 2000 AC30/4. Not only are the component values the same for the newer 2008-era AC15 Handwired (HW), the component values are also the same values used in the 1960 Vox AC4!! (where the component values differ, they are listed in parenthesis).
The pentode/triode switch is a Handwired option. If the switch is omitted and the EF86 is to be used in triode mode, R5 and C3 can be omitted as well.
C4 and R3 is really part of the power supply circuit, the B+ on R4 should free of signal AC from other stages.
R5 1M (4.7M AC15HW,
5.5M 1960 AC4)
R6 2.2k ( 1.5k 1960 AC4)
This is a potential point-to-point layout using a tube socket and 2 mini-tag strips. Tube socket pins 4&5 are for filament voltage, and pins 2&7 are the internal shield and they should be grounded. R3 and C4 can be omitted if the power supply section has already done this, however there is little harm in doing it this way. This layout uses the center pin on the tube socket which does nothing on a EF86. Some sockets don't have this, but its an extra lug that is handy in point-to-point wiring.
This can be constructed on a small board, or in an enclosure, and then added to an amp as a 2nd input. Splicing it into an existing circuit should be harmless. There are a few to be aware of, however:
Below is an example of how it might be added to a bargain bin Gregory 800 amp. The amp has a nice output section. The cathodyne phase inverter doesn't really provide any gain (an AC15 or AC10 would have used a long tail phase inverter which does provide gain to both signals). The signal either needs to be applied at the input to the phase inverter or at the grid of the the second gain stage.
Important note about schematic: the light green highlighted resistors on the cathode of the second stage are differences between this schematic and actual version of this amp as Gregory built it. It is not a change made in conjunction with the addition of the EF86 box.
At the "X" spot in the above melded schematic is the 'tie-in' point. This is where the EF86 circuit ties in to the old Gregory circuit. This is the one modification to the original circuit: a mix resistor must be added between the volume control center lug and the grid.
The mix resistors are R5 and R6 in the above schematic. They need to be high enough so that the volume control on one circuit doesn't effect the volume control on the other circuit (don't make R5 and R6 too low). Another important aspect of the mix resistors is how it changes the effective grid leak resistor of the triode (don't make R5 and R6 too big).
A good place to start would be 100K, 220K, or even 470K.
Below is a gut shot of the completed box. A 1M pot was added to the output signal to attenuate the signal as needed. The toggle switch turns on the B+. A DIN-5 connector carries (1) ground (2) signal out (3) B+ supply and (4) and (5) are the 6.3v filament voltage.
A 1.5M ohm resistor was selected for R5, and C1 was changed to a .001uf capacitor.
Above: the box is mounted to a plate to allow the the box to installed in the bottom of the cabinet.
If you don't want to make screw holes in your cabinet, some pedal velcro would do the job (or duct tape!).
The back of the box has three holes drilled to mount a 6.3v transformer if the amp's transformer doesn't have enough extra amperage to run the EF86's filaments.
Once the unit was completed, splicing it into the existing circuit took just a few minutes (and I was not very familiar with the layout, and it was still a quick job).
The result was exactly what I had hoped for. More gain and more high end clarity, EF86 style.. Using the EF86 stage instead of the original Gregory 12AX7 made for a tremendous difference and added nice voice option for already unique amp. And with the modular box and slice-in wiring harness, it was not a 'gut job' and I didn't drill any tube socket holes either.
Being able to A/B the two in the same amp with the same analogs from the phase inverter to the speaker makes this box very useful. In the above schematic, I designated two potential tin-in locations, and I tried them both. Inserting the signal at the phase inverter worked fine, but the volume of the amp was too low (cathodyne inverter doesn't provide gain like a long tail inverter). Inserting the signal at the grid of the second stage was the magic. This let the volume knob on the front of the amp remain functional and the tone knob as well (with the 300k volume knob at high volume levels, the tone knob functions normally, even though the signal comes in at the back end of the tone circuit).
The attenuator pot is very much needed in this arrangement. provided as much gain as the amplifier could handle. I used the pot on the EF86 box to dial in the gain (50% position on a 1M pot, so its 500k/500k voltage divider).
The wiring harness is a DIN-5 connector with a twisted pair for the filaments, a ground wire, and a B+ wire. The audio signal is carried in a coaxial sheilded cable. The amp is no more noisy with this arrangement than it was prior to adding the EF86 box.
Also seen in this amp: Above to the right is a pair of switched jacks that allows an external speaker to be used, or for another amp to plug into this speaker.
Other than the caveats about the power transformers discussed previously, there aren't really any tube amplifiers to which this could not be adapted. Some would be more difficult than others. As easily as this box is plumbed into a circuit, it can be removed. A pristine vintage amplifier is restored to originality in five minutes time (your wait longer for your soldering iron to heat up!). If the box is secured with velcro (or even duct tape), it'll leave no traces.
This can, of course, be done with a 12AX7 instead of an EF86, or a 6SJ7 for that matter. Any it doesn't have to be a Gregory Apollo 800, it could be a Vox AC15CC1, an Epiphone VJR, or an old Silvertone or an old Supro (you don't gut that, right?).