Guitar Tube Amp From Scratch (Part 1, Transformers)

Guitar amps are really pretty forgiving to build.  There are plenty of ways to do it wrong, but those ways are rarely self-destructive, just kinda crappy-sounding.  Figure out the crap, change it.  That's the great thing about point-to-point valve amps:  they're easy to fiddle with.
This is going to be the first in a series on Building a Guitar Tube Amp From Scratch.   The series will demonstrate the creation of a guitar amp from a random set of transformers.

I've been volunteering at a recycling/reusing facility that takes donations of appliances (computers in particular), refurbishes and remixes them, then sells them. The facility pursues the noble goal of providing computers and internet access to people who otherwise wouldn't be able to afford them. This effort is subsidized by rendering broken donations down to their base elements and selling these raw materials to other recycling facilities.

Step 1:

To business! At three dollars a pound I get to take home the shiny stuff. Exhibit A: three transformers taken from a tube amp of some sort. When reusing parts there are three useful circumstances to take note of:
I: Model numbers of the amp you are dismantling:
A: This lets you look up schematics. Just about any tube amp is halfway to being a guitar amp; usually fiddling with the gain structure will get you something musical so the chassis, sockets, and components can be reused.
B: In this case the amp had at some point made someone feel sad and impulsive, so most of the chassis was no longer usable or recognizable. Remember kids: the gear's not the @#%#, you're the @#%#.
II: Model number on the transformers:
A: Lacking a schematic, one can always search for the transformers. The internet is just one big pile of information. And cat pictures . Even a scan of an eighty year-old electronics catalog could have the info you need. Why is there a scan of an eighty year-old electronics catalog online? Who knows? It's the internet.
B: My Google-fu was no match for this iron. No matter: I have one final refuge.
III: What valves the chassis was populated with. This lets you know:
A: Roughly how many watts the Output Transformer can handle
B: How much heater current the Power Transformer can supply
C: Roughly how much high voltage (B+) the Power Transformer can supply. Different configurations and biases will use up different amounts of current, but this should get you quite close. On Hi-Fis the Choke will likely be able to handle whatever B+ the Power Transformer can dish, but not always. For our little demonstration the Choke will be downstream of the output stage, so if they came from the same amp the Choke should be hardy enough.
D: It can get complicated. Ask questions and check out some of the great forums and libraries available online . Keep in mind that all forums have search functions.
E: I'm not trying to scare you. Ask questions but GET TO WORK.

Step 2: Power Transformer

  Five valves are scribbled over my metacarpals: 12AX7s *2, 6V6 *2, and 6X5. So I need to find their datasheets and add up their current draw.  Duncanamps.com will distill most datasheets to a few key pieces of information, plus supply links for additional reading.
  Most commercial amps use all available amperage, so the current the existing valves drew is all you can expect to use.  Adding up the valves tells me that I have 2.1A at 6.3v for heater current and about 77mA for the B+. Note that the datasheets tend to list the B+ at idle, or when there is no sound to amplify.  This simplifies things-- kind of like how high school physics formulas are excempt from gravity-- but it'll get us in the area.  If another tube is substituted, just make sure the valve has an equal or lower idle rating.  If all else fails; power the thing up and keep a finger on the power transformer.  If it gets hot to the touch, try, try again.

Step 3: DC Resistance on the Output Transformer

 All we need is a small AC wall wart (anything between 4 and 10 volts AC would work), a multi-meter (we'll only need the AC voltage and the ohm-meter), a 1 Ohm / 1 Watt resistor, and that transformer of unknown quality.
 If the wires on the transformer lack any distinctive coloring, it would be handy to create labels.  This can be as simple as stripes of color using markers or folded tape with numbers.
 Using the ohm-meter, test all the possible combinations of wires to see what connects to what.  Usually the reading will be very low,  as a high DC resistance is rarely desirable in a transformer.  Draw out the connections and make note of the resistances. 
 On the output transformer I found four separate windings: two with two connecting wires each and two more that each had a center tap.
  For example: The yellow wire was 17.6 Ohms away from the red wire, but 8.8 Ohms away from the red/yellow.  I checked the resistance between the red/yellow wire and the red (8.8 Ohms) and confirmed that the red/yellow wire was directly in the middle of the red and yellow wires.
  Notice that the black-red-green winding has a slight imbalance.  This is fine.  The DC resistance is a good indicator but it only hints at the winding ratio; and the winding ratio is what we're interested in.

Step 4: Impedance Ratio on Output Transformer

 To find the exact ratio I hooked the AC transformer (with the safety 1 Ohm /  1 Watt resistor in series) to the black and green wires. I used them because they had the highest resistance, which hinted at the highest ratio. This way, I could be fairly confident that the voltages formed on the other windings would be lower.
  Take note first of the exact voltage the wall wart is giving out. The transformer and resistor you're attached to won't present much of a load so the voltage will be higher than the wall wart is quoted for. In my case, the wall wart rated for 9vAC was giving out 10.27vAC.
  Let this be a lesson in transformers! They are not a load; they transform loads. This is why tube amps need to have a speaker plugged into them when turned on. A load must be presented to the output valves, or they trip over themselves.  It's like expecting one more step at the top of a dark staircase; only several thousand times a second.
  With 10.27vAC on black-green I found 317.2mvAC on yellow-red. Blue-black/white had 2.273vAC and green/yellow-black/yellow had 1.760vAC.  This confirmed that black-green had the highest ratio, which suggested that it was the primary winding. 
  I first turned my math to the yellow-red winding.  An impedance ratio is the voltage ratio squared, or, (VoltageIn / VoltageOut)^2.  10.27vAC (the voltage on one winding) divided by 0.3172vAC (the voltage on the other), squared, equals 1048.  My goal is to hook it up to an 8 Ohm speaker-- so 8 Ohm times 1048 equals 8384 Ohms.  This is a push-pull Output Transformer so each output tube will see half of that, or 4192 Ohms (we'll call it 4k2 Ohms).  This is a pretty good load for most valves, with a bonus of having that red/yellow tap which would allow us a 4 Ohm load.
  If a 16 Ohm speaker was the goal, then the primary would have a 2k1 Ohm winding per tube, with the red/yellow tap offering an 8 Ohm speaker connection.
  The blue-black/white and green/yellow-black/yellow windings we'll keep in mind for feedback, or simply keep unconnected.  With a typical speaker cabinet they offer loads far too low or too high to be feasible.  For instance, with an 8 Ohm speaker on the blue - black/white winding we'd have a  ~163 Ohm load for the valve.
  Most valves would have a hard time driving anything that low.  It would be possible, but not without what I'd call unreasonable sacrifices--superfluous and complicated, etc.  I won't be unkind tho-- it might have some interesting sounds to it, or allow an unusual valve-- please post what you find.

Step 5:

The output transformer is a back of the envelope kind of formula, and a power transformer is even easier to work out.  Just plug the known quantity (utility voltage, 120vAC in the US) into the voltage ratios we worked out (10.18 : 41.7 : 0.617 = 120 : 492 : 7.27).
With fullwave rectification we'd attach one high voltage secondary to each anode (492/2 = 246) giving us 344vDC (246*1.4 = 344), not counting losses like rectifier drop and any filtering.
We know that the heater voltage should be 6.3v, but here it is at 7.27v.  A couple of things contribute to this: Using a low voltage wall wart on a possibly fluctuating wall outlet, testing with an old multimeter, and extrapolating upwards quite a bit with those divulged numbers would let some gremlins in.  Finally, utility voltage in the US has been slowly creeping upwards.  The transformer might be expecting a lower voltage.
No matter.  There are plenty of ways to fix the heater supply once I know how it'll act in situ.
I know now what wires go where and the approximate voltages.  Looking at the amp the transformers came from I know the amp they will become could have a valve rectifier, two preamp valves and a low wattage push-pull power amp.  I could reshuffle a bit, remove the valve rectifier to free up some heater current, or use lower-current power amp valves and and apply that to additional preamp valves--I'll go over some of that in future instructables.

This is the first step in a series on building a guitar tube amp from found transformers. Let me know what you found confusing or helpful in this stage, and what you hope to see in the future.

Thanks~
ASDTerror!