I would like to create a PCB for a "Pignose" style guitar amp. I always kinda liked this old transistor radio style design because it's super quiet but as soon as you touch a string on the guitar it explodes [1]. Fidelity sucks but this is for guitar.
Anyway, here is my LTSpice model so far which is the classic transistor radio style design with a slightly modern twist of using an op amp for the preamp:
But before I can go further I need some design help. I'm not an engineer and I have discovered that using an interstage transformer like this is sensitive to inductance values (it's very sensitive to leakage inductance) and I need to figure out some parts issues.
Generally my feeling is that a power amp should be designed in reverse. Meaning pick a wattage, then an output transformer and work backwards to the preamp. But there aren't a lot of transistor output transformers which from the simulation and from deductive reasoning I figure needs to have a primary of only about 100R or so. For example, the LTSpice model above uses the Hammond 149Q which is 200R:3.2R. This is not ideal because in practice the load is 8R which makes the primary is going to be more like 500R so I'm just not going to be able to push that much current into the OT unless I use a relatively high voltage (it's also unnecessarily expensive). So it would be nice if the PCB could work with whatever old transform I can find large or small. Ideally I would like something that is like 100R:8R and 5W. Does anyone know of a source for such a transformer?
Would you agree that the above design is sound? There are many many examples of this type of circuit on the Internet but virtually all of them are illustrative. They are just back-of-a-napkin type sketches. I need a real circuit with real parts and real values. If I add any leakage inductance (coupling value 0.98) the circuit as a strong tendency to oscillate so it would be great if I knew what methods are used to combat oscillation in this circuit.
Another parts problem is the transistors. At first I was thinking about trusty BD139. But it seems they are way too fast because the circuit is hard to stabilize when the circuit is overdriven (as guitar amps often are). If I sub with old 2N3055, it works much better. But if the transistors are going to be soldered to the PCB and also mounted on a heat sink, I need a TO-220 package. What is a decent *slow* transistor in a TO-220 package?
Again, I'm not looking for hi-fi, high performance and linearity. This is for guitar so it is more important what happens when the amp is NOT linear. OTOH it would be nice if it were possible to make it more linear. At least I would like to understand if there were some simple changes that could be made to improve fidelity.
[1] I have other reasons for exploring this design. In particular I am thinking about adding a rectifier circuit to control the bias so that it is low when there is no signal and then high when there is signal. So it's got a builtin noise gate. Similarly, because it's single supply and has a high ratio of peak current compared to the quiescent current (1A / 0.002A maybe), I could get a builtin compression effect by making the supply sag.
Anyway, here is my LTSpice model so far which is the classic transistor radio style design with a slightly modern twist of using an op amp for the preamp:
But before I can go further I need some design help. I'm not an engineer and I have discovered that using an interstage transformer like this is sensitive to inductance values (it's very sensitive to leakage inductance) and I need to figure out some parts issues.
Generally my feeling is that a power amp should be designed in reverse. Meaning pick a wattage, then an output transformer and work backwards to the preamp. But there aren't a lot of transistor output transformers which from the simulation and from deductive reasoning I figure needs to have a primary of only about 100R or so. For example, the LTSpice model above uses the Hammond 149Q which is 200R:3.2R. This is not ideal because in practice the load is 8R which makes the primary is going to be more like 500R so I'm just not going to be able to push that much current into the OT unless I use a relatively high voltage (it's also unnecessarily expensive). So it would be nice if the PCB could work with whatever old transform I can find large or small. Ideally I would like something that is like 100R:8R and 5W. Does anyone know of a source for such a transformer?
Would you agree that the above design is sound? There are many many examples of this type of circuit on the Internet but virtually all of them are illustrative. They are just back-of-a-napkin type sketches. I need a real circuit with real parts and real values. If I add any leakage inductance (coupling value 0.98) the circuit as a strong tendency to oscillate so it would be great if I knew what methods are used to combat oscillation in this circuit.
Another parts problem is the transistors. At first I was thinking about trusty BD139. But it seems they are way too fast because the circuit is hard to stabilize when the circuit is overdriven (as guitar amps often are). If I sub with old 2N3055, it works much better. But if the transistors are going to be soldered to the PCB and also mounted on a heat sink, I need a TO-220 package. What is a decent *slow* transistor in a TO-220 package?
Again, I'm not looking for hi-fi, high performance and linearity. This is for guitar so it is more important what happens when the amp is NOT linear. OTOH it would be nice if it were possible to make it more linear. At least I would like to understand if there were some simple changes that could be made to improve fidelity.
[1] I have other reasons for exploring this design. In particular I am thinking about adding a rectifier circuit to control the bias so that it is low when there is no signal and then high when there is signal. So it's got a builtin noise gate. Similarly, because it's single supply and has a high ratio of peak current compared to the quiescent current (1A / 0.002A maybe), I could get a builtin compression effect by making the supply sag.