Tweet
While I've got my mate's RI LTD on the bench (having just fixed it with new output tubes and a new (mojo) OT), here I am sitting twiddling on it with my strat thinking "what it is I don't like about the sound?" and it occurs to me that it has a rather ice-picky high end when the presence is dimed. I spoke to him about the ice-pickiness, and he also finds it annoying and is happy for me to experiment a bit and see it it can be tweaked. (Yay!). The PC-board technology in the RI makes changing out any resistor values a real PITA, but apart from gutting the whole amp and doing a 5F6A eyelet board, which I don't really want to be bothered with since its not my amp, I want to think about this carefully about what mods to tweak before diving in.
What I have tried is a 'speed-up' cap tacked in parallel with the 27k NFB series resistor in the RI. The RoT Merlins suggests would mean 74pF is about right (I went with 100pF - and the presence control does sound better for this)
However, to pontificate a bit more, I notice that there are a few differences about the RI circuit compared to a vintage 5F6A. To compare the two (starting from front to back):
The RI http://www.el34world.com/charts/Sche...N_59_REV_E.pdf * has the same input stage as the 5F6A http://www.el34world.com/charts/Sche...SSMAN_5F6A.pdf (more or less) and the same DC-coupled pair driving the tone stack.
(* I couldn't find a schematic for the LTD, but I'm assuming its much the same as the '59 Bassman RI, albeit that the LTD has adjustable bias and a 5AR4 tube rectifier)
The differences between the 5F6A and the RI begin at the tone stack. The TS in RI has a 100k 'slope' resistor (compared with 56k in the 5F6A) and a .1uF cap going to the bass pot input (compared with .022uF in the 5F6A), but the pot values, and the treble and mid caps are the same as the 5F6A. I guess the 56k slope resistor in the 5F6A shifts the mid-scoop point up a couple of hundred Hz and decreases the relative amount of mid scoop available (and is potentially a bit lossier than the 100k slope resistor - not that the latter point should make much difference when being driven by a CF stage). I also guess that the .1uF bass pot cap in the RI interacts to either shunt more bass to ground when the pot is cut, or allows more bass to pass to the next stage when the pot is dimed. The tone controls in the RI circuit give a reasonable range of control, but then again so does the 5F6A tone stack. Horses for courses. A PC-board PITA if I was to change the tone stack to a 5F6A version.
The LTP stages between the RI and the 5F6A are also different, and the RI owes more to the blonde bassman (6G6B) http://www.el34world.com/charts/Sche...SSMAN_6G6B.pdf circuit. The LTP tail resistor is 6k8 in the RI (same as the 6G6B) whereas its 10k in the 5F6A (giving the RI slightly more gain I guess). This is also probably because the RI borrows the presence control from the blonde bassman (but in the RI it is really trebly-peaky when dimed - maybe to do with the OT and speakers in the RI being '5F6A spec'). I note that the LTP plate-to-plate shunt in the 6G6B is 100pF, whereas the RI has retained the 47pF that the 5F6A has. I guess this wouldn't help with the peakyness in the RI, and it may be a simple matter of tacking an extra 100-250pF across the 47pF (that is in the RI) in order to stabilise this somewhat. So that's something relatively easy to try (since an extra 47pF - 100pF can be tacked between Pins 1 and 6 of the LTP socket). (Any suggestions/comments about that?)
Next; the NFB loop in the RI employs a 27k/4k7 voltage divider (similar to the 5F6A), presumably because its riding off the 2000:1 impedance ratio of the RI/5F6A OT. This is probably on a par with the amount of NFB in a 6G6B (but could be why the plate-to-plate shunt in the RI needs upping maybe?). I was wondering whether there could also be a simple opportunity here for 100-200pF shunt caps from the grid-to-ground (as Merlin outlines in Chapter 9 (Fig 9.11) in his first book (to 'slug' the dominant NFB pole as he puts it). It would be a hell of a lot easier to simply put these caps on the output tube socket pins between the grid and the cathode (but the cathode is sitting on top of the 1R bias resistor - does that matter in this regard?).
The output stages of the RI and the 5F6A are virtually identical, but the RI employs 47R grid resistors, which are mounted on the main board (- miles away from the 6L6 grid pins, and exactly what they expect 47R to achieve as grid stoppers for the 6L6s is beyond me). Seems to me that putting 1k5-4k7 grid stoppers at Pin 5 might be an improvement.
Also the 470R screen resistors in the RI are mounted on the main board (also miles away from the screen socket pins). I wonder if it would be better (in terms of making the noise floor of the amp a bit quieter - in the typically high-interference gigging environment - considering stray coupling, oscillations'n'all that) by adding more (~47-100R) g2 resistance in series with each 470R, but mounting these additional resistors on the output tube socket pins? Would (extra) 47R g2 resistors at Pin 4 help with stray RF coupling/stability?
In terms of the power rail, there isn't much difference between the RI and the 5F6A electrically I guess (other than series filter caps with equalising resistors in the RI), except that there is a lot more filter/decoupling for the pre-amp in the RI (possibly to do compensate for something wishy-washy in the grounding design of the PC board?). Would this may make the RI input a bit stiffer/bitier? Again such a pain to have to take the board out to change this.
Sorry about the long-winded post. What do others think?
What I have tried is a 'speed-up' cap tacked in parallel with the 27k NFB series resistor in the RI. The RoT Merlins suggests would mean 74pF is about right (I went with 100pF - and the presence control does sound better for this)
However, to pontificate a bit more, I notice that there are a few differences about the RI circuit compared to a vintage 5F6A. To compare the two (starting from front to back):
The RI http://www.el34world.com/charts/Sche...N_59_REV_E.pdf * has the same input stage as the 5F6A http://www.el34world.com/charts/Sche...SSMAN_5F6A.pdf (more or less) and the same DC-coupled pair driving the tone stack.
(* I couldn't find a schematic for the LTD, but I'm assuming its much the same as the '59 Bassman RI, albeit that the LTD has adjustable bias and a 5AR4 tube rectifier)
The differences between the 5F6A and the RI begin at the tone stack. The TS in RI has a 100k 'slope' resistor (compared with 56k in the 5F6A) and a .1uF cap going to the bass pot input (compared with .022uF in the 5F6A), but the pot values, and the treble and mid caps are the same as the 5F6A. I guess the 56k slope resistor in the 5F6A shifts the mid-scoop point up a couple of hundred Hz and decreases the relative amount of mid scoop available (and is potentially a bit lossier than the 100k slope resistor - not that the latter point should make much difference when being driven by a CF stage). I also guess that the .1uF bass pot cap in the RI interacts to either shunt more bass to ground when the pot is cut, or allows more bass to pass to the next stage when the pot is dimed. The tone controls in the RI circuit give a reasonable range of control, but then again so does the 5F6A tone stack. Horses for courses. A PC-board PITA if I was to change the tone stack to a 5F6A version.
The LTP stages between the RI and the 5F6A are also different, and the RI owes more to the blonde bassman (6G6B) http://www.el34world.com/charts/Sche...SSMAN_6G6B.pdf circuit. The LTP tail resistor is 6k8 in the RI (same as the 6G6B) whereas its 10k in the 5F6A (giving the RI slightly more gain I guess). This is also probably because the RI borrows the presence control from the blonde bassman (but in the RI it is really trebly-peaky when dimed - maybe to do with the OT and speakers in the RI being '5F6A spec'). I note that the LTP plate-to-plate shunt in the 6G6B is 100pF, whereas the RI has retained the 47pF that the 5F6A has. I guess this wouldn't help with the peakyness in the RI, and it may be a simple matter of tacking an extra 100-250pF across the 47pF (that is in the RI) in order to stabilise this somewhat. So that's something relatively easy to try (since an extra 47pF - 100pF can be tacked between Pins 1 and 6 of the LTP socket). (Any suggestions/comments about that?)
Next; the NFB loop in the RI employs a 27k/4k7 voltage divider (similar to the 5F6A), presumably because its riding off the 2000:1 impedance ratio of the RI/5F6A OT. This is probably on a par with the amount of NFB in a 6G6B (but could be why the plate-to-plate shunt in the RI needs upping maybe?). I was wondering whether there could also be a simple opportunity here for 100-200pF shunt caps from the grid-to-ground (as Merlin outlines in Chapter 9 (Fig 9.11) in his first book (to 'slug' the dominant NFB pole as he puts it). It would be a hell of a lot easier to simply put these caps on the output tube socket pins between the grid and the cathode (but the cathode is sitting on top of the 1R bias resistor - does that matter in this regard?).
The output stages of the RI and the 5F6A are virtually identical, but the RI employs 47R grid resistors, which are mounted on the main board (- miles away from the 6L6 grid pins, and exactly what they expect 47R to achieve as grid stoppers for the 6L6s is beyond me). Seems to me that putting 1k5-4k7 grid stoppers at Pin 5 might be an improvement.
Also the 470R screen resistors in the RI are mounted on the main board (also miles away from the screen socket pins). I wonder if it would be better (in terms of making the noise floor of the amp a bit quieter - in the typically high-interference gigging environment - considering stray coupling, oscillations'n'all that) by adding more (~47-100R) g2 resistance in series with each 470R, but mounting these additional resistors on the output tube socket pins? Would (extra) 47R g2 resistors at Pin 4 help with stray RF coupling/stability?
In terms of the power rail, there isn't much difference between the RI and the 5F6A electrically I guess (other than series filter caps with equalising resistors in the RI), except that there is a lot more filter/decoupling for the pre-amp in the RI (possibly to do compensate for something wishy-washy in the grounding design of the PC board?). Would this may make the RI input a bit stiffer/bitier? Again such a pain to have to take the board out to change this.
Sorry about the long-winded post. What do others think?
Comment