Real TrainWreck mojo!

[Alan0354]

I ask this in a separate post again:

People said the amp is not stable and can oscillate. I want to confirm they are not talking about guitar easily feedback, that the amp can burst into uncontrollable squealing when you turn the volume up without guitar plugging in?

This is important because if you said the amp actually burst into oscillation, then there must be accidental or intention coupling from output back to input. That definitely change the sound.

If that is the case, find out where is the coupling by looking at the wiring. This is not a high gain amp, only 3 stages. You have to make it oscillate, it's not easy to oscillate by design. Trace the signal carrying wires and see where it couple back. Then recreate that.

I did search on web, most pictures are wiring of clones, not the real thing. The wiring diagrams I saw do not show some of the wires. eg, from the PI to G1 and G2 is not show, it just said going to the power tube.

I myself have no interest in building one as it has no MV and the mojo is from flat out driving the amp. I don't think I can appreciate the amp if I put in a MV, it has to be flat out. I just join in to learn something.

[Chuck H]

The PI to power tube leads are run from the front end side of the board, under the board to the power tube grid stopper at the tube pins. Most diagrams DO show this.

I can't say if any TW amps ever left Ken's workshop with an oscillation problem. I do know that it's not uncommon for the amps to exhibit some instability. Not pickup squealing acoustic feedback, actual lead dress instability within the amp. Ken hand picked tubes for each amp from a stash of NOS. He's also said to tune the amps to the edge of stability in some cases. A tube change could therefor cause instability if no other compensatory changes were made. I expect amps that exhibit the problem weren't serviced by Ken at TW.

[Alan0354]

I am just curious and join in this thread.

That's the thing about instability. That is electronics. In order to have oscillation, you have to have enough feedback from the later stage back to the preamp stage so the forward gain times the feedback is greater or equal to unity.......1) Gain X FB >=1 AND 2) phase shift is 360 deg. This is very important for this case. Maybe if you can find the point of coupling and recreate that, you might find your answer.

From the schematic, it's a pretty much a typical Fender BF/SF vibrato channel with the 3.3M and the cap removed and put in a 0.002 cap so you have a lot more signal going to the third tube. Gain is not super high, you have to make the amp oscillate. Only when you get to 4 stages preamp where things start to get dicey.

1) you have to have means of coupling. You need to find that. This is not a high gain amp, only 3 preamp stages and the last stage has gain less than 10. So the wire that couple back has to be quite obvious. From the drawing, I just don't see it. Could it be the placement of the OT and feedback through magnetic coupling? Remember induce EMF= area of the loop times flux density through the loop for single turn loop. To me, that might be more prominent that capacitor coupling between two wires.

2) Remember the phase shift has to be 360 deg. This means you can look at the phase shift of each stage and the tone stack to deduce out where is the most likely point of coupling back.

This part, there is no magic, you have oscillation, you got to have coupling. Any of these kind of coupling definitely changes the sound.......I don't know about making it sounds better, in my case with the first Bassman prototype, because of cross wiring in the progress of changing the amp, it gave me a strange kind of sound that I don't like. It is not in the final build in the KMD as I layout the circuit with good component flow.

I hope this might help your quest a little. Because the lack of MV, I am not really interested in building one, I am a lot more intrigued by the Dumble design that I build into the Bassman.


BTW, since this thread stalled for a day, I post this question in Amp garage as they have a Trainwreck forum. You might be interested in reading that if they reply. I am a believe of mojo, but I do believe there is a scientific reason for that. I am curious to hear the result.

The Amp Garage :: View topic - Is the real Trainwreck amp sound much better than clones?

[jazbo8]

Hmmm, really? Do you not see the mis-information that you are giving, who are you trying to help/confuse?

[Alan0354]

Hmmm, really? Do you not see the mis-information that you are giving, who are you trying to help/confuse?

Like how? If it is oscillation, there got to be a reason.

What misinformation? It's not as if you have better guess.

[loudthud]

The Trainwreck acts like it has more gain than you think it would. Probably due to the lack of interstage attenuation except for the loading of the second stage and absence of grid stoppers. (Some clones add a grid stopper on the second stage.) Depending on layout, it really wants to oscillate. It's a fight as you turn up the Treble and Presence controls.

What the Trainwreck lacks in gain, it makes up for in feel. Some clones are built with a PPIMV or VVR but these things have to be used in moderation to preserve the feel. Unless you build one, you won't understand.

[jazbo8]

Did you even bother to read all you can on the TW before making your grand pronouncement?! If you wan to help others - do not spread mis-information, and kindly get your terms and definitions right - look up phase margin for example. You are old enough to know better...

[Alan0354]

Did you even bother to read all you can on the TW before making your grand pronouncement?! If you wan to help others - do not spread mis-information, and kindly get your terms and definitions right - look up phase margin for example. You are old enough to know better...

Who are you to tell me? I was talking to Chuck SPECIFICALLY ON THE FEEDBACK( oscillation) issue. This is how you trace the feedback, do you even know what we are talking about? read the post, read the definition of positive feedback, THIS IS STANDARD IN ALL TEXT BOOKS IF YOU EVER CARE TO READ AND GET OUT OF THE TUBE MOJO!!! This topic is very well research since the 50s and books are written on stability issue.

If you have oscillation, you don't have phase margin, You even understand stability?

Or you pretend to know electronics JUST BECAUSE YOU WORK ON TUBE AMPS.

[Alan0354]

The Trainwreck acts like it has more gain than you think it would. Probably due to the lack of interstage attenuation except for the loading of the second stage and absence of grid stoppers. (Some clones add a grid stopper on the second stage.) Depending on layout, it really wants to oscillate. It's a fight as you turn up the Treble and Presence controls.

What the Trainwreck lacks in gain, it makes up for in feel. Some clones are built with a PPIMV or VVR but these things have to be used in moderation to preserve the feel. Unless you build one, you won't understand.

Oscillation is only one possibility to look into. I was just carry on a conversation with Chuck. When amp is at the edge of positive feedback ( oscillation), the sound change a lot. Maybe or may not be the sound people want. But that is one possibility to look into. If people find where the coupling point is, then they can manipulate the coupling to see whether that's where the amp get the desirable sound. This is just a possibility, nothing more.

[Alan0354]

Did you even bother to read all you can on the TW before making your grand pronouncement?! If you wan to help others - do not spread mis-information, and kindly get your terms and definitions right - look up phase margin for example. You are old enough to know better...

I see you try to bring the fight over the the other forum. You better read up what I said before you talk about oscillation. try to keep up.

[jazbo8]

try to keep up.

Nice! You are truly a clueless schmuck, and I'm really tired of correcting your mistakes and mis-conceptions. Are you a "40-year EE" or just pretending to be one?!

[Alan0354]

Nice! You are truly a clueless schmuck, and I'm really tired of correcting your mistakes and mis-conceptions. Are you a "40-year EE" or just pretending to be one?!

Correct what? I challenge you to name the mistake. Did you even read my post clearly and look at the condition of oscillation? Read it. You can't even present an argument. Do you even study electronics? Again, do you really think you work on tube amp and you know electronics?

[Chuck H]

That's the thing about instability. That is electronics. In order to have oscillation, you have to have enough feedback from the later stage back to the preamp stage so the forward gain times the feedback is greater or equal to unity.......1) Gain X FB >=1 AND 2) phase shift is 360 deg. This is very important for this case. Maybe if you can find the point of coupling and recreate that, you might find your answer...

1) you have to have means of coupling. You need to find that. This is not a high gain amp, only 3 preamp stages and the last stage has gain less than 10. So the wire that couple back has to be quite obvious. From the drawing, I just don't see it. Could it be the placement of the OT and feedback through magnetic coupling? Remember induce EMF= area of the loop times flux density through the loop for single turn loop. To me, that might be more prominent that capacitor coupling between two wires.

I can't answer on a technical level regarding the principals of feedback. I WILL say that in most guitar amps the cause is seldom obvious and is usually confounded such that any single node cannot, by itself, be located and pinned down. One exception would be signal coupling in the power supply. Some Fender HR series amps do this. The power supply decoupling nodes feed same phase stages and if the decoupling fails the amps will oscillate. Because of the "mojo" factor surrounding the TW amps there has been much discussion of the filters and decoupling nodes, but I don't think they're responsible in this matter. More likely a lead dress and/or layout issue. In this case the cause is confounded by multiple stages, all of which could be involved to a lesser or greater degree, rather than one location of non obvious coupling.

From the schematic, it's a pretty much a typical Fender BF/SF vibrato channel with the 3.3M and the cap removed and put in a 0.002 cap so you have a lot more signal going to the third tube. Gain is not super high, you have to make the amp oscillate. Only when you get to 4 stages preamp where things start to get dicey.

Not exactly. I've written similar statements re: the TW design in the past. But it's not accurate. In my own evolution as an "amp guy" I first saw details as over complex and very important. Once I learned a bit more about how the circuits were working I started seeing in broader strokes and omitting some degree of detail in my considerations. Now I'm back to recognizing the details again. As it turns out the details are always relevant. My point is that the TW amp isn't a slightly skewed Fender type design. About the only relevant thing they share is where the tone stack is located and the number of gain stages. A BMW sedan and a Ford pickup both have a motor, four tires and a steering wheel too, but they cannot be substituted effectively for each other for specific purposes.

The TW amp indeed doesn't produce much preamp clipping. But it doesn't use a master volume and the PI stage and power tubes are intended to add to the clipping. Effectively making five series stages that are tuned for clipping. Not unlike some modern preamps. Things you missed about the difference between the TW amps and the typical Fender vibrato channel include: tone stack coupling cap values, tone stack resistance values, power supply topology and values, power tube mu, preamp stage bias points and plate voltages. These may seem like small matters individually, but all combined they make for a bigger difference. Consider that early on the only differences between a Fender 5f6a Bassman and Marshall jtm45 were the transformer brand and speakers. Yet they are very different amps.

This part, there is no magic, you have oscillation, you got to have coupling. Any of these kind of coupling definitely changes the sound.......I don't know about making it sounds better, in my case with the first Bassman prototype, because of cross wiring in the progress of changing the amp, it gave me a strange kind of sound that I don't like.

PFB induced by lead dress and layout can have many effects, as you noted. IMHE they CAN be good. I'm recalling an amp I built where rotating the OT 30* changed the clipped tone of the amp completely. This difference in OT location was also the difference between the amp being stable or not. The closer I got to unstable, the better the OD sounded with a lot more hair on it and touch sensitivity. I'm certainly not saying this would always be the case. Just offering an example to illustrate that it's possible.

[Dave H]

Correct what? I challenge you to name the mistake. Did you even read my post clearly and look at the condition of oscillation? Read it. You can't even present an argument. Do you even study electronics? Again, do you really think you work on tube amp and you know electronics?

Careful, your phase margin is approaching zero there.

[Alan0354]

I can't answer on a technical level regarding the principals of feedback. I WILL say that in most guitar amps the cause is seldom obvious and is usually confounded such that any single node cannot, by itself, be located and pinned down. One exception would be signal coupling in the power supply. Some Fender HR series amps do this. The power supply decoupling nodes feed same phase stages and if the decoupling fails the amps will oscillate. Because of the "mojo" factor surrounding the TW amps there has been much discussion of the filters and decoupling nodes, but I don't think they're responsible in this matter. More likely a lead dress and/or layout issue. In this case the cause is confounded by multiple stages, all of which could be involved to a lesser or greater degree, rather than one location of non obvious coupling.
First of all, this is only a suggestion as there are too many talks. Regarding to power supply filtering, unless the layout is really wrong which I don't see, the first two stages has a 20uF with 9.1K resistor. The audio is NOT feeding back from that point. I did experiment the values before. Yes, it will oscillate if the cap fails for sure.


Not exactly. I've written similar statements re: the TW design in the past. But it's not accurate. In my own evolution as an "amp guy" I first saw details as over complex and very important. Once I learned a bit more about how the circuits were working I started seeing in broader strokes and omitting some degree of detail in my considerations. Now I'm back to recognizing the details again. As it turns out the details are always relevant. My point is that the TW amp isn't a slightly skewed Fender type design. About the only relevant thing they share is where the tone stack is located and the number of gain stages. A BMW sedan and a Ford pickup both have a motor, four tires and a steering wheel too, but they cannot be substituted effectively for each other for specific purposes.

The TW amp indeed doesn't produce much preamp clipping. But it doesn't use a master volume and the PI stage and power tubes are intended to add to the clipping. Effectively making five series stages that are tuned for clipping. Not unlike some modern preamps. Things you missed about the difference between the TW amps and the typical Fender vibrato channel include: tone stack coupling cap values, tone stack resistance values, power supply topology and values, power tube mu, preamp stage bias points and plate voltages. These may seem like small matters individually, but all combined they make for a bigger difference. Consider that early on the only differences between a Fender 5f6a Bassman and Marshall jtm45 were the transformer brand and speakers. Yet they are very different amps.

I am referring to gain wise, of cause we all know a change in value of one component can change the sound. I am just saying the preamp is from modification of the vibrato stage and the extra gain is from changing the 3.3M area. My whole point is the gain is not very high that usually the preamp is not very prone to oscillation. More likely from output area back to the preamp.


PFB induced by lead dress and layout can have many effects, as you noted. IMHE they CAN be good. I'm recalling an amp I built where rotating the OT 30* changed the clipped tone of the amp completely. This difference in OT location was also the difference between the amp being stable or not. The closer I got to unstable, the better the OD sounded with a lot more hair on it and touch sensitivity. I'm certainly not saying this would always be the case. Just offering an example to illustrate that it's possible.


Yes, the lead dressing can couple back. But I don't think the capacitance coupling is the cause. I am more suspect the magnetic coupling. More like loop antenna. Moving the lead on the output side might show some change of effect even though the wires are no where close to the preamp. Magnetic coupling ( more exact, EM coupling) happens even if the two loops are farther away.

Yes, I did say in the post that the feedback from OT is definitely a possibility. And that can be very hard to pin down other than moving the OT. Turning the OT 30 degree can make a day and night difference. Magnetic flux is very very directional. It is cook book practice that if you layout two inductors, the minimal coupling is if you make it in "T" shape. The worst is parallel. You might want to look into the field pattern of a transformer

I know it's not easy to trouble shoot and experiment. But so far you must have tried a lot of different ways. Nothing is more important to have your hands on the amp.....which I don't. I am just commenting on the amp easily oscillates and maybe, just maybe you can get something out of this part.

The first statement of feedback gain is common sense. The important part is the phase. In order to have oscillation, the phase shift from input to the feedback point has to be 360 deg ( in phase). Even if you have feedback gain of greater than 1, if the phase is not almost 360 ( or 0 deg), you won't have oscillation. That's the key that you can eliminate part of the circuit that is 180 deg off phase(each gain stage is inverted which is 180 out of phase. Every other stage is 360 out of phase. So theoretically, half the stages is not the coplit.) . In another word, you have to have PHASE MARGIN OF 0 DEG., or no phase margin to sustain oscillation.

As I repeat, this is only a suggestion that might or might not be relevant. Just a suggestion to look into, nothing more.