I do, all the time.
IRFP240-250-640-930 among others.
But take "tube characteristics" label with a grain of salt.
They do, sort of, when compared 1 to 1 as raw components on a bench or on Proto board but no component is used on its own.
Amplifier architecture has a BIG say on sound:
* Typical tube amp is made out of individual gain stages , loosely capacitor coupled, there is a literal phase splitter which puts out 2 out of phase signals to drive out of phase power tubes, whose signals are combined in an output transformer, also there is little or no negative feedback to compensate for errors.
Output impedance is high which interacts with weird impedance Guitar speakers a lot
* SS amps including MosFets of any kind are typically direct coupled which is way tighter, have no phase inverter but different polarity outputs or drivers which in due turn amplify positive or negative peaks, typically have high negative feedback which makes them "HiFi" and brittle, have high damping unless a special feedback is used, have no output transformer, etc.
In my book "MosFet amps are just robust SS bipolar {transistor) ones".
You can get closer if you mimic classic Tube amp architecture though: FET phase splitter, out of phase power MosFets, output transformer, little negative feedback.
But you need more linear "Audio/Lateral" MosFets there, think 2SK/2SJ series, VERY expensive and hard to find.
Common Vertical/Switching MosFets, cheaper and abbundant, such as those I mentioned above are esy too non linear for simple amplifier use
IRFP240-250-640-930 among others.
But take "tube characteristics" label with a grain of salt.
They do, sort of, when compared 1 to 1 as raw components on a bench or on Proto board but no component is used on its own.
Amplifier architecture has a BIG say on sound:
* Typical tube amp is made out of individual gain stages , loosely capacitor coupled, there is a literal phase splitter which puts out 2 out of phase signals to drive out of phase power tubes, whose signals are combined in an output transformer, also there is little or no negative feedback to compensate for errors.
Output impedance is high which interacts with weird impedance Guitar speakers a lot
* SS amps including MosFets of any kind are typically direct coupled which is way tighter, have no phase inverter but different polarity outputs or drivers which in due turn amplify positive or negative peaks, typically have high negative feedback which makes them "HiFi" and brittle, have high damping unless a special feedback is used, have no output transformer, etc.
In my book "MosFet amps are just robust SS bipolar {transistor) ones".
You can get closer if you mimic classic Tube amp architecture though: FET phase splitter, out of phase power MosFets, output transformer, little negative feedback.
But you need more linear "Audio/Lateral" MosFets there, think 2SK/2SJ series, VERY expensive and hard to find.
Common Vertical/Switching MosFets, cheaper and abbundant, such as those I mentioned above are esy too non linear for simple amplifier use
This is something that interests me greatly. I see no good reason why solid state amps can't come close to tube amps, they just have to be built right (similar asymmetrical distortion, the various things people call, "sag," etc) The thing is, its not like there's a book out there that tells you how to make modern and shiny discrete transistors sound like old, inefficient tube amps - probably -, but there's still plenty of literature out there on how to make a great sounding tube instrument amp. I think one just has to find a way to bridge that gap.
I have a Quilter Overdrive 202 which claims to recreate this effect with a solid state output. To my ears, whatever they are doing. it sounds good. Maybe I'll break it open and have a look at what's going on inside one day. There are a fair amount of small, tiny SMD parts, so I can't imagine it would be easy to trace.
The Sony VFETs and Tokin SITs have very triode-like characteristic curves and do sound like triodes. The Sony VFETs are getting hard to find but the Tokin SITs can still be found, although not cheap. But still cheaper than a DHT like a 300B or 2A3, and they last a long time.
Or you can use the Lovoltech LU1014D power JFETs for triode like performance as well. These are still available and are quite inexpensive.
Or you can use the Lovoltech LU1014D power JFETs for triode like performance as well. These are still available and are quite inexpensive.
Runoffgroove does have cool resources alright. Not necessarily at the same depth of books published on the design of old tube amps but that's an extremely unfair comparison. I think finding an authentic implementation of an output section combined with ROG's preamp stuff could give nice results.
For a solid state power amplifier section, I would have a look at the Amp Camp Amp designs of Nelson Pass. The ACA#1 is about 5W, distortion characteristic is progressive, with primarily low order 2nd harmonic up to about 2W then additional 3rd harmonic above that. That would be a good basis for a "Champ style" guitar amp, of course with a preamp section.
I've always wondered if those cheap and relatively easy to find PA system distribution matching transformers could be used in an output stage. Some have plenty of winding taps, which I believe can be symmetrical for 2 primary windings, center tapped. Of course, they have the multiple secondary taps also, for speaker connection.
The trick would be to design around the transformer you can actually still buy. I'm pretty sure I've read where folks have done this using tubes; I would think a SS device would be far more forgiving in an attempt to match whatever the impedance is, presented by a hopefully available center tapped winding designated as "70V" in its original usage intention.
One would think, such a topology would preserve at least some of the speaker impedance interaction with the output stage, going back through the magnetic coupling of the windings. I'm imagining Drains of FET devices connected to each side of a symmetrical primary winding, with a B+ voltage on the center tap. Biased to some appropriate quiescent current flow, with two signals from a phase splitter modulating that bias current.
I suppose here is where simulation becomes your friend. The B+, bias current, power dissipation and protection of the output devices from being connected to an inductor with constant current flow in case of disruption (like the speaker becomes accidentally disconnected - the FETs arent big metal plates in wide evacuated space) would be some design challenge.
The trick would be to design around the transformer you can actually still buy. I'm pretty sure I've read where folks have done this using tubes; I would think a SS device would be far more forgiving in an attempt to match whatever the impedance is, presented by a hopefully available center tapped winding designated as "70V" in its original usage intention.
One would think, such a topology would preserve at least some of the speaker impedance interaction with the output stage, going back through the magnetic coupling of the windings. I'm imagining Drains of FET devices connected to each side of a symmetrical primary winding, with a B+ voltage on the center tap. Biased to some appropriate quiescent current flow, with two signals from a phase splitter modulating that bias current.
I suppose here is where simulation becomes your friend. The B+, bias current, power dissipation and protection of the output devices from being connected to an inductor with constant current flow in case of disruption (like the speaker becomes accidentally disconnected - the FETs arent big metal plates in wide evacuated space) would be some design challenge.
I would think the 70 volt transformer should be fine with Fets. I have an amplifier from way back that uses transformers to match the load to the transistors. Eventually wanted to try going that direction (give me a reason to use some of the caps I salvaged from an organ, only rated for 100V).
Yes that should not be too difficult. Just deciding on what is a reasonable B+ to work with, you may only want 2:1 on the OT or decide for a higher ratio. Given that it starts to get more expensive for higher voltage supply transformers, capacitors, and power transistors that may come into play as well.
GaN Fets, SiC Fets can be an option here as they can provide the higher operating voltage (550V to 1.5 kV range!). They have also come down in cost due to demand in automobile electrification etc..
Is that a novel implementation or has that been done before? Just trying to imagine how to build that myself (i.e. what parts to use) because I am very intrigued by this.
For solid state I'd prefer to lose the output transfo, work with the ACA mini which has current feedback. I forget what the damping factor is on that design, though. was it 10? Anyways 10 Watts output power, complementary push-pull and you'd only have to build 1 channel for a guitar amp. The IRF Mosfets are readily available, inexpensive. I could add a two or three stage preamp to that, something like a Princeton but running JFET and/or compound JFET-BJT.
Maybe run the preamp at 18 V, and the power amp at 24 V.
10 watts is pretty loud using a good guitar speaker. I ran a TDA2030 at about 14W through a Jensen 12" Ceramic.. more than enough volume for a practice amp.
Maybe run the preamp at 18 V, and the power amp at 24 V.
10 watts is pretty loud using a good guitar speaker. I ran a TDA2030 at about 14W through a Jensen 12" Ceramic.. more than enough volume for a practice amp.
I am certainly interested in both approaches! Maybe I'll give the ACA mini a go first.
Are there any good resources on this subject matter you would recommend me to take a look at?
