Sortaflex: D.I.Y. Tube Bass Preamp

Way back in 2013, I built a clone of an Ampeg B15n Portaflex “fliptop” head unit. What I came up with was a sort of average based on several schematics I’d found, all with slight differences. That project was fiddly and expensive, but I’m glad I saw it through because it was my first D.I.Y. tube bass amp. I tried to be as “canonical” as possible, given cost concerns and available resources.

So after emerging from under my rock a few months ago, I decided that I wanted to build a tube bass preamp that would run on AC power. I chose the Portaflex preamp section as a model mostly because I’d had a good experience with it before, but I wasn’t super-strict about sticking to the original. What follows is a brief record that project.

For questions and comments, email th@thallenbeck.com

History of the Ampeg B-15 Portaflex

Note for D.I.Y.ers

If anyone wants to make one of these, I can put up PCB images, a bill of materials, and other resources. Both PCBs are single-sided. Although I chose to have a few manufactured, they can be made with transfer / ferric chloride, or similar. See the end of this post for more information about components and such.

And now, two action-packed videos just for you!

Schematics

Schematic for the main board
Schematic for the power transformer and RC filtering

The main differences between this gizmo and the venerable Ampeg Portaflex are:

  1. 12AU7/12AX7 nine-pin tube instead of a 6SN7/6SL7 octal tube. Most of the fliptop schematics I’ve seen specify 6SL7 for the preamp tubes; one showed 6SL7s. I used a nine-pin tube here to make it easier to stand the PCB up and stuff everything into a Hammond enclosure.
  2. Different plate resistor values (R3 and R14 in the schematic) and a different biasing voltage. I used whatever DC voltage was present at the end of the RC-filter section, which I’ve measured as about 206VDC (loaded). R3 and R14 could be tweaked for different gain and different impedance, but since the output signal is pretty hot to begin with, I’ve tried to keep the overall gain down.
  3. Addition of capacitor on cathode of V1A. Can be omitted. See schematic.
  4. No phase inverter or power stage (obviously, since it’s a preamp). More about this below.
  5. 68k grid-stop resistor at the grid of V1B. I’m not sure why I thought that should be there, but it’s part of the 2013 project as well. Maybe it’s a holdover from a previous guitar-amp project.
  6. The volume control is after the Baxandall EQ and before the V1B input. In a typical fliptop amp, the volume control would be just before the phase inverter.
  7. Different values for the coupling capacitors. I used mostly big WIMAs and small WIMAs. Most of mine are 1uF, but “canonical” values (22nF) could be used instead.

So, does the Sortaflex sound like a Portaflex? Sort of. Maybe some… or not. Hence the name. Using a different tube probably makes it apples to oranges by default.

And it’s just the preamp section. Although I can’t prove it, I’ve always thought that the phase inverter stage of the Portaflex contributes to a quality in its sound that I can’t describe properly: strech-y? Rubber-band-y? Slightly compressed?

What I like most about the Sortaflex:

  • With a 12AU7, it’s got plenty of clean headroom.
  • The low-end response is huge and for that matter, the EQ is wide and responsive.
  • I’m not hearing any AC hum – it’s quiet as a church mouse.
  • The balanced (XLR) output sounds pretty good from what I can tell.
  • It’s giving me a chance to teach myself how to powdercoat small enclosures. (As you can see from the photos, I’m still learning.)
  • It looks cool in the dark.

What I like least about the Sortaflex:

  • It’s big and heavy, and seems like overkill for a preamp.
  • It’s expensive to make, even without the $80 Jensen transformer.

About the PCBs

They are 0.062″ with a 1-oz. copper thickness, not the thicker kind (0.094 or 0.125 with 2-oz. copper thickness) that one often sees in conjunction with tube amps. I made the traces nice and wide though – most of them are 0.04″, or 0.07″ for the ground traces on the RC-filter board. If power tubes had been involved, I probably wouldn’t have used PCBs at all. But so far I haven’t had anything go wrong with either of them.

The small holes in the boards are for the screws that hold the standoff posts in place. On the back sides, you’ll see clearance in the groundplane so that the standoffs are isolated from any traces. On the main board, the big hole toward the bottom allows the pair of filament wires from the power transformer (the thick green twisted ones in the photos above) to pass through and solder to the heater connections from the back.

About the Balanced Output

The balanced XLR output is achieved by a 12:1 step-down Jensen JT-DB-E transformer, the cost of which (about $80 US) is way the **** out of proportion with a typical D.I.Y. project’s budget. I sprung for a couple of them because I really wanted to see how this would turn out. Although I’m quite pleased with the result, I should probably have a psychological workup for spending that much money on something that’s going to sit in my basement until civilization collapses… which could be any day now I guess.

The transformer, having a 12:1 (primary to secondary) ratio, pretty much does it own isolation. The primary impedance of the JT-DB-E is something like 140kOhms. The first and second versions of this project (see photos below) used an IRF820 MOSFET to buffer the signal to a 1:1 transformer. They worked okay but were fiddly as heck even after I worked the buffer onto the PCB (that version is not pictured).

There might be cheaper audio transformers with high primary impedances, but I settled on the JT-DB-E after I opened up my oft-used Radial DI box and saw that one inside it. However, I left what I think is enough space in the the center-right section (looking at it from the bottom) for chassis-mounted transformer, in case I want to experiment in the future. For a version with no balanced output, the PCB could be populated without the transformer section, by omitting C11, R18, R19, the transformer, R20, C12, and the ground-lift switch. Or, for an offboard (chassis-mount) transformer, a wire can be run from where C11 meets R14 and V1B, or from R19 if R18 and R19 are needed.

About the Ground Wiring

Since this is a high-voltage project, I tried to use star grounding even though it’s just a glorified pedal. That’s why you see so many thick black wires connected to the RC-filter PCB (see photos above).

About star grounding: http://www.geofex.com/article_folders/stargnd/stargnd.htm

About amp grounding in general: https://www.aikenamps.com/index.php/grounding

For the RC filter board, the diode rectifier, the filtering capacitor that follows it (C1), the next RC filter stage (R1 and C2), and the voltage divider for the LED all have their own own ground wires (see schematic and photos above). They run to the terminal strip along with the ground wires for the power transformer’s filament supply and the tube circuit board. The two articles referenced above will explain why such a seemingly fussy approach can be desirable. It’s worked well for me when I build tube amps. Or what I should say is: I always use this approach for tube amp projects –> I usually get very little hum noise in said projects –> I’m superstitious.

By the way, referring to the two photos above, those two wires extending in perpendicular from the PCB are for the bright blue LED. A voltage divider prevents the LED from being fried (see schematic above).

About the Components

The tube in the videos is a JJ ECC82 / 12AU7. I’ve also tried an Electro-Harmonix 12AU7 and a Tung-Sol 12AU7, with similar results, and 12AX7’s from Electro-Harmonix and Tung-Sol, which yielded much higher gain. That could be nice if you want tube bass fuzz.

The power transformer SKU is TR-PW-13. I get them from TubeDepot. Their website has a data sheet for it. It can drive a single tube, with ~= 200VDC/5mA at the secondaries (not center-tapped) and 6.3V/300mA for the filaments (center-tapped).

The RC-filter board uses F&T electrolytic capacitors for smoothing the oscillating AC into a useable DC voltage. They’re pricey but I like them. I get them from TubeDepot or Amplified Parts.

The AC switch assembly came from someone who sells on Amazon. There are many similar listings but not all are appropriate for this project. The fuse is worked into the unit. Separate components for the plug, the switch, and the fuse would also suffice. This sort of mousetrap should always be fused. And there should always be an Earth ground, distinct from any other ground points (see schematic and photos).

Most resistors are Vishay 2-Watt metal-film. I like them better than 1/2-Watt. I get them from Mouser.

Most nonpolar capacitors are WIMA, also from Mouser. The larger ones are in positions where a higher DC tolerance (> 100V) is required.

The tube socket is from TubeDepot, SKU SK-9PINPC2.

The enclosure is a Hammond 1590D.

The PCB standoff posts usually come from Angela Instruments. The RC-filter PCB has 1/4-inch standoffs and the ones for the main board are 1-1/2 inches, to make room for the pots and the 1/4″ jacks.