Search the Forum
|
(Advanced Search)
|
|
Single-knob Power Scaling
|
Posts: 522
Threads: 58
Joined: Aug 2018
Location: CANADA
Country:
Hi Guys
The London Power Power Scaling kits come with two controls for the front panel: Power Scale and Drive Compensation. These two controls allow four ways to use the amp, one of which is Power Scaling where the tone is retained with loudness changes. The other modes are simply versatilities made possible.
Many players still want a 1-knob Power Scale solution, and there are situations where this is possible with the kits. Power Scale is a single-section pot where Drive Comp is supplied as a dual-pot and this is what you use for the single-knob solutions below.
If the amp has a master-volume preamp, such as the original "800" does, then the dual pot can replace the stock MV and the new Power Scale control. The Power Scaling goal is achieved as drive is in step with Power Scale and tone is maintained.
A Vox amp may come to mind here as those with an MV use a single-section pot. However, the MV is the cross-line type, the worst configuration possible, which would be greatly improved in function were it replaced by a 2-section pot wired in the more usual manner. This negates using the supplied DriveComp pot as a 1-knob Power Scale control.
Amps that do not use a Schmitt splitter and instead have a gain stage followed by a concertina splitter are candidates for a 1-knob Power Scale solution. In this case, one section of the pot is placed between the gain stage and the concertina, with the other section used as the Power Scale control. It does not matter if feedback is wrapped around the PA.
If the MV is a post-PI type that is typically a dual-pot itself, then the above solution does not work. One would need a 3-section pot, which are available, but it is easier to find 4-sectiuon pots that are fairly common now for home theatre applications.
In fixed-biased amps, you still need bias controls per usual, mounted on the tube plane of the chassis and set ONLY at full power. These are not performance pots; rather, set and forget with each tube change or new PS installation.
Posts: 176
Threads: 27
Joined: Oct 2018
Location: United States
Bio: I'm 34 years old and work on MI amps and do some cattle ranching.
Country:
04-26-2023, 11:32 PM
(This post was last modified: 04-27-2023, 02:09 PM by makinrose.)
Great post! Question: how would recommend setting a drive compensation/limit on an amp with see saw inverter like those found on early tweed and pre-tweed amps? I've experimented with few setups but haven't had the success I've had with Schmitt inverters and a dual pot. It seems like drive compensation ends up changing the character of the second drive line's that is being fed from the first.....It may be that the design is too archaic, but I do like the sound of the see saw inverter for certain styles of music and would like to find good way to apply power scaling to those designs without resorting to a cross line master volume.
Posts: 11
Threads: 1
Joined: Dec 2020
Location: Canada
Country:
Hi makinrose,
I don't have a tonne of experience with seesaws but I was able to get adequate results in a custom cathode-biased amp where the power tube grid leaks were replaced with a dual-pot. Power tube grid-stoppers were of course added.
That amp also had a 'balance' pot as part of the divider between the output of the first seesaw triode and the second which the owner often used in conjunction with the drive comp knob.
That said, we weren't trying to replicate any specific tones, just tried to be a little bit different. Our "adequate" solution might be different to your goal.
Posts: 522
Threads: 58
Joined: Aug 2018
Location: CANADA
Country:
Hi Guys
There are a number of similar-seeming 2-triode splitters that might all be loosely referred to as "see-saws".
In TUT Fig..6-11, the first figure shows two independent gain stages where the first stage drives one output tube and also a voltage divider. The voltage divider feeds the second stage to drive the other side of a push-pull output stage. There, the voltage divider is specified to match the free-running gain of the second stage - pretty chancy.
In the context of the single-knob Power Scale solution, a pot sect can be placed at the output of the first stage which would feed both its side of the power stage and the voltage divider. The latter would have to be comprised of high-value resistors to not later the sweep of the pot too much.
The right-hand figure uses a free-running stage to drive one output tube and to drive a feedback-controlled inverter stage. The inverter has a gain of -1 and drives the second half of the output stage.related This circuit is intended to provide accurate signal amplitudes to the output tubes. Again, half the PS pot can be used at the output of the first stage, driving both its related power tube grid and the input to the inverter stage. To not alter the pot sweep too much, the gain-set resistors for the inverter have to be quite high.
Any variations of splitter design where the stages are interconnected at the cathode will require a 2-section pot for the drive compensation function unless a single-section is used ahead of the entire circuit.
Have fun
Posts: 176
Threads: 27
Joined: Oct 2018
Location: United States
Bio: I'm 34 years old and work on MI amps and do some cattle ranching.
Country:
(04-27-2023, 01:15 PM)K O'Connor Wrote: Hi Guys
There are a number of similar-seeming 2-triode splitters that might all be loosely referred to as "see-saws".
In TUT Fig..6-11, the first figure shows two independent gain stages where the first stage drives one output tube and also a voltage divider. The voltage divider feeds the second stage to drive the other side of a push-pull output stage. There, the voltage divider is specified to match the free-running gain of the second stage - pretty chancy.
In the context of the single-knob Power Scale solution, a pot sect can be placed at the output of the first stage which would feed both its side of the power stage and the voltage divider. The latter would have to be comprised of high-value resistors to not later the sweep of the pot too much.
The right-hand figure uses a free-running stage to drive one output tube and to drive a feedback-controlled inverter stage. The inverter has a gain of -1 and drives the second half of the output stage.related This circuit is intended to provide accurate signal amplitudes to the output tubes. Again, half the PS pot can be used at the output of the first stage, driving both its related power tube grid and the input to the inverter stage. To not alter the pot sweep too much, the gain-set resistors for the inverter have to be quite high.
Any variations of splitter design where the stages are interconnected at the cathode will require a 2-section pot for the drive compensation function unless a single-section is used ahead of the entire circuit.
Have fun I was referring to the type of splitter on the left hand side of TUT Figure 6-11. Most usually in Fender, Valco and Gibson amps of that era I'll see R1 and R2 act as the grid leak for output tube one and a voltage divider for the second drive line.
I have tried the single pot approach you suggested but wasn't happy with the sound...I may need to further play with the values I used though.
If I were to use a dual pot rather than a single pot would I use a R1 and R2 (from Figure 6-11) as a voltage divider but not a grid leak for output tube and then replace the two grid leaks with pots? Maybe a bit like figure 6-38 in TUT 4 but with a dual pot for the 100K grid leaks?
Posts: 522
Threads: 58
Joined: Aug 2018
Location: CANADA
Country:
Hi Guys
Re: TUT Fig.6-11 right side:
R1+R2 can only be used as a grid-leak for the output tube if the output stage is cathode-biased. My comments above are looking at the splitter in isolation. With fixed-bias, there would be a coupling cap into each power tube grid.
In the context of cathode-bias, the description above applies.
If you want the net grid-circuit resistance to be 220k, start with a 250kA pot (very common value) and make R1,2 around 2M5. This would likely be a 2M2 for R1 and then something between 22k and 240k depending on the gain of the second triode. You might want to use a 1M trimmer for R2 just to dial in the closest output balance.
For either bias method, this addendum applies
POP shows other splitter variations that have cathode-coupling between the triodes (Fig.4-20), forcing one to use two pot sections at the output or one pot section at the input for the single-pot PS solution.
Regarding old amps: The resistors used might be E24 20% cracked carbon if very old, carbon film if not so ancient. In any case, 20% resistor values around the tubes do not control things too tightly. In your new builds, you likely have 5% values, and even using 1% will not make the tubes work exactly as you might hope. Since the left-side circuit has free-running stages, balance of output is at the mercy of the tube gain spread (do not waste money on "matched sections") and on the circuit tolerances. Presumably, using such a splitter today means you are looking for some of that imbalance to add fatness to the tone?
When you look at the right-hand circuit and those from POP, you would expect excellent output matching, but that's not what you get. Despite throwing away all of the tube's voltage gain to have unity-inverting, the gain will be a bit lower because the amount of gain sacrificed is still relatively small compared to what ideal feedback control demands. This might be okay.
The input resistor becomes a load to the free-running input stage for the PA, along with the grid-leak circuit of the first power tube. The free-running stage may not exhibit its maximum or desired gain, but at least it is running free. The see-saw stage is supposed to provide exactly -1 gain and you might have to tweek one of the feedback Rs to get there.
In any case, because the signal goes through an extra tube to get to one half of the output stage, that signal will be a bit more distorted, or "harmonically rich" compared to the drive into the other side. This will produce a warmer tone without there being noticeable imbalance between the stages.
*As far as I know, the triode sections cannot be tested until they are in a vacuum, so once they are within the full glass envelope the deed is done.
Posts: 176
Threads: 27
Joined: Oct 2018
Location: United States
Bio: I'm 34 years old and work on MI amps and do some cattle ranching.
Country:
04-27-2023, 08:40 PM
(This post was last modified: 04-28-2023, 12:11 PM by makinrose.)
(04-27-2023, 04:48 PM)K O'Connor Wrote: Hi Guys
Re: TUT Fig.6-11 right side:
R1+R2 can only be used as a grid-leak for the output tube if the output stage is cathode-biased. My comments above are looking at the splitter in isolation. With fixed-bias, there would be a coupling cap into each power tube grid.
In the context of cathode-bias, the description above applies.
If you want the net grid-circuit resistance to be 220k, start with a 250kA pot (very common value) and make R1,2 around 2M5. This would likely be a 2M2 for R1 and then something between 22k and 240k depending on the gain of the second triode. You might want to use a 1M trimmer for R2 just to dial in the closest output balance.
For either bias method, this addendum applies
POP shows other splitter variations that have cathode-coupling between the triodes (Fig.4-20), forcing one to use two pot sections at the output or one pot section at the input for the single-pot PS solution.
Regarding old amps: The resistors used might be E24 20% cracked carbon if very old, carbon film if not so ancient. In any case, 20% resistor values around the tubes do not control things too tightly. In your new builds, you likely have 5% values, and even using 1% will not make the tubes work exactly as you might hope. Since the left-side circuit has free-running stages, balance of output is at the mercy of the tube gain spread (do not waste money on "matched sections") and on the circuit tolerances. Presumably, using such a splitter today means you are looking for some of that imbalance to add fatness to the tone?
When you look at the right-hand circuit and those from POP, you would expect excellent output matching, but that's not what you get. Despite throwing away all of the tube's voltage gain to have unity-inverting, the gain will be a bit lower because the amount of gain sacrificed is still relatively small compared to what ideal feedback control demands. This might be okay.
The input resistor becomes a load to the free-running input stage for the PA, along with the grid-leak circuit of the first power tube. The free-running stage may not exhibit its maximum or desired gain, but at least it is running free. The see-saw stage is supposed to provide exactly -1 gain and you might have to tweek one of the feedback Rs to get there.
In any case, because the signal goes through an extra tube to get to one half of the output stage, that signal will be a bit more distorted, or "harmonically rich" compared to the drive into the other side. This will produce a warmer tone without there being noticeable imbalance between the stages.
*As far as I know, the triode sections cannot be tested until they are in a vacuum, so once they are within the full glass envelope the deed is done.
Thanks for all the help. That will definitely help me design the splitter. The see-saw/paraphase splitter is really good for harmonica amps and for guitar players wanting an early fat blues tone. It certainly isn't a very modern sound but tends to be fat and gritty. I regularly build these as non-master volume amps for customers.
The drawback to the amps that is that they rely on whole amp overdrive. The clean (cleanish) sounds are interesting but lots of people like to push them to overdrive which only happens at high volumes. I'm wanting to make it possible to get those tone without being so loud.
I have also thought it is probably possible to get a similar sound by adding gain stages to the typically simple pre-amps those types of amps have and use a more usual splitter. Traditionally these early tweed circuits use octal pre-amp tubes which are rather noisy at least with the circuit values I usually see. That noise may be problematic if I am adding gain stages. Miniature tubes may be the way to go. I've got a lot of experimenting to with to see what works best for the goal.
Posts: 522
Threads: 58
Joined: Aug 2018
Location: CANADA
Country:
Hi Guys
An alternative to the see-saw splitter is the concertina. Jut wire the splitter as self-bias and there is still the free-running stage ahead of it. The concertina output is slightly asymmetric and warmer than with a Schmitt, so this is as "conventional" as you might want to update the splitter
Most harmonica amps work best with lower-mu tubes - the 12AY7 is a common sub into Fender amps, reducing two stages from mu=100 to mu=50. 12AT7 is in between at mu=70. As TUTs state, low-mu tubes sound mellower than high-mu tubes. A harmonica has pretty high and steady output, akin to a signal generator plugged into the amp, so the milder tone of these tubes works to remove some of the harshness of the instrument.
You can also use an asymmetric output stage with dissimilar tubes. If you find the concertina has less fatness than the pre-see-saw 2-stage style, this output tube mix might compensate for it.
Posts: 176
Threads: 27
Joined: Oct 2018
Location: United States
Bio: I'm 34 years old and work on MI amps and do some cattle ranching.
Country:
04-28-2023, 08:53 PM
(This post was last modified: 04-28-2023, 08:59 PM by makinrose.)
Moving over to concertina is essentially what Fender did with it's "E" series Tweeds. All the previous iterations were see-saw inverters and most used the standard "Tweed Deluxe" type circuit that other manufacturers also used during that period.
To my ears the concertina amps sound different. Not better or worse but more refined and the breakup has a different voice. That said I have built those with the typical E series values so I might not be comparing apples to apples. Maybe using the higher plate resistors (250K v.s 100K) with higher value cathode resistors and a grid-leak input stage will bring it closer to sound I am wanting and ability to PS more easily? The mix tubes in the output stage would be easy to try as well. Thanks for the suggestion!
Posts: 7
Threads: 2
Joined: Jan 2024
Location: The beach...
Bio: Playing guitar badly since 1986 but a little bit better since 1989...
Self-exiled to Japan following a study abroad in 1998...
Proud owner of a then-new '93 G&L Legacy, Vox Virage 2, lots of rare (unloved) pedals and bunch of VHT Specials sometimes for money (ALWAYS for love.)
Country:
I think single-knob is the way to go for my VHT Special 12/20RT because of its effects loop RETURN control that effectively acts like MV before the phase splitter that looks like a concertina/cathodyne circuit. Are there better alternatives for the PI (like this kit)?
https://londonpower.com/electronics/spli...er-volume/
Posts: 522
Threads: 58
Joined: Aug 2018
Location: CANADA
Country:
Hi Guys
Looeselectron: The VHT1220 has trem which needs to be controlled. The Watts control is a dual with one section controlling power and the other section keeping the trem modulation of the grids in proportion. To have a one-knob Power Scale solution here requires a 3-section 1MA pot.
Your SV2 has been modified to use a 1M for Power Scale to accommodate use of the stock pot.
The MV is 250kA just like the usual Power Scale and DriveComp pots are in the SV kits. If you could find a triple or 4-section 250kA that might work for you - with a caveat.
The VHT schematic shows the 1MA in series with the ground-referenced 250kB Depth control, forming a voltage divider. The 1MA is wired as a rheostat and can look like any value between zero and 1M. At the extreme end, there is a voltage division of 5:1, where the modulation signal amplitude comes out at 1/5 of what went in. With 1MA replaced by 250kA, the minimum signal is much higher at 1/2. This would cause the trem depth to increase as you dial down power.. This suggests that changing the Depth control itself to 62k5 would keep everything proportioned as it should be. Of course, you won't find such a value of pot but you can use a 100kA paralleled across pot-X to pot-0 by 167k.
The loading on the tremolo circuit ranges from 1M25 to 250k as stock. In the modified form with the triple 250A for Power Scale, the loading would range from 500k to 250k. This likely would not effect things too noticeably.
Note that the SP-MV referred to is only for Schmitt splitters. It is not really a preferred MV type and is meant to be the proper form of such MVs if one wishes to use it.
|
|
Come in where it's warm!
|
A warm welcome to tube amp modding fans and those interested in hi-fi audio! Readers of Kevin O'Connor's The Ultimate Tone (TUT) book series form a part of our population. Kevin O'Connor is the creator of the popular Power Scaling methodology for amplifiers. |
Please remember these three principles: respect, sharing, community. |
Not familiar with The Ultimate Tone book series? See discussion topics, or click here to visit London Power/Power Press Publishing. |
|