Circuit Edits

So, while turning the amp on for the first time, I found it squealing and way noisier than could be used in any kind of setting. I ran through my notes and discovered I had wired the primary of the output transformer out of phase, that is the secondary was feeding the feedback loop out of phase, so that it wasn't taming the amp - it was making it feedback on itself. Hence the squeal and super loud noise. I remedied this by swapping the transformer primary leads on the tube sockets. Problem solved.

Another fix - the control grid of the 7591 tubes was much lower than I would have liked - sitting at around 305V, rather that 370V-375V that it should be. The extra PI filter that I had running of one of the HT leads was knocking down the voltage a bit too much. I stuck a pot across the existing resistors and varied it until I had approximately 365V-370V at that point. There was a sequence of adjustments, because as I changed that value, the plate voltage changed, and so did the bias. All 3 were reacting with each other. I eventually reached a happy medium. I connected a 22k-ohm and 33k-ohm resistor in parallel with each other.

Feedback - The Fender circuit has an 820-ohm resistor feeding back into the phase inverter stage (pre-7591), while the Fisher has a 2200-ohm in parallel with a 1000pF cap. Again, a happy medium. I have a decade resistor box that I ran across the feedback point and listened to the circuit with different feedback resistors. 1k-ohm sounded good. I replaced the jumper with a 1k-ohm resistor in parallel with an 1800pF cap.

Reverb make-up gain - tube replacement. While testing out the amp, I found that a 12AX7 (at least any of the ones I had) were far to noisy to be used in the make-up gain section for the reverb return. I threw in a used 12AU7 in there and that cured my problem, without making the amp sound noticeably different. I may consider a JFET return amp, instead of the 1/2 of a 12AX7, leaving the other half for the mixer circuit. I might be able to use the unused (when modified) section for gain for the solid state preamp I intend to use as a channel option to the tube preamp.

Project End

I finally put the amp completely together, and this project is at an end. Through many mistakes, mostly with wiring, I managed to build a working amplifier.

The bias board that I had originally built on perfboard was rebuilt. I was not happy with the quality of the work on this board. I decided to cut up some panelboard and mount some eyelets to the board, using a layout that I quickly drew up on graph paper, as a guide. The result is a more uniform look, as well as studier construction. I don't have to worry about the hot glue coming off, or perfboard becoming brittle and breaking off while in use.

I also added another PI filter after the choke in the power supply, for the 7591 grids (pins 4/8). With the bias voltage sitting around 23V, I was able to monitor the voltage while adjusting a potentiometer for that voltage. I mounted the extra sub circuit on terminal strips, mounted to the main fiberglass board.

Tomorrow (and beyond the time frame of my electronics class) I intend to make further investigations into the workings of this amp. I am overall satisfied with the outcome of my hard work. This is a far cry from the quality of my first tube amp project. As with anything, good practice makes for better results.

The last few days were probably the most intense. Building the amp (actually putting the pieces into place) is very time consuming. More so than anyone would first assume. There are commercial kits one can buy, costing anywhere from $800-$1500. I'd say, based on the time it takes to get everything together, that price is actually quite fair. I am happy that I laid this amp out from my own layouts, however. It gave me the advantage to use parts that I had and to make arrangements for custom circuitry.

Sat 12/4 (6HRS)

Sun 12/5 (4 HRS)

Mon 12/6 8:30-12:30 & 8:30PM-12AM (7.5 HRS)

Tues 8PM-12:30AM (4.5 HRS)

Putting things together

Worked on (the start of) finalizing this amp:

• Mounted transformers properly
• Mounted/installed capacitor board
• Installed RCA jacks for reverb
• Installed 1/4" speaker output jack
• installed power switch (not tight)
• Installed all front panel controls (potentiometers, switch)
• Installed 4 1/4" input jacks
• Installed/connected/soldered front panel controls ground wire
• Installed front panel wiring (shielded multi-pair or single, where appropriate)
• Installed all tube sockets
• Installed ground point at power transformer mounting screw
• Mounted front panel to chassis
• Connected choke wires to capacitor board
• Drilled extra hole in chassis for terminal strip for rectifier circuit

12/2-12/3 (7 HRS)

Solid State Power Supply

Quick bread boarding this morning - 9V power supply for solid state preamp stage. I considered a chip-regulated power supply, however, the simplicity of a zener diode resolved that choice. Here is the schematic:

I am using an integrated bridge rectifier, not individual diodes, but I don't have a simulation symbol for a bridge in LTSPice (that I know of).

7:30-8:30AM (1HR)

Painted Chassis, Heater Voltage Testing

Black painted chassis - $1 generic flat black spray paint. The sun was out long enough for the paint to stick, even if it is a California winter.

I mounted the back with 6-32 machine screws and star washers to "dig" into the chassis metal, making a good ground connection. The 6-32 threads sticking up into the inside of the chassis also lets me install terminal strips, should I need them.

I also tested the tube filament voltage with the power transformer I am using for this project, to determine if there will be any large drop. I had old tube sockets that I wired up together, on the respective filament pins. 2 7591 output tubes and 4 12a_7 type tubes. I measured around 7VAC with all tubes installed. Looks good. Now I can install the sockets I intend to use, and wire up the amp.

7PM-9:30PM (2.5HRS)

More Metal Work - Fabrication

Just when I thought all the metal shavings were cleaned up, I noticed I forgot to drill some holes:

• Power Switch
• Fuse holder
• AC inlet (grommet/strain relief)
• Speaker output (1/4" female mono)
• mounting holes for circuit boards

All of the above just happen to be somewhat important to the overall project. I had to drill some holes for the wiring entering the chassis, so I took advantage of the time to drill all the other holes needed. I am fairly certain that I have (NOW) taken care of any holes or metalwork that need to be attended to. As I type this, the cleaned panels and chassis are drying - I intend to cover them with a quick coat of flat paint, just to make the project look more homogeneous.

While drilling the last of the holes, I marked and drilled holes for an extra power transformer, if needed, to supply extra current for the tube filaments. I am uncertain that the power transformer I have will be adequate for all the tubes in the circuit. Tomorrow I will wire all the heaters up and do a quick test of any voltage drops/excess current. I have spare tube sockets that will allow me to perform this test.

Here are some pictures of tonight's mess:

5PM-9:15PM (4.5HRS)

O/P Xfmr measurements, CAD layout, chassis work, board building

Been busy since my last post.

Populated the main circuit turret board, components not yet soldered in.

 

 

METAL WORK

• drilled all holes for front panel components (pots, switches, jacks, etc.)
• aligned and drilled holes for placement of front panel
• countersunk philips head screws bolted to main chassis
• punched out holes for tube sockets
• re-drilled to compensate for existing ventilation holes
• removed bulk metal from section of main chassis behind front panel
• there is a gap between the two pieces of metal because of the way the chassis box was manufactured
• laid out and marked chassis for transformers and capacitor board
• need to drill before placement of main circuit board

I used CADStd again to assist in drawing out the placement of controls and tube socket layout. It makes it easier to place components when you can work with a 1:1 drawing and move them as needed, knowing that the dimensions do not change.

CAPACITOR EYELET BOARD

I did, however, use good old graph paper for laying out the capacitor bank board and the main rectifier diode board. Much faster.

The method I am using for these boards is similar to what Fender used in their original tube guitar amps. A layered board system with components on the top board, and the bottom one used as insulation. This way, the board can be mounted directly to the chassis without standoffs, and without danger of shorting out.

The metal holes that the components are placed into are called "eyelets". I was able to mount about half of the eyelets when the staking tool broke. I resorted to shortening the eyelets and using brute force to expand the eyelets once in their holes. The 1/8" holes that hold the eyelets were made with a hand whitney punch. There's no comparison for getting a job done using the right tool for the job. An example is most of the work I am doing now - I am trying to reuse old parts in an attempt to be "green" and to save money. A lot of the work done thus far could have been completed much faster with all the money in the world, and all the tools that go into said money.

I digress...

The capacitor board and diode board were then populated without soldering.

I also twisted some stranded insulated copper wiring for the tube filament wiring: Here is the method I chose to twist the wire:

• place the 2 wires in a drill chuck
• clamp the end of the 2 wires in a vise
• holding the wire taut, activate the drill, spinning the wires
• after the wires are sufficiently twisted, remove from drill and vise

Some thoughts on the latest work.

Metalwork is very messy. My garage workspace is now littered with metal shavings. It is also dangerous. I have had enough scrapes in the past to know how to do the work that I am doing without too much injury.

OUTPUT TRANSFORMER

Before mounting the output transformer I wanted to be sure the wire connections that I intend to make would be the correct ones. I took the transformer to school and ran a 1kHz sine wave signal into a wire I was absolutely sure (and later would be verified) was a speaker connection (secondary). I took the following steps:

• measure input voltage
• measure output voltage
• calculate voltage ratio
• SQUARE both sides (impedance ratio)
• multiply both sides by expected speaker impedance

My calculations verified the proper wire connections to make (note I had extended the original wiring). This also confirmed the "thought" around the web that this particular transformer has a primary impedance of roughly 6k-OHMS.

NEXT STEPS

• drill transformer mounting holes
• drill/mark placement for power switch
• drill/mark placement for RCA reverb connections
• drill/mark placement for main circuit board
• fabricate standoffs for boards

OVER 3 DAYS (16 HRS)

Transformer Mounts and Bias Circuits

This weekend I worked on the following:

• power transformer chassis mounts
• fabricated from angled aluminum stock I picked up at the local metal supply (Blue Collar Supply)

The power transformer I intend to use for this project is meant to be mounted semi-flush with the chassis. In it's stock form, a rectangular shaped hole is meant to be made to allow the bell-end of the transformer casing to fit into, and the transformer laminations rest on the chassis. I do not particularly enjoy metal work, probably mostly because I do not posses the tools to easily manipulate the pieces. I decided to fabricate some mounts for the transformer, so I can escape machining a large hole in the chassis. This way, I only have to drill 4 holes for mounting, and cut some angled aluminum. It went quite well.

I haven't made the holes for mounting yet. I plan to draw up a simple outline in CADSTD to have a visual size representation to work with on the chassis.

• Output tube bias circuit
• Adjustable, negative voltage power supply, full wave bridge, PI filter into 2 sealed trim potentiometers

On the turret board layout I made, I kept the output tube grid supply resistors on separate lines, with the intention that I should have provisions for separate bias supplies, to allow for compensation for mismatched tubes. I began with a voltage doubler circuit that I ended up not using, because the range of voltages was much too negative than I needed for the Fisher output stage. According to the 500-C schematic, the bias voltage in stock form is a set -17 V. I was getting anywhere from -40V to -102V DC from the first circuit.

I started over with a full wave bridge rectified power supply, grounding the positive side of the bridge. With this circuit I was able to obtain a range of -13V to -35V DC. I will have to recheck when the amp is powered up, as there was no other load on the power transformer.

I used a variable AC power transformer to "ramp" up the primary of the transformer that fed the power supply circuit. I did this because I am using old capacitors (from an old tube oscilloscope) and wanted to make sure I did not damage the electrolytic capacitors.

Here are some pics of the second version bias circuit:

 

• Output Transformer wiring
• Extended the wiring for the output transformer

Because the output transformer I am using is an old, used unit, the wires were somewhat shorter than I would like to work with. I found some spare wire and extended the existing wiring. I had to open up the transformer casing and expose the wires connected to the core.

• Recycled some parts
• I have some old terminal strip boards from previous junk-yard sales and such, and decided to put them to use. I desoldered all the old wiring/component residue from them and cleaned them up. Good as 60 years ago!

(8HRS)