17.1.09

ga-5f1

i call this amp the ga-5f1 because the fender champ 5f1 circuit and the gibson gA-5 circuit are so similar that only 5 components differ! well that is certainly true for the vintage versions. these subtle differences, however, create two different amps. the gibson has an extra 7db of gain and can really start to break up when cranked.

prototype a

this series of four builds started when i did a self-study course on vacuum tube audio frequency amplifiers for my university coursework. i needed a project, so i set out to find an amp to build. i had read some good stuff about the electrar tube 10, one of gibson's offbrand amps. learning more about this amp led me the the kalamazoo model one, another gibson off-brand amp quite similar to the ga-5. i decided that was the amp to build - the model one with a tone control.

this was my first amp build and it was the first amp that i actually built and designed from scratch. sure, i had a schematic, but i had to buy a lot of tools to drill the chassis and stake turrets on the fiberglass board. I used hammond tranformers, orange drop caps and jj tubes to build the amp.



when the amp was first fired up, i noticed a lot of hum - the old 100 ohm resistors trick fixed that one. i had some other problems with parasitic oscillation. i never figured that one out, although i did recently solve a problem that may have also solved this amps problem. the strange thing was that the amp would ring even if a dummy load was connected. it must have been an actually mechanical oscillation!

I really loved the clean sound of this amp, especially after adding a bright cap across the volume pot and with my second attemp of a tone circuit in. The first tone circuit was from the model one, but i later changed it too the tone stack used in the gibson ga-8. It sounded much better. The amp did always have a pretty harsh distortion sound, though. I eventually pieced this amp out and use the parts to build other amps because it just wasn't seeing much use since it was in a naked chassis... not that i didn't have big plans for the amp. i had designed an antique cathedral radio style cabinet for it, but the build of it didn't turn out good.



there were several issues with this amp. first the layout wasn't very good. this likely led to the oscillation problems. these problems were mostly solved by putting a 12dw7 tube in the preamp to reduce the gain a bit. not a bad way to play. the poor layout also didn't leave room for all of the components, some needed to be retrofitted to the amp with terminal strips.

in the end, this was good for my first build. it sounded beautiful when clean - clear as a bell. i certainly learned ways to improve my next build.

prototype 1

one of my aspirations is to teach classes on building tube amps, building ga-5f1 kits and hopefully extending those courses to the internet. i began this endoevour by building the first of three prototypes. the ga5-f1 is based on the classic ga-5 circuit, with a solid state rectifier and el84 output tube- much like the 2003 gibson ga-5 les paul junior reissue.

this protoype was all about functionality and low cost. it was one of my first builds to incorporate the use of radial capacitors instead of axial capacitors. (axial capacitors have leads coming out of both ends, radial capactors have both leads coming out of only one end.) radial capacitors are much less expensive, but all the layouts i have seen are made for axial capacitors. as you can see from my circuit board layout, i didn't exactly accomodate for them in my layout either.



another cost-cutting measure was the use of 70 volt transfomers as output transformers. transformers are simply magnetic cores wound with magnetic wire and have turns ratios and power ratings. power is realted to voltage and current, and current is dependant upon impedance. the 70V transformers is fed a constant voltage and there are several 'power taps' which have different primary impedances to change the power going to the speaker. theses 70 v transformers have a 50Hz-15Khz frequecy response, so they work great for guitar amp transformers.

i used a 10W transformer and with a few calculations using a calculator, signal generator and a volt meter, i found the power tap that would give me the correct impedance for an el84.

i wanted to have plenty of room for this build so i once again built it in a cakepan.

for a power transformer i used a fender standalone reverb unit power transformer. it gave voltages close to those i was looking for, so a dropping resistor was installed before the first filter capacitor. for this build i started with a 100 ohm resistor to see where that got me.

one of the unique features of this amp was that i installed two output tube sockets one for a el84 and one for a 6v6, wired in such a way that each would be biased correctly by simply swapping tubes. i did some recording with each type of power tube to see if there was a great difference in sound. i really couldn't tell much of a difference, so i opted to go with the el84 for the build because they are easier to find and modern production el84 tubes more reliable at high voltages. i got about 5.5 watts from the 6v6 and 5 watts from the el84.

the tests of this amp were a success! it sounded great, had no parsitic oscillation and the cost cutting measures didn't seem to affect the tone in a negative way.

prorotype 2

this prototype was an attempt to build prototype 1 in proper sized chassis. for the chassis i used bud 9x5x2- close to the specs of the orignal champ chassis.

this build also incorporated a switched volume pot. turning the pot down past zero flips a 'piggyback' switch that turns off the amp. this is pretty nice because now i don't have to drill a 1/2" hole in the chassis for an additional power switch.

i also added some cost-cutting jacks to this amp. in the end, they ended up being a great pain becuase they are quite difficult to plug into or unplug from. it's defintately worth the extra 25 cents to get some good jacks like neutrik jacks (still less than switchcraft).




i also tried a different wire with this amp. it had a thinner insulation and looked a little easier to work with. i only bought one color, light blue. the wire seemed to work fine, but i soon realized that it is impossible to trouble shoot an amp without multiple wire colors. being that the wire was not commonly produced in many colors, i decided against using it again.

you will also notice that i had found a way in this layout to incorporate the radial capacitors in a way that was reminicent of axial capactitors. this avoids the silly look of radial capacitors with thier butts in the air on a circuit board and makes them easier to work with and integrate with other components.



the circuit board material is used on this build is grade xx garolite from mcmaster carr. the board is 3 x 6 x 1/8" cut from a 4 foot length. its is much less expensive than fr4/g-10 garolite and more rigid than the fender type vulanized fiber boards (all of which can be found at mcmaster.)

the eyelets i used are some that i found in my boss's electrical kit. he gladly let me have them. the are smaller than the classic fender sized turrets, and they fit into the same holes as do the hoffman turrets. you will notice a mixture of eyelets and turrets on this board based one the needs of the particular electrical node.

the dropping resistor used in this build is 82 ohms, the closest to the value that gave the closest desired voltage from the duncan amps power supply designer software. using the power and output transformer from protype 1, i got a little more juice out of the el84 this time around.

another thing you will notice on the turret board is an axial fuse. the reason this fuse is there is simple- if this fuse blows something bad has happened and the amp needs to be looked at. this also frees up some room on the front panel in case any extra features required an additional pot, switch, etc.

you may notice on this particular layout that i did not use the entire 3" of space. this is cimply because many modern componets won't span the full 3".

ultimately the layout was an improvement, except for the fact that i actually had the layout flipped vertically. this also made troubleshooting difficult because nearly every wire crossed the board on it way to where it needed to go!

intital tests of this amp showed severe problems with parasystic oscillation. shilelded wiring eventually solved the problem for a while and the amp sounded great. by this time i had also found a cheap speaker to use - the jensen mod 8-20. it didn't sound as good as the weber signature 8, but it did have a bit of a smoother distorted tone than the weber.

this build was a partial success. i had attributed the parsitic oscillation to the inverted layout that cause all of the wires to cross the circuit. i would later find out that was not the case, but that story is yet to come...

prototype 3

this prototype improved upon the first two in a few ways. first, the layout problem was fixed and the layout was tidied up a bit. the bridge rectifer doesn't use as much space and the mounting holes were place in areas that were much easier accessed. other than that, this layout is pretty close to the classic 5f1 layout. i wanted a tried and true layout in order to build this amp without parastic oscialltions.

for this amp i used a new power transformer. i used a hammond 261g6 this was a bit more expensive than the one i used previously, but it had better specs and was available through my usual supplier.

you will notice that i am back to nuetrik jacks on this build also. much better!

i am also using a new wire, which you can get as a complete kit of 8 colors from radiodaze. its a great deal and very sturdy wire!

i also bought a proper cabinet for this amp- a 5f1 champ cab from mojotone. it looks great. the tweed covering isn't my favorite, but it would have cost more to go with a nice cream tolex. you will notice that the 2" chassis leaves a little gap between the top of the chassis and the control panel cutout. that problem was easily solved with some large gold piping. i later found that there exists a 9.5 x 5 x 2.5 chassis that would have avoided this problem!


zero signal electrical tests of the amp showed the amp dissipating 6.5 watts, but no redplating was occuring so i didn't bother. i did use a different 70.7 v transformer for this amp and its higher dc resistance likely played a part in this.

i was pretty dissapointed when i fired this amp up and it had the same oscillation problem as prototype 2! i spents hours trying to find the problem, even poking around with the oscilloscope. something was causing a signal to be induced into the inital gain stage, but how? these questions led to a revision of prototype two to find the answers.

prototype 2a

this time around, i wondered if the 5f1 layout was that great of a layout afterall. i wanted to modifiy it so that all of the wire paths exited the board on the same side they were destined to go. once again the layout was modified. this time i had run out of xx garolite. i had order a vulcanized fiber 5f1 board from mojotone to see what the classic fender stuff looked like, so i hacked up the "insulator" fiberboard and applied my new, improved layout.



i worked really hard to clean up all the wiring and get everything just right.

when i fired up the amp, i got the same oscillation! what? what is going on here? again, poking around with the oscilloscope showed a voltage being induced in the first gain stage from somewhere.

eventually i figured it out... it was 'blocking distortion.' the grid of the el84 was being pushed to the point that it would start to conduct, this in turn was creating a strange signal to be induced on the first gain stage that would make the problem worse. after reading some technical info, i installed a 1.5k resistor on the grids of the el84s and instantly the probems went away on protypes 3 and 2a. what a relief! now they both sound great!

in the end, these builds helped me learn a lot about tube amps. one of the most important things i learned however was take your time and do things right the first time around. rushing things only leads to poor design choices. luckily most of the time they won't matter much, but every once in a while they do..

so now that i've built 4 ga-5f1 amps, what is next? well, i just happen to have enough sapre parts layoung around to build a gibson les paul junior / fender 5c1 champ ( both have a 6sj7 pentode preamp and 6v6 power amp.) this amp will also incorporate another crazy idea i had- using a power transformer as an output transformer! stay tuned!

8 comments:

Ray said...

Too bad the cathedral amp never came to fruition; I thought it would've been sweet.

It's pretty enjoyable to read about the history of the amps that existed in our apartment, thought I don't completely understand everything. I guess that's what I get for not paying enough attention while learning about circuits in my general physics class....

Anonymous said...

Having just replaced two EL84 tube sockets on a Gibson Goldtone amp(2000 production date)and thinking how simple it looked inside...then to see your creation in it's simplicity, I'll be building an EL84 as soon as I can order the parts!
Thanks for the detailed history of your trial runs.

NewTubeFan

lookah said...

hi...your amp looks sweet. can i ask a newbie question? i'd like to install a bright cap on my gibson ga-5 reissue. (i've removed the 1M/22pf cap tone filter but it's too dark at low settings with my p-90's.) which terminals do i solder it too? i couldn't see it in your pics...

thanks!

luca

davidandbetsy said...

luca,

I actually installed a bright cap on my first GA-5 build. I just put a capacitor across the 'Hot' terminal and the wiper on the volume pot. Buy a couple of values and see what you like best. I can't remember the value I used something around 200pF. Try values between 100pF and 500pF. I do remember it having a beautiful bell like tone when it was clean - putting in a 12DW7 tube gave it a lot more clean headroom also. I just had to swap tubes to go from a clean amp to a dirty amp.

lookah said...

thanks! i will do that. i have both a 100pf and 250pf sitting around. btw/ i'm also going to try adding the 47k negative feedback resistor. will either a 1/2 watt or 1 watt work? and does the type of resistor matter much for negative feedback circuits...carbon comp or carbon film?

-l

davidandbetsy said...

lookah, i used a 1/2 resistor, it shouldn't matter wether it is carbon comp or carbon film.

lookah said...

great, thanks again!

Dan said...

Thanks for posting this, I was looking for comments on using radial caps with eyelet boards and found you. It is crazy to pay $7.00 for a $2.00 cap just because the leads stick out the ends. I am going to redesign my eyelet tonight and use the radial caps I already have.