Download mikede30.pdf

Electar Tube 30:
Rebuilt with Reverb and Tremolo
by Mike Donovan
Background
I bought this amp from Musicyo.com a couple of years ago. In stock form the clean tone was muddled, and the
overdrive was simply awful. I first replaced the speaker. This helped somewhat. I tried to do some mods, but
the cheap PCB construction led to much frustration so I gutted the amp.
At the time, Steve Ahola was experimenting with the Trainwreck Express circuit in his Electar 30, so I tried
some of the ideas he had generously presented to fellow Ampagers. I ended up with something like Steve’s,
except with a paraphase inverter instead of the split-load. It sounded pretty good, but I wanted reverb and
tremolo. So I gutted the beast again and here is the latest revision.
Power Supply
Pretty standard affair. I added the two 100 ohm 5 watt resistors to knock the B+ down from 390v to 355v. This
also simulates the “sag” characteristics of a tube rectifier, and limits the initial surge of current when firing up
the amp. I added a standby switch. I tied the filaments to the power tube cathode resistor through 100 ohm
resistors, an old hum reducing trick.
Preamp
Based on the blackface Fender, with a few mods to reduce the bass response. I reduced the 1st stage cathode
bypass cap to 1uf. The “Marshall” values of 2k7 and 0.68uf work good here too. For the tone stack, I went
with the “Marshall” values for the bass and mid caps (22n). Sometimes I prefer the Fender values of 100n bass
and 47n mid, sometimes I go with 33n for both, it depends on the rest of the amp, the speakers, etc. I prefer to
wire the midrange pot blackface-style. I used a 82pf bright cap rather than the usual 120pf mainly because I was
fresh out of 120s. It works fine.
The 2nd stage cathode cap is reduced to 4.7uf. The coupling cap coming off the 2nd stage is 10n. These things
combined help reduce the “flatulence” of the stock Fender circuit.
Phase Splitter
Standard Vox long-tailed pair. I fed the reverb into the normally unused half of the splitter like the Matchless
Chieftain. I also fed some negative feedback there in order to cut down on some hum and hiss and to tighten up
the power amp stage. It did sound pretty cool without the NFB, but it was noisy and unrefined. The 2M2
feedback resistor forms a voltage divider with the 100k reverb pot, so it’s roughly equivalent to the Marshall
100k / 5k divider. I tried a cathodyne splitter and used the other half of the 12AX7 as a gain stage, like the
Princeton Reverb topology. That did produce much more reverb, but I didn’t like the overdriven sound as much
as the long-tail.
Power Amp
Cathode biased. I originally had a fixed bias / cathode bias switch, but found I was using cathode bias most of
the time for that “Vox” feeling. Also, at this point I became distressed at the general “spaghetti” situation and
wanted to simplify things, so I ripped out the fixed bias. I used 1k screen resistors because I was concerned
about the relatively high voltages on the EL84 plates and screens. I first tried a cross-line master volume
control, but settled on the dual pot method. I found the junction of the two pots a convenient place to hook in
the tremolo. I moved the output transformer to a point in between the power tranny and the EL84s. It originally
was sitting over by the input side of the amp with 18” leads connecting back to the power tubes and speaker. I
added a speaker jack that grounds the output when nothing is plugged in.
Reverb
At this point, I had used up both 12AX7s. I wasn’t sure if the Electar power transformer could safely supply any
more filament current, so I went with a solid state reverb driver and recovery stage. I’ve had great success with
the JFET / MOSFET cascode circuit. Notice the voltages on the reverb recovery stage, they come very close to
the 12AX7s fed from the same supply node. The driver stage is wired a little differently to accommodate a
wider input voltage swing. Due to space constraints, I built the circuits on perfboard with flea clips. Ideally, I
would have used eyelets. The MOSFET in the driver stage requires a small heatsink, the recovery stage does
not. I used a standard Fender reverb transformer. The performance is excellent, as good as the blackface tube
circuit, to my ears. Unfortunately the Electar cabinet will not accommodate the long tank so I am forced to use a
short tank. Also, by feeding the reverb directly into the phase splitter, there is not as much reverb gain as the
Fender blackface circuit. Normally I only use a small amount of reverb, so this is really not too bad.
Tremolo
I was faced with the same problem, out of tubes. I hunted around for some solid state tremolo designs but most
of them used op amps or low voltage components, so I figured, why not give the standard JFET / MOSFET
cascode circuit a chance? It worked perfectly on the first try. It can be adjusted to a very intense tremolo.
The sound
Well, it’s a Fender preamp with Marshall voicing into a Vox AC-15 power stage. Good all around “old school”
rock and roll type of sound, touch-sensitive and all that good stuff. Nice Vox-like “chime” due to the cathode
biased EL84s. The master volume can be used to dial in a crunchy sound at lower volumes but it does not do the
saturated Mesa Boogie thing. When the master is dimed, it is effectively out of the circuit and you can get the
classic power amp distortion. The amp is plenty loud for an enjoyable listening experience in a civilized band
situation.
Other Stuff
I didn’t use any fancy components, mostly carbon film resistors and polyester caps, whatever I had laying
around. I use the original Sovtek tubes that came with the amp. I did try some other tubes but it didn’t make a
whole lot of difference to me. I used a star ground method. The hum and hiss on this amp are comparable to the
average blackface Fender. I can usually do much better, but with the horrible way this amp is laid out, I
considered average to be a victory. Due to space constraints, I went with a very minimalist power supply, I
would normally decouple every stage. This amp is a beast to work on due to the positioning of the controls, I
wanted to add some other stuff like a presence control but ran out of panel space and dreaded running any more
wires.
Here are pictures of the whole inside, the reverb section and the power supply / tremolo. The schematics for
both the cathodyne and long-tailed version follow on the next pages.
D
100k
1W
179v
250p
2
250k-A
Treb
100k
1M-A
Vol
33n
68k
1
3
1.43v
1M
D
1uf
100k
1W
177v
7
1.46v
4.7u
25v
250k-A
Bass
1k5
C
47p
10n
100k
1W
285v
1M
470k
6
8
10n
1
2
10k
33n
2k7
10k-A
Mid
To
From
Rev. Tank Rev.Tank
220k
330k
172v
2
500p
1
68k 3
1M
Power
2N5485
470
IRF820 Footswitch
10/50v
3
12V
1
100n
100 /
5w
N.C.
100 /
5w
120 VAC
Fuse
1.5A SB
1
E
100 / 1W x 2
Filaments
100k
1W 10n
1k5
1k5
7
9
130
5W
11.8v
47u
50v
3
2
56k
1W
352v
1k 5W
E
B
A
1k 5W
9
7
220k
250k-B
100n Intensit
y
100k
10u
330k
50v
1W
2
1
2N5485
1.08v
250k-A
dual
22n
8
Master
59v
1M
2
3
B
2
IRF820 100k-A
3
Rev
68k 3
2
1k5
252v
6
D
IRF820
12V
1
10/50v
330k
22n
FB
12V
B
47
352v
56k
1W
22v 7
3
3.2v
10u
25v
1k5
C
68k
2
3
2N5485
1M
1k5
1.5v
1k5
355v
338v
310v
275v
A
B
C
D
Standby
1k 5W 4k7 1W
47u
450v
47u
450v
22u
450v
22n 100k
3
1
22n
22n
1M
1M-B
Speed
2
Voltages in tremolo
circuit vary due to
oscillation
Footswitch
10k
22u
450v
Electar Tube 30
( Modified )
Revised 11/27/01 Page 1 of 1
Michael J. Donovan
FB
D
100k
1W
179v
68k
22n
3
1.43v
1M
250k-A
Treb
100k
1uf
1M-A
Vol
7
10n
22n
2
6
47k
1k2
1k5
1M
10n
2M2
To
From
Rev. Tank Rev.Tank
172v
2
2
IRF820
IRF820 Footswitch
2N5485
470
1M
Power
1
3
1.08v
1uf
100 /
5w
N.C.
100 /
5w
E
100 / 1W x 2
Filaments
C
100k
1W
22n
1k5
245v
3
130
5W
E
11.8v
47u
50v
3
9
2
12V
1k5
IRF820
1
68k
2
355v
338v
310v
275v
A
B
C
D
47u
450v
4k7 1W
47u
450v
22u
450v
1k 5W
7
3
22n
22n 100k
3
1
2N5485
1M
1k5
1k 5W
A
250k-B
100n Intensit
y
100k
1W
100k-A
Rev
1.5v
1k5
Standby
B
220k
2
100k
1W
1k 5W
FB
10/50v
3
2N5485
2
120 VAC
Fuse
1.5A SB
68k
1
12V
10/50v
3
12V
68k
250k-A
dual
Master
10u 330k
50v
500p
1
100k
1W
B
1k5
3
7
100k
1W 10n
330k
1k5
1
249v
D
220k
335v
6
8
1k5
V3, V4 EL84
352v
7
2
9
352v
10k-A
Mid
330k
1
65v
63v
22n
B
3
1M
8
1.46v
4.7u
25v
250k-A
Bass
1k5
V2 12AX7
100k
1W
177v
250p
1
2
D
V1 12AX7
22n
1M
1M-B
Speed
2
Voltages in tremolo
circuit vary due to
oscillation
Footswitch
10k
22u
450v
Electar Tube 30
( Modified )
Revised 11/29/01 Page 1 of 1
Michael J. Donovan
FB