Gibson GA45-RVT low power

[BigE]

Had a guy bring me this GA45 RVT that somebody had gotten inside and put a non orig PT and also added addition Filter caps and rewired the existing ones.  It was a mess.  The owner wants it put back to factory specs, wiring, etc.  So I put in a new PT, put in new filter caps and fired it up.  Thing is I got hardly any volume on either channel.  You have to turn it up to 10 to get any kind of audible signal.  and even with the volume turned to zero you can still hear signal bleeding through the speakers from somewhere, but I can't find where!   All my voltages look good and the wiring looks correct.  One thing i did notice is that if I disconnected the roach from the vibrato channel at Pin 5 V4 it got more volume.  But the signal I'm getting is more of an AM-Radio type sound.  Not a normal good healthy signal.

Anybody got any suggestions on where to start/check? (other than recheck your wiring.)

Thanks in advance.  this one is making me pull my hair out.

http://www.webphix.com/schematic%20heaven/www.schematicheaven.com/gibsonamps/ga45rvt.pdf

[HotBluePlates]

You have any way of injecting a known good signal? Like an audio oscillator or signal generator?

If it were me and I had such a gizmo, I'd figure out how big a signal would be needed to get a good loud output from the output tubes/OT/speaker, and would inject that signal at, say, the phase inverter input.

Verify a good signal through to the speaker from that point to the speaker, then move backwards towards the input jack. The injected signal must be made smaller as you get more gain stages involved. When you go from good strong output to weak output, you've found a stage that needs a closer look.

If I wanted to make the process very fast, I'd pick a spot halfway between the input jack and speaker, and inject an appropriate signal there. If good, find the halfway point between that spot and the input jack, and inject there. If still good, inject halfway between that new spot and the input jack, etc. This verifies circuit sections as fast as possible.

[sluckey]

Is this the same amp you asked about back in July?

[BigE]

Yes, it's the same amp.  I'm just now getting back to it.

[BigE]

did some more signal tracing this AM on the vibrato channel.  I loose signal on pin 5 of V4A. ( If I lift the leg of the roach that connects to Pin 5 I get a lot of volume back).  Weird thing is i also have strong signal on the left side of R72 and none on the right.  and i don't have the reverb pan hooked up.

On the Normal channel I get a stronger signal all the way through the amp, but you have to almost turn it up to 10 to really hear it.  And it sounds nasily, almost out of phase-ish.  Not full and normal.

I did notice that there is no V7 tube.  There wasn't one when the amp came to me.

I would think it's got to be something in the power section since that's the only thing I've really touched, i just can't figure out how to pinpoint what it is.  Rechecked the wiring again and all looks OK.  Voltages run a little high but nothing i would think would cause all this.

any ideas?

[silverfox]

I occasionally pull apart my Univox amp. If I forget to connect the Reverb circuit when putting it back together again, the amp won't work. Something to do with feeding the Reverb back into or the signal path is cut when the Reverb is not connected.

I looked up the OA2 tube- V7. It is a voltage regulator for the Reverb section. Hmm... I went back and looked at the circuit and this is definitely different then what I've seen before. I don't get the Presence circuit at all. It looks like just a very basic resistance filter.

Anyway, try putting V7 in.

My 2 cents and I hope someone with better accounting will help me understand the books on this one.

Silverfox.

[quayhog]

Go back to basics.  get one channel working, disconnect all tubes from untesting channel.  Concentrating on Power supply.  the OA2's either work or they don't If it fires its likely OK.  regardless they're cheap.

If power supply is working properly, and the 6L6GC are conducting current and have reasonable plate voltages and you have nearly -50 VDC bias voltage on the grid and all grounds are verified good You should have the makings of a good and reliable amp.

My experience with Gibson/Epi amps are they can be a bitch to work on. The tag strips dry out and become brittle over time.  They break easily, somtimes breaking a connection with it.

After you get the power amp running get a preamp working, one at a time.

If all power amp voltages are as they should be, check all grounds.  I've seen busted grounds, and they seem to be very common in these old Gibson amps.  Take it one step at a time. 

[HotBluePlates]

... I loose signal on pin 5 of V4A. ( If I lift the leg of the roach that connects to Pin 5 I get a lot of volume back).  Weird thing is i also have strong signal on the left side of R72 and none on the right.  and i don't have the reverb pan hooked up. ...

I did notice that there is no V7 tube.  There wasn't one when the amp came to me.

Problem with Reverb channel signal. Missing tube in Reverb channel. Reverb tank not hooked up.

Maybe plug in all the Reverb components?

... i just can't figure out how to pinpoint what it is.  ...

Signal Injection, which I already described.

[Jack1962]

V7 regulates power for the reverb and also the the roach.

[sluckey]

V7 regulates power for the reverb and also the the roach.

It only supplies the tremolo circuit, not the reverb circuit.

[HotBluePlates]

V7 regulates power for the reverb and also the the roach.

It only supplies the tremolo circuit, not the reverb circuit.

Yes, but the trem roach is on the reverb channel, right?

Question: do those roaches usually become a big or small resistance with a lot of light?

If V7 is yanked, Node C voltage has to pull up to a voltage equal with Node B, right? I dunno if that might light up the roach brightly (might depend on the setting of R73, which may be an internal trim). But if we assume it causes the roach resistance to become small, wouldn't that tend to shunt R72 and V4A grid to ground, killing the reverb signal and Reverb channel dry signal?

Seems worth at least looking at...

[sluckey]

Question: do those roaches usually become a big or small resistance with a lot of light?

More light would equal less resistance.

Your thoughts about the increased node c voltage due to the missing V7 make sense to me. With the increased supply voltage to V3A the lamp would burn brighter due to increased current thru V3A. I think the fact that the regulator tube and the R73 bias pot are used indicates that it's important to set the quiescent current thru the lamp.

I would just pull V3 and see if the volume comes up. If so, then V7 is necessary and R73 may need a tweak.

However, there are issues with the normal channel also. I suspect a common problem (such as power supply or power amp). I would just forget about the reverb channel and make the normal channel happy. Once the normal preamp and power amp are working properly, the reverb channel may very well come to life too (well except for the missing V7 issue). 

[BigE]

got around to checking voltages on the power tubes tonight.

Got 436V on the Plates/Screens, 0v on the Cathode, -50 on the grids.  But current on one tube is 27ma and 35ma on the other.

Now keep me straight here...  isn't that way low current?

[HotBluePlates]

...  But current on one tube is 27ma and 35ma on the other.

Now keep me straight here...  isn't that way low current?

Does it matter?

Low idle current, big bias voltage, big input signal for big plate voltage swing might be the intent for the OT in that amp. -50v is dead-nuts on the schematic bias voltage.

I don't think this is why the amp don't play. At least, not by itself. Again, signal injection would tell ya where the amp stops passing a strong signal.

[BigE]

Regarding Signal Injection...   How do you know what an acceptable/normal size signal for each stage is?  and what hz signal would you use?

(I'm learning here, so bear with me.)

thanks again...

Also, I still don't have V7 in cause I don't physically have one.  I have one on the way though.

[HotBluePlates]

... what hz signal would you use? ...

400-1000Hz, if you have a signal generator/oscillator of some kind.

Any kind of music output (ipod, CD player, etc) if you don't. (determining level will be a little harder.)

... How do you know what an acceptable/normal size signal for each stage is?  ...

Work output to input, just like designing an amp.

- Output tubes have a bias of -50v, so 49-50v peak to each tube; or 98-100v peak-to-peak or ~70v RMS from grid-to-grid on the output tubes.
   * You might not be able to develop that directly, so you apply a signal to the driver stage or phase inverter input until you get that total voltage at the output tube grids.

- V6 is a 6FQ7, so gain is probably around 16 or so. 50v peak at the 6L6 grids requires 50v/16 = 3.1v peak or 2.2v RMS into each grid of V6. That's a push-pull signal, so 6.2v peak or 4.4v RMS grid-to-grid.

- V5A is a 6EU7 split-load inveter, so gain will be less than 1. Call it 0.9 out of convenience, so 3.1v/0.9 = ~3.4v peak or 2.44v RMS on this grid should drive the 6L6's to 50v peak each grid.

-Stuff between R18, R19 and V5A grid is a notch filter, so ignore that. Or if you suspect a serious problem with it, run a wire from R18/R19 to V5A grid to bypass.

- R18, R19 are mixer resistors but also form a voltage divider, so Normal channel strength is reduced to 1/3 while the Reverb channel strength is 2/3.
   * For Normal channel, you need 3.4v / 0.33 = 10.3 v peak or 7.4v RMS output at the plate of V4B for full 50v peak at 6L6 grids.
   * For Reverb channel, you need 3.4v / 0.66 = 5.15v peak or 3.7v RMS output at the plate of V4A for full 50v peak at 6L6 grids.

- Normal Channel: V4B is a 6EU7, but high plate load puts gain at around 80 (maybe a bit more or less). 10.3v peak / 80 = ~130mV peak or ~93mV RMS for full 50v peak at 6L6 grids.

- Normal Channel: stuff between volume control and V4B grid is another notch filter; prior to that is a James tone circuit and a volume control. The James circuit will probably be ~-20dB loss, or 1/10th output (assumes Treble/Bass controls at half-way point).
   * Assuming volume control full-up, you need 130mV peak / 0.1 = 1.3v peak (0.93v RMS) at V1B plate for full output at 6L6 grids.
   * For volume at 1/2, loss will be another 0.1, or 1.3v / 0.1 = 13v peak (9.3v RMS) at V1B plate for full output at 6L6 grids.

- V1B is a 6EU7, and with value shown gain will be ~61.
   * Full-up volume requires 1.3v / 61 = 21mV peak (15mV RMS) for full 50v peak output at 6L6 grids.
   * Half-up volume requires 13v / 61 = ~213mV peak (~152mV RMS) for full 50v peak output at 6L6 grids.

Similar for Reverb channel, but trem complicates things at V4A's grid (might need a little more signal to smack the 6L6's fully). Also, reverb circuit levels are their own thing, apart from the dry signal path.

Sluckey is rightly pointing out that "hardly any volume on either channel" means to start looking at V5A and V6 for problems, probably by injecting a strong signal at V5A's grid and seeing if you get strong output to the 6L6's and at the speaker.

[BigE]

ok, tried my hand at a little signal injection tonight and here's what I came up with.

Everything looks pretty good until I put a 10v peak on pin 8 on V4b.  when I did that I got a nasty distorted 150v Peak on the power tubes.

I also noticed that when i injected signal on the grids of V6 that pin 2 was louder than pin 7.

I'm kinda stumpped now.  I would think with symptoms of no volume the signal would just quit not get super loud and distort the wave??

[HotBluePlates]

Everything looks pretty good until I put a 10v peak on pin 8 on V4b.  when I did that I got a nasty distorted 150v Peak on the power tubes.

- Normal Channel: V4B is a 6EU7, but high plate load puts gain at around 80 (maybe a bit more or less). 10.3v peak / 80 = ~130mV peak or ~93mV RMS for full 50v peak at 6L6 grids.

Don't forget to read to the end for the final voltage to use! 

I also noticed that when i injected signal on the grids of V6 that pin 2 was louder than pin 7.

The easier thing is probably just inject at V5A grid, and let it develop the 2 push-pull outputs to test everything from that point to the speaker.

I would think with symptoms of no volume the signal would just quit not get super loud and distort the wave??

Since you now know you had way too much signal at V4B, it seems you can continue moving backwards towards the input. At this point, for the normal channel I'd just be injecting the signal at the input jacks.

[BigE]

Tonights update:  I injected a 130mv signal into V4b Pin 8 and only get a 10-12v peak signal at the 6L6 grids.  So I'm assume we're getting closer.  gotta be something between V4 and V5 right?  How do I narrow it down further?

thanks in advance! 

[HotBluePlates]

Tonights update:  I injected a 130mv signal into V4b Pin 8 and only get a 10-12v peak signal at the 6L6 grids.  So I'm assume we're getting closer.  gotta be something between V4 and V5 right?  How do I narrow it down further?

Look at the schematic.

V4B's cathode bypass cap is a 0.1uF, so if you're using anything like a "too-low" frequency then the level at the 6L6's could be reduced. I figure -3dB for the 0.1uF and the cathode resistor (in parallel with the resistance looking into the cathode of the 6EU7) is ~2900Hz.
   * Sorry for overlooking that before. Call gain 1/2 what I figured for a typical 400-1000Hz test tone; 130mV peak -> ~25v peak at the 6L6 grids.

See C12 (100pF)? If you use a "too high" frequency then the level at the 6L6's could be reduced. You're probably not testing up around 10-20kHz, so don't sweat that part.

See C13, C15, C16, R20, R21, R22? These are a twin-T filter, which places a steep cut at a specific frequency. If you use a frequency  "too close" to that notch frequency, then the level at the 6L6's could be reduced. With an online calculator, I get the notch frequency around 70Hz.

If you think those parts might be bad, or having an impact, hook up a jumper wire between the output of C14 and V5A's grid to bypass the filter.

[BigE]

jumpered from the out of C14 to pin 5 of V5B and no change.  <insert head scratching here>