I've built 3 Fender clones from scratch - Champ. Princeton and Bassman, which taught me a lot about tube design. But when I look at this schematic, I can't make heads or tails! The preamp is cathode biased triode, but so un-Fender, but the power section....the power tube plates go to each side of the OT primary, so that means push-pull? But where is the phase inverter? V2 seems to take separate inputs from the normal and tremelo channels, but then go separately to each power tube. What am I missing? And what is an "8K2" resistor? And the rectifier is full wave? I will dig into Valve Wizard and the Marshall forums, and know I will be learning (unlearning?) a lot of what I thought about tube amplification. Thanks!
Attachments
The second stage's "long tail phase inverter" gives push pull operation for the normal inputs.
The tremelo input to the other side of the phase inverter also results in push pull outputs for those inputs.
The reason is the high cathode resistor on V2. It makes the plate outputs equal and opposite for any inputs.
Using the principle of superposition, we see that both sets of inputs can work independently or at the same time.
http://www.valvewizard.co.uk/acltp.html
Last edited:
OK looks like V2 is the PI..., but getting separate inputs from the normal and tremelo channels is so un-Fender...
Yes, absolutely. The design of the phase inverter forces the two plates of V2 to always be equal and opposite,
even if there is an input to only one of the two grids.
Those are just for proper grid biasing and don't affect the AC operation much. Imagine the 56k cathode resistor greatly increased AC wise.
Then the plate current of the first half of V2 would be forced to flow into the second half of V2. The two plate currents would be equal,
since they would actually be the same current. This means the AC drops across the plate resistors, and hence the two V2 AC plate voltages,
will be equal (but opposite, since the two plate currents flow in opposite directions).
Last edited:
Thanks! I need to revisit phase inversion (valve wizard), to visualize what the signals are doing as they alternate...
It's also important to understand that it is actually inverting signal polarity, not changing phase.
Unfortunate term, a phase inverter doesn't actually change the phase.
Last edited:
Think opamp - one input is inverting, one input is non-inverting - this is the same, except you also have two outputs (the inverse of each other).
Little side note, you may notice some long tail pair phase inverters don't have matched plate resistors. There is one side slightly lower meant to balance the voltage output of each side. Some do..and some don't, and for musical amp it is not a question of right/wrong. Audio you want balanced, and guitar amps well...
That's because the cathode circuit lacks a high enough impedance, usually because of the lack of a negative supply.
Sure ain't a Marshall either, it's originally a Watkins Dominator design..before it became the bluesbreaker. Either way, cool circuit that won't break back or eardums!
Hi Guys
A lot of "economical" designs used the two inputs to the differential splitter as INPUTs from separate preamp stages. Later, almost universally all amps use this same circuit but with the preamp tied to one input and the feedback signal tied to the other.
While Vox, Marshall, Gibson and many others used the diff-splitter inputs independently running the power amp without feedback, Fender was using the concertina splitter. Around the same time, it seems every manufacturer shifted their thinking to the same method, which is what we have had for the past few decades. The modern method takes advantage of the "differentiating" capability of the diff-splitter to compare a signal to the feedback and reduce overall distortion.
Have fun
A lot of "economical" designs used the two inputs to the differential splitter as INPUTs from separate preamp stages. Later, almost universally all amps use this same circuit but with the preamp tied to one input and the feedback signal tied to the other.
While Vox, Marshall, Gibson and many others used the diff-splitter inputs independently running the power amp without feedback, Fender was using the concertina splitter. Around the same time, it seems every manufacturer shifted their thinking to the same method, which is what we have had for the past few decades. The modern method takes advantage of the "differentiating" capability of the diff-splitter to compare a signal to the feedback and reduce overall distortion.
Have fun
Concertina splitter...? Do you mean Cathodyne? Like in the Fender Deluxe ans Princeton? Uses only one triode of an 12AX7?
Hi Guys
Also called a "split-load" phase inverter, for obvious reasons.
The concertina has many advantages over the Schmitt: far better stability; separation of gain and signal splitting; lower output impedance overall. Then some disadvantages: supply needs better filtering; output impedances are not matched; frequency responses of the outputs are not matched; less voltage swing for a given supply voltage.
Personally I prefer the concertina because it is more stable and allows many more options for other features. The asymmetric output is particularly well-suited to guitar amps as this provides for a fatter tone. The imbalance of the standard stacked-Schmitt also allows this.
The Schmitt is no more balanced with a negative rail for the current source resistor than when using the same resistor value with a positive grid reference voltage equal to the negative rail value. Imbalance will be present in any differential stage that does not have balanced input signals. Real life and sims demonstrate this amply: any "balanced" circuit will show asymmetric output levels with any input signal other than a precisely balanced input signal. For guitar amps this is completely unimportant, so a few less things to worry about.
The only thing that needs to be balanced in a push-pull amp for guitar is the output transformer (OT) hum. The two halves of the primary are not identical, so balancing the DC idle currents does not cause hum balance. DC balance is close to the hum balance point but usually off by a few milliamps. There are many ways to hum-balance the output stage whether the stage is fixed-biased or cathode-biased. TUTs show how.
Have fun
Also called a "split-load" phase inverter, for obvious reasons.
The concertina has many advantages over the Schmitt: far better stability; separation of gain and signal splitting; lower output impedance overall. Then some disadvantages: supply needs better filtering; output impedances are not matched; frequency responses of the outputs are not matched; less voltage swing for a given supply voltage.
Personally I prefer the concertina because it is more stable and allows many more options for other features. The asymmetric output is particularly well-suited to guitar amps as this provides for a fatter tone. The imbalance of the standard stacked-Schmitt also allows this.
The Schmitt is no more balanced with a negative rail for the current source resistor than when using the same resistor value with a positive grid reference voltage equal to the negative rail value. Imbalance will be present in any differential stage that does not have balanced input signals. Real life and sims demonstrate this amply: any "balanced" circuit will show asymmetric output levels with any input signal other than a precisely balanced input signal. For guitar amps this is completely unimportant, so a few less things to worry about.
The only thing that needs to be balanced in a push-pull amp for guitar is the output transformer (OT) hum. The two halves of the primary are not identical, so balancing the DC idle currents does not cause hum balance. DC balance is close to the hum balance point but usually off by a few milliamps. There are many ways to hum-balance the output stage whether the stage is fixed-biased or cathode-biased. TUTs show how.
Have fun
I notice also the Marshall 18w preamp tube shares one 100K load resistor. This reduces the gain (effectively putting only 50K on each stage?), but increases headroom? And the EL34s share one 820R cathode resistor. Seems fairly conservative. So the amp should be pretty clean, except there's no feedback... Man, I gotta build this!
I've seen a slightly different 18W schematic, with a "single" and "parallel" input, which sends the guitar into either one or both triodes. With a shared load resistor will, in single mode, the one triode still be sharing the load resistor, or see its full load (meaning higher gain)?
Interesting question. I've used equivalent circuits to calculate the following (assuming fully bypassed cathodes and ignoring the loading from the subsequent stage):
(Case 1, parallel 12AX7 triodes sharing 100k plate load)
Gain = mu x Ra / (Ra + ra/2) = 100 x 100 / (100 + 64/2) = 76
(Case 2, parallel 12AX7 triodes sharing 100k plate load, but no signal into one of the grids)
Gain = mu x Ra||ra / (Ra||ra + ra) = 100 x 39 / (39 + 64) = 38
The gain in Case 2 is low because the active triode is loaded by the ra of the idling triode.
The above calculations do not take into account any change in potential dividers at the input grids. (Not sure if the circuit you are looking at has that.)
(Case 1, parallel 12AX7 triodes sharing 100k plate load)
Gain = mu x Ra / (Ra + ra/2) = 100 x 100 / (100 + 64/2) = 76
(Case 2, parallel 12AX7 triodes sharing 100k plate load, but no signal into one of the grids)
Gain = mu x Ra||ra / (Ra||ra + ra) = 100 x 39 / (39 + 64) = 38
The gain in Case 2 is low because the active triode is loaded by the ra of the idling triode.
The above calculations do not take into account any change in potential dividers at the input grids. (Not sure if the circuit you are looking at has that.)
Last edited:
- Status
- Not open for further replies.