Yeah - this thing is rolling all original everything, so it can certainly use new caps.
Very well could be... the original grounding scheme is less than spectacular IMO.
Enjoy the benefits of a premium membership!
Upgrade Account
-
TalkBass has been independent since 1998. Add your voice.
Create a free account to reply to discussions, view embedded media, and browse with fewer display ads.
Join free Log in
Want zero display ads or expanded classifieds tools? Compare plans.
Ampeg B-18 issue/debug
- Thread starter aborgman
- Start date
Yeah - this thing is rolling all original everything, so it can certainly use new caps.
Very well could be... the original grounding scheme is less than spectacular IMO.
Here is a bad copy of the original for your schematic. There are a few variations. I was referring to a different version (the second schematic) when I made my comment above.
"]
[/URL]
"]
[/URL]
Mine is more like half way in between those...
It has the preamp section more like the first (tonestack between halves of 6SL7), except it has 1M treble pots not 4M.
The power section is like the second with a tube rectifier, etc.
I got it now. Man, they were always tinkering with those designs. So many variations are out there.
Yeah, it makes it a pain trying to follow everything.
I'm going to sit my butt down this afternoon and draw up an accurate schematic.
Very nice. Thanks for the copy.
Looks like the 22K feedback resistor from the OPT to the PI is missing.
Looks like the 22K feedback resistor from the OPT to the PI is missing.
So I'm looking at the ground scheme, and here is how it is currently laid out:
30uF/600V cap - grounded to one lug of can cap
Power tube cathodes and hum pot ground - grounded to another lug of can cap.
Standby switch - grounded to third lug of can cap
Eyelet board ground is connected to the same can cap lug as 30uF cap.
Ground traverses one edge of eyelet board to other end of amp - here it connects to OT output side ground and connects to the chassis through channel 2 input jack.
Sound mostly normal?
30uF/600V cap - grounded to one lug of can cap
Power tube cathodes and hum pot ground - grounded to another lug of can cap.
Standby switch - grounded to third lug of can cap
Eyelet board ground is connected to the same can cap lug as 30uF cap.
Ground traverses one edge of eyelet board to other end of amp - here it connects to OT output side ground and connects to the chassis through channel 2 input jack.
Sound mostly normal?
Yes that ground scheme sounds about right.
Important to note that the only ground point is at the input jack. The cap can is isolated from the chassis.
I've tried a number of different ground schemes on these amps and nothing seemed to work any better. These amps are normally dead quiet. I leave them as originally laid out with the following exceptions.
Rather than use a disjoint set of wires soldered along the eyelet board, I use a continuous piece of slightly heavier gauge wire, bent like a "T" at each eyelet. Same for the ground bus. It makes the buses more robust.
On the B15's I connect the OPT secondary ground and the speaker return together at an eyelet and run a wire from there to the point where the phase inverter is grounded to the bus. This is necessary because there is a feedback resistor from the OPT secondary. The ground point for the B18 is slightly different.
I remove the standby switch from the center tap of the HV secondary. The PT secondary HV center tap goes directly to the (-) of the first power supply cap. I install a bleed resistor on the first cap. The standby switch is installed after the first power supply cap. I also boost the value of the first cap. In order to avoid flashover on the tube rectifier, I increase the peak inverse capacity by installing a diode on each leg of the HV secondary. The diodes are mounted on unused tabs of the tube socket and then go to pins 4 and 6, the HV taps connect to the other end of diodes. It is still tube rectified, the diodes serve to increase the PIV and protect the tube.
Important to note that the only ground point is at the input jack. The cap can is isolated from the chassis.
I've tried a number of different ground schemes on these amps and nothing seemed to work any better. These amps are normally dead quiet. I leave them as originally laid out with the following exceptions.
Rather than use a disjoint set of wires soldered along the eyelet board, I use a continuous piece of slightly heavier gauge wire, bent like a "T" at each eyelet. Same for the ground bus. It makes the buses more robust.
On the B15's I connect the OPT secondary ground and the speaker return together at an eyelet and run a wire from there to the point where the phase inverter is grounded to the bus. This is necessary because there is a feedback resistor from the OPT secondary. The ground point for the B18 is slightly different.
I remove the standby switch from the center tap of the HV secondary. The PT secondary HV center tap goes directly to the (-) of the first power supply cap. I install a bleed resistor on the first cap. The standby switch is installed after the first power supply cap. I also boost the value of the first cap. In order to avoid flashover on the tube rectifier, I increase the peak inverse capacity by installing a diode on each leg of the HV secondary. The diodes are mounted on unused tabs of the tube socket and then go to pins 4 and 6, the HV taps connect to the other end of diodes. It is still tube rectified, the diodes serve to increase the PIV and protect the tube.
Sounds like what I was planning to do.
Where do you usually attach ground for the ext. amp jack?
The external speaker jack is isolated with washers from the chassis, otherwise there is a ground loop. The main speaker out is also isolated since it is a cable. If a jack is installed, it should be isolated as well.
The ext amp jack is not isolated so the ground is connected to the chassis at that point. The signal going to that jack is taken from the PI grid, I use a shielded wire that is grounded at one end at the ext amp jack.
It's cheating a little bit to do this but when you are not using the jack, the ground doesn't come into play. It seems to work fine like this when the jack is being used. There aren't any currents in the chassis between the jack and the main ground that could add noise. You could isolate the jack and run the ground to a pre-amp ground point to keep the pathway as clean as possible.
If you were using switchcraft jacks, like the original, it would mean enlarging the jack mounting hole a bit to accept the isolating washers. Some people frown on modifying a vintage chassis.
The ext amp jack is not isolated so the ground is connected to the chassis at that point. The signal going to that jack is taken from the PI grid, I use a shielded wire that is grounded at one end at the ext amp jack.
It's cheating a little bit to do this but when you are not using the jack, the ground doesn't come into play. It seems to work fine like this when the jack is being used. There aren't any currents in the chassis between the jack and the main ground that could add noise. You could isolate the jack and run the ground to a pre-amp ground point to keep the pathway as clean as possible.
If you were using switchcraft jacks, like the original, it would mean enlarging the jack mounting hole a bit to accept the isolating washers. Some people frown on modifying a vintage chassis.
I may have found the culprit.
After pulling it apart to get ready to replace all the electrical components I found the following:
A bent over resistor pin making contact with another eyelet... that it definitely isn't supposed to.
This little piece of metal is basically a jumper around the 270k load resistor on the pentode gain stage of the 7199.
A jumper right around R14.
After pulling it apart to get ready to replace all the electrical components I found the following:
A bent over resistor pin making contact with another eyelet... that it definitely isn't supposed to.
This little piece of metal is basically a jumper around the 270k load resistor on the pentode gain stage of the 7199.
A jumper right around R14.
okcrum
in your chest
You're definitely on to something. That would completely swamp the negative feedback from the output. Good find!
This is the plate side, not the cathode side - so it shouldn't have a huge effect on the feedback loop.
My guess is that because the pin was just barely touching, and is well and truly oxidized - it was acting in a diode like fashion, and that is why it only effected one side of the waveform.
Good news! Sometimes all it takes is one errant strand to cause havoc.
Has it resolved the problem?
Has it resolved the problem?
Who knows... I was detaching the whole eyelet board in preparation to replace all the electrical components, so I'm not going to bother re-wiring everything until after that. It's certainly pretty suspicious though.
Since you are replacing all the components, have you considered replacing the eyelet board as well? You can buy a piece of G10 circuit board and use eyelets or even better turrets. Mark at Link Removed has details on his web site as to how he did this with a B-15.
The reason why I mention this is that these old boards can become conductive between the eyelets as they age. The old Fender boards are more prone to this because of the material that they were constructed out of. This allows you to leave the original board intact in case you want to reverse the process.
Nah, I was just going to re-use the original eyelet board.
I'll bake it in a reflow oven just to drive out any moisture before re-using it.
okcrum
in your chest
Current flow through the tube would rise, and the tube is then a cathode follower when the pin is shorting the plate resistor. The signal across the cathode resistor is in phase with the stage input, and would be way higher in the fault case than the normal feedback from the OT. Both are good candidates for an oscillation condition.
Good call on the point contact behavior.
