Building a "Bottlehead" Cavalli-Jones Headphone Amplifier

My friend Steve is a regular at the AA and Bottlehead sites. He's built a number of Bottlehead kits even though he is not trained in electronics. Steve noticed Alex Cavalli's Morgan Jones Headphone Amp at the Headwize site and asked if it was possible for someone like him, who can't completely understand the schematic diagram, to build one. Furthermore, Steve wanted to build one that cosmetically matched his Bottlehead gear. A couple of emails later, Steve and I decided that we would work on the project together. We present the project both from my perspective and Steve's. Steve and I would also like to thank Alex, who made major contributions to this project.

Steve Says:"I think it’s important to make an effort to learn schematics — or at least the one you’re trying to build. I can now read a simple schematic, and while I still might struggle with putting some of it into a physical layout, it made this process easier and more enjoyable for me."

Step one: The Layout When I built my own Morgan Jones, I used my CAD program to draw a full-sized pictorial drawing of the MJ amp. I did this to be sure how the parts would fit into the rather small chassis that I had chosen. A little more work on that original drawing and Steve and I had come up with an idea of how to lay out the amp. Central to the construction of this amp was the use of a star grounding pattern that had originally been worked out by Headwize member Prune. Also, the Bottlehead-style layout on a flat chassis plate puts all the parts out there to see with lots of space to work safely.

Steve Says:"Without this CAD drawing, I would never have been able to build this amp. My actual chassis plate is 9”x11”. I drew a rectangle to those dimensions on a piece of paper, then physically placed the components, one building block at a time, and marked their location. I then taped the paper to the back of the plate (remember, this view is from below), and used a steel awl and hammer to punch-mark the hole locations."

The circuit depicted here is for the "optimized version with feedback" whose schematic is shown in Figure 10 of the original Headwize article. It has been slightly modified to make the feedback variable as described by Alex in subsequent disussions at the Headwize DIY forum

Steve's power supply initially was the same one I used in my MJ. But, pictured to the left is a safer power supply, especially for beginners. It has a voltage divider at its end to bias the filament supply in order to prevent the Heater to Cathode voltage of the tubes from being too high. The other andvantage of this voltage divider is that it serves as a "bleeder" that will bleed off the voltage from the power supply capacitors within 5 to 10 minutes of shutting off the amp. This will make servicing and tweaking somewhat safer.

The idea for the power supply came from Alex after some discussion at the Headwize DIY forum. I worked out the details using the Duncan Amps PSU designer.Click Here to see the Duncan Amps PSU designer output for this PSU.
The pictorial drawing at the top of the page now depicts wiring the amp with this power supply although the pictures of Steve's unit do not.

Building Blocks We broke the amp down into "building blocks" that represented the various parts of the amp. Shown on the left below are the power supply building blocks and on the immediate left are the tube circuits.

In my version and in Steve's we wired V2 and V3 as mirror images of each other. We did this to keep the passive components of each tube located on the outside edge of the chassis. This made more efficient use of the space on the chassis.

Several months later, we discovered that the two sections of a 6DJ8 (but NOT a 6922) are not exactly identical with respect to the heater/cathode potential that they can withstand. In building this amp, the B+ should always connected to the plate that is connected to pin 1 of V2 and V3 as shown in the schematic drawing above. The pictoral drawings have been changed to reflect this modification.

Steve then assembled each of the building blocks on the chassis to be sure of how they would fit. Final construction involved simply interwiring the building blocks to each other.

Steve Says: Remember I talked about learning to read a schematic? I actually caught a wiring error Bill had made in an earlier layout. Don’t think I’m going to let him forget it, either…

Bill's Comment It's tough being a guru when your student insists on catching your errors.....How embarassing! To make matters worse, I insisted through several Emails that he was reading the schematic wrong :-)

Step 2: Developing a Parts List We used good parts throughout; Vishay/Dale resistors, Panasonic electrolytic and polypropylene capacitors and ceramic tube sockets. Like me, Steve had also decided to wire his unit with long-grain magnet wire from Doc Bottlehead. If you're a believer in the goodness of wire, consider that there's 15 feet of wire crammed onto this little board. It should be good wire!
Not shown in any of the drawings is the 4.7uF film capacitor listed in the parts list. It is used as a bypass capacitor across the 470uF output capacitors. New builders may wish to try the amp without the bypass capacitor first since not everyone feels that bypass capacitors belong in the signal path. You can see the bypass capacitor installed in the picture of Steve's completed amp below.

Note also that there are two 5W resistors (0.5R and 0.25R) in the parts list. These are the optional resistors shown in the PSU schematic for adjusting the filament voltage. It is likely that you will neeed both if you use the Shottky diodes specified. My unit has 0.6R in the filament supply and Steve ended up with 0.75R in his filament supply. Make sure you leave room during construction to accomodate two resistors in the filament power supply. The second resistor is not shown in the pictures of Steve's unit becasue it was added after the pictures were taken

The parts list also has two 5W resitors for the power supply, 1.8K and 2.0K. Depending on local conditions and parts tolerances you will need one or the other to produce the correct voltage from the power supply. We recommend that you start by installing the 2.0K/5W resistor and then replace it with the 1.8K unit if the PSU voltage is too low.

Step 3: Contruction Begins Steve laid out the amp on an 9 X 11 aluminum plate after slightly altering the location of the 3 tubes. After making all the holes, Steve began by builidng the power supply and filament supply. Note Steve's careful construction techniques and clean wiring. Steve's attention to detail is what made this collaboration possible!

Steve Says:"Ah, the black art of grounding. Run all ground, heater, and B+ wires close to the plane of the chassis. My signal wires are all elevated. Wires should always cross at 90 angles. I used 4-nines long crystal copper magnet wire, and placed it in Teflon tubing. I started with the top half of the layout, the power supply. Run the wiring first, then on go the caps and resistors and other implements of destruction. Next, I did the B+, filaments, and most of the ground wiring for the bottom half."

Step 4: Wiring the Tubes Steve next wired up each of the tube sockets using terminal strips to hold the components. Notice the high quality of Steve's solder joints; This baby's gonna fire up without a problem!

Steve Says:"Do them one at a time, and take your time. Like grammar school penmanship, neatness counts!"

Step 5: Miscellaneous Wiring Having finished the tubes, Steve turned his attention to wiring the input and output circuitry.

Steve Says:"Don’t rush through the home stretch. I built this over the course of weeks, not hours. I doubled and triple-checked my work to make sure everything was where it was supposed to be. Oh — I use heat sinks (alligator clips) when soldering caps and resistors and diodes."

Step 6: Ignition and Liftoff!!! Because of Steve's careful work, the amp came to life immediately. The only problem he encountered was a blown power fuse as the PSU capacitors charged up for the first time (not unusual). When the PSU voltage stabilized it was a respectable 221VDC! Way to go, Steve! And we had only one teensy-weensy problem, we wired the pots BACKWARDS; the old drawing mirror images in the CAD problem :-) The pictiorial drawing above now shows the correct wiring for the pots!

The only other "tweak" that was required was to add a second resistor to the filament supply to adjust the voltage to 6.3V. With a single 0.5R resistor, Steve was getting about 6.9V on the filaments. Adding another 0.25R brought the filement voltage down to 6.35V.That secoond resistor in the filament supply is not shown in these pictures since it was added later

Steve selected maple for his base. Not only is he a pretty good electronic technician, he's a darn good woodworker to boot!

Steve Says:"The chassis is stock aluminum plate brushed with a circular metal brush that attaches to a drill. Finished with low-lustre spray-on lacquer. The base was sanded from 150 through 380 grits, then wiped with a tack cloth. That’s a six-layer hand-rubbed finish: 1 coat dye stain, 2 coats gel stain, and 3 coats of polymerized tung oil."

Safety First and Always! Steve Demonstrates using a 10K-50K 3-Watt resistor to discharge the PSU capacitors before working inside the amp. Without a bleeder resistor, These capacitors will hold a lethal charge for several days after the amp is turned off. The two large output capacitors must also be similarly discharged. Don't ignore this warning!

Note Even though the new power supply shown here now has a bleeder resistor in it, it will take over 5 minutes after turning the amp off for the power supply voltage to decay to safe levels. ALWAYS measure the voltage on the PSU capacitors before working inside the amplifier and discharge them with the resistor as shown if necessary. Also, the bleeder resistors do not discharge the large output capacitors, which still must be discharged manaally before working inside the amp.