I’ve got this amp back together, so it is time to provide some observations on the circuit on the circuit and other stuff. There have been a few surprises.
I started working on the preamp a few weeks ago. See before and after pictures. I basically shotgunned the circuit board, replacing all caps and resistors. I also cleaned and lubricated the pots. As near as I can tell, this was the first time the preamp had every seen any service. There were many bad resistors, and one dead tube (presumably related to the bad resisters).
During this process, I noticed that this amp is a bit odd, at least for an Ampeg. It is a two channel amp, with the only real difference between the channels being the input impedance. I did note that there are no grid stoppers on the input tube, which is something that I find odd, and am mildly inclined to change. There are two preamp tubes, a 6SL7 and a 6SN7. The 6SN7 is not present in the 835N schematic available on this site, so the obvious question is: why is it there? I was quite surprised to discover it wired as a dc-coupled cathode follower used to drive the tone stack. This is not very Ampeg, at all. In fact, the two channel cathode-driven tone stack configuration (see my schematic here) is almost identical to an early tweed Fender Bassman, such as the 5e6 (different tubes, though). Of course, the tone stack is Bandaxall, so that much remains true to Ampeg. I put the tone stack into Duncan’s Tone Stack Calculator, and it shows a pretty significant trough in the upper bass range, ranging from about 100-300hz, depending on bass knob selection. Kinda odd, considering it’s a bass amp.
In any event, having identified the workings of the preamp, I decided that I needed to go back and look at the phase inverter. I had originally assumed that it was a standard long-tailed PI. But given that cathode followers are typically followed by a gain stage, I expected to find the 6SL7 in the power chassis configured as a gain stage followed by a Cathodyne PI, like in the Bassman. What I found was something more complex.
The phase inverter is pretty similar to that found in the 835N, although the feedback mechanism is different. It’s actually a “closed loop” PI, which you can read about here, albeit with a higher parts count (not quite sure why they did that). There were issues with this stage. It turned out the negative feedback resistor from the output jack had drifted badly, so the cathode of the first half of the triode was dumping current on the speaker/output transformer.
In the closed loop PI the grid of the second triode is fed through a voltage splitter tied to the two plates. The idea is that part of the signal sent to the first power tube is mirrored to the grid of the triode, which is amplified (and inverted), and sent to the other power tube. The linkage back to the grid through the voltage splitter is supposed to provide automatic balancing, as any growth in the second signal will cancel out part of the input signal. As can be seen in the article linked above, the voltage splitter typically utilizes a 510k and a 470k resistor, which this amp did. But the 510k had drifted badly, and so things weren’t working well at all. Just for amusement, I created a spice model of the PI (LTSpice model here), and sure enough, given the correct values the PI provides remarkable balance to the power tubes, excellent headroom, and about 10x in voltage gain, which eliminates the need for a separate gain stage after the cathode follower.
The PI is working well. Here is a reading at the two power tube grids, using my scope.
OK. So I have moved on to (I) verifying that the speaker actually works (it does); cleaning up the cabinet (it is cleaner but still looks like a battle-scarred warrior); and reassembling. Oh, I also put a new handle on it. In a way, I am done, and it looks like this (front and back).
That said, I’m not super happy with where things stand. First, I am not a bass player and don’t have a bass. This amp was supposed to be the bass amp for the music room I am setting up with the family, but a bass is a future purchase. I grabbed a guitar (Parker Fly), and it works fine, but is kind of sterile, and really much brighter than I anticipated. It could be the circuit itself; it could be that the speaker needs to loosen up. It does have a microphonic tube (most likely the 6SN7 that I just kind of selected at random and stuck in there). But I supposed I will rummage through my tube stash and see if swapping out the preamp tubes doesn’t make the amp more lovable. Other suggestions are welcome.
[Just a quick note: I reviewed the schematic that is posted above, and I see at least one thing that I think is incorrect. The voltage divider that connects the two channels is showing as having resistor values 270k and 100k, but my memory is telling me that those resistors have equal value. I just can't remember if they are both 100k or 270k.
I'll have to open the amp up and check.]
Thanks, I'll consider it. Mostly it's easier for me to just report things as I go, but I don't know who will have the patients to read all the way through my rather disjoint posts.
I started working on the preamp a few weeks ago. See before and after pictures. I basically shotgunned the circuit board, replacing all caps and resistors. I also cleaned and lubricated the pots. As near as I can tell, this was the first time the preamp had every seen any service. There were many bad resistors, and one dead tube (presumably related to the bad resisters).
During this process, I noticed that this amp is a bit odd, at least for an Ampeg. It is a two channel amp, with the only real difference between the channels being the input impedance. I did note that there are no grid stoppers on the input tube, which is something that I find odd, and am mildly inclined to change. There are two preamp tubes, a 6SL7 and a 6SN7. The 6SN7 is not present in the 835N schematic available on this site, so the obvious question is: why is it there? I was quite surprised to discover it wired as a dc-coupled cathode follower used to drive the tone stack. This is not very Ampeg, at all. In fact, the two channel cathode-driven tone stack configuration (see my schematic here) is almost identical to an early tweed Fender Bassman, such as the 5e6 (different tubes, though). Of course, the tone stack is Bandaxall, so that much remains true to Ampeg. I put the tone stack into Duncan’s Tone Stack Calculator, and it shows a pretty significant trough in the upper bass range, ranging from about 100-300hz, depending on bass knob selection. Kinda odd, considering it’s a bass amp.
In any event, having identified the workings of the preamp, I decided that I needed to go back and look at the phase inverter. I had originally assumed that it was a standard long-tailed PI. But given that cathode followers are typically followed by a gain stage, I expected to find the 6SL7 in the power chassis configured as a gain stage followed by a Cathodyne PI, like in the Bassman. What I found was something more complex.
The phase inverter is pretty similar to that found in the 835N, although the feedback mechanism is different. It’s actually a “closed loop” PI, which you can read about here, albeit with a higher parts count (not quite sure why they did that). There were issues with this stage. It turned out the negative feedback resistor from the output jack had drifted badly, so the cathode of the first half of the triode was dumping current on the speaker/output transformer.
In the closed loop PI the grid of the second triode is fed through a voltage splitter tied to the two plates. The idea is that part of the signal sent to the first power tube is mirrored to the grid of the triode, which is amplified (and inverted), and sent to the other power tube. The linkage back to the grid through the voltage splitter is supposed to provide automatic balancing, as any growth in the second signal will cancel out part of the input signal. As can be seen in the article linked above, the voltage splitter typically utilizes a 510k and a 470k resistor, which this amp did. But the 510k had drifted badly, and so things weren’t working well at all. Just for amusement, I created a spice model of the PI (LTSpice model here), and sure enough, given the correct values the PI provides remarkable balance to the power tubes, excellent headroom, and about 10x in voltage gain, which eliminates the need for a separate gain stage after the cathode follower.
The PI is working well. Here is a reading at the two power tube grids, using my scope.
OK. So I have moved on to (I) verifying that the speaker actually works (it does); cleaning up the cabinet (it is cleaner but still looks like a battle-scarred warrior); and reassembling. Oh, I also put a new handle on it. In a way, I am done, and it looks like this (front and back).
That said, I’m not super happy with where things stand. First, I am not a bass player and don’t have a bass. This amp was supposed to be the bass amp for the music room I am setting up with the family, but a bass is a future purchase. I grabbed a guitar (Parker Fly), and it works fine, but is kind of sterile, and really much brighter than I anticipated. It could be the circuit itself; it could be that the speaker needs to loosen up. It does have a microphonic tube (most likely the 6SN7 that I just kind of selected at random and stuck in there). But I supposed I will rummage through my tube stash and see if swapping out the preamp tubes doesn’t make the amp more lovable. Other suggestions are welcome.
[Just a quick note: I reviewed the schematic that is posted above, and I see at least one thing that I think is incorrect. The voltage divider that connects the two channels is showing as having resistor values 270k and 100k, but my memory is telling me that those resistors have equal value. I just can't remember if they are both 100k or 270k.
I'll have to open the amp up and check.](12-15-2017, 09:28 PM)Hangman Wrote: If you're interested in writing a blog post about your restoration process for this site, let me know.
Thanks, I'll consider it. Mostly it's easier for me to just report things as I go, but I don't know who will have the patients to read all the way through my rather disjoint posts.