The finished installation. It is very difficult to tell that there is a modern silicon diode hiding back there.
This amp was given to Perry by a friend. (I should have such friends).
Evidently it was found in a garage somewhere, an old lady's husband had bought it new, along with a "matching Fender guitar", which she sold. (Imagine what that was...)
They plugged it in, exactly the wrong thing to do to an old amp, and it blew a fuse. So they asked Perry if he wanted it. Perry said yes.
Note the rust on the snap-on tube shields. I cleaned those up a bit, not too much. Also note the Selenium bias rectifier (the green square thing). More on that later...
After discussing the project for about a year, Perry Nichols finally relented and sent me his 1958 tweed Tremolux for restoration.
The amp has many stories to tell. It was made in 1958, and was soldered together by Lily - her name is on the little piece of tape inside the chassis.
The Jensen P12Q speaker was made in Chicago in 1957. The amp is almost completely original. The output transformer was replaced sometime around 1960, with a Stancor A-3880 (also made in Chicago). The Tremolux was known for blowing up transformers, it might have something to do with the fact that the original Triad 108 transformer was rated 15 Watts, and the Tremolux routinely put out more than 18 Watts.
This era of tweed Fender amps used a selenium diode in the bias section. The selenium diode is the little green square thing to the left of the eyelet board, off by itself. The diode consists of 8 plates in a selenium "sandwich". Each plate represents about 10 Volts. So this bias diode supplies about 80 Volts. (negative 80 Volts, this is the bias voltage).
Selenium was a reasonable technology in 1958, but it is untrustworthy today, and if it fails and burns it emits a very toxic, foul smelling gas that is not good for humans. I replace these as a general rule. I put a 1N4005 diode and a resistor in series behind the selenium rectifier, leaving the selenium unit there for original appearance.
Here is the modified selenium diode: the silicon diode and resistor are hidden on the back side, covered by a little square of fish paper to insulate in from the chassis. The diode/resistor is also in shrink tubing.
I tested this selenium diode in circuit and it delivered -78 Volts. I selected the resistor to go with the 1N4005 voltage drop, in order to result in -78 Volts. The resistor is 100 Ω.
I use the original upper solder tag for connection to the bias voltage source from the transformer.
I checked all the electrolytic capacitors, as well as key resistors and caps for value and condition. I check the capacitors for ESR and leakage. As expected, the filter capacitors were not reliable - original Astron Minimite 20/500s. The cathode bypass caps were also Minimites as original, but they had drifted to more than 50 µF from the specified 25 µF. The bias Minimites were also bad.
All the Minimites were restored with new capacitors inside the original shells. This leaves the amp looking exactly as it did when it left the factory. You can read about my method of restoring old Astron Minimites HERE.
Stay tuned for further updates. I am waiting for the re-coned speaker, new tubes, and some miscellaneous items, and I have some tuning and restoring to do on this great amplifier.
The Fender 5G9 circuit has 470Ω, 1W screen resistors on the 6V6s - but they are absent in this amp. Instead, this amp has 1500Ω. 1/2 W grid resistors. They appear to be original to the amp, installed at the factory, so I suspect this was an example of "Change on the Fly" design method of Fender in the Tweed years.
Once I had everything sorted, I finally brought the amp up slowly on the Variac, and arrived at 122 VAC (wall voltage) without any big noises, sparks, or smoke. Yay.
The output tube plate voltages were quite high at ~415 V. I quickly checked the bias and then shut off the amp, as the tube current was more than 50 mA! Yikes. 400+ Volts on the plates is not unusual for these amps, but the bias was way off. I temporarily clipped a 100K potentiometer in front of the 56K bias resistor, and adjusted the pot until the amp sounded good and was a more reasonable 28mA. I read the value on the pot, and found a resistor of close value to solder in place. I am ordering new TAD 6V6STR tubes for the amp, they are designed for high voltages, and so I will probably have to adjust the bias again when they get here.
I was playing though my shop speaker, which is a 1956 Jensen P12P, and the amp sounds really good. The tremolo is superb. With the tremolo on, even a low level, you do not miss reverb at all. The tone is wonderful.
The chassis and everything else was coated liberally with grime. As with most restorations, thorough cleaning is the first order of business. I usually blow off most of the loose crud, and then clean with alcohol, using a toothbrush and a little paint brush. If there is corrosion, bad solder joints, broken wires, etc., I check for it at this time. I take the amp completely apart. The cabinet is cleaned separately with Murphy's Oil Soap. I have tried every other cleaning agent on old tweed, and Murphy's works the best, without loosening the glue, staining, or removing "patina". Plus, it leaves the cabinet smelling lemony fresh!
The original speaker is a Jensen P12Q from 1957. During testing, the speaker displayed a problem. On low-pitched, decaying notes, there was a noticeable raspy buzz from the speaker. Sounded like voice coil rub. And, of course, that's what it was.
I removed the dust cap from the cone and blew out the gap, no change. The gap was very small - about 0.004" - I could barely get a piece of paper in there to shim it. These old Jensens had tight voice coil gaps. that's one of the reasons they sound so good, in my opinion.
I tried an old hack to try to re-center the voice coil: I shimmed the side where the rub seemed to be worst, and then saturated the whole cone with isopropyl alcohol. Repeated several times while moving the cone around manually, this sometimes removes stresses from the cone and acts to re-center it, eliminating the rub. Except not this time. (sigh)
This speaker is off the Brian at Weber for re-coning. I specified a 30 Watt coil and 8 ribs, and no doping, as original. Brian has done a number of Jensen re-cones for me in the past, and he always does a superb job.
The tube chart indicates that the amp was produced in September of 1958. The circuit is designated as "5E9", but a look at the layout shows that this amp is a 5G9, the "new" model in 1958. Leo Fender had a habit of using old tube charts until he ran out of them before going to the newer, correct tube chart. The Tremolux circuit had many of the same features as Fender's top-of-the-line amps - like long-tailed pair phase inverter, bias tremolo, power choke, etc.