This is something i need to learn now, please educate me people!
What is a cascode? What does it do? How do you cascode?
This is what i think, bear with me as i really dont know.. The books i have, dont give me a clear answer.
Cascode was something often used when hig power was a design criteria and devices with high SOAR didn´t exist. It halves the voltage and current the devices sees. You make a cascode when you series-parallel devices.
If this is right, why not use it now when even higher power is wanted?
What is the downside with it? It has to have one as people has left it behind them
For me, my novice thinking, why not use it, ramp up the voltage, use suitable parts that can handle the voltage, and voila, you have a very powerful amp
I know its not that easy, but that was a quick version of my thinking about cascode.
Can someone please educate me correctly about what a cascode is, what it really does, how you make a cascode, and why it is not used anymore? Also, what is the downside with it?
Sorry for the novice question but this really hinder me from higher levels of understanding
What is a cascode? What does it do? How do you cascode?
This is what i think, bear with me as i really dont know.. The books i have, dont give me a clear answer.
Cascode was something often used when hig power was a design criteria and devices with high SOAR didn´t exist. It halves the voltage and current the devices sees. You make a cascode when you series-parallel devices.
If this is right, why not use it now when even higher power is wanted?
What is the downside with it? It has to have one as people has left it behind them
For me, my novice thinking, why not use it, ramp up the voltage, use suitable parts that can handle the voltage, and voila, you have a very powerful amp
I know its not that easy, but that was a quick version of my thinking about cascode.
Can someone please educate me correctly about what a cascode is, what it really does, how you make a cascode, and why it is not used anymore? Also, what is the downside with it?
Sorry for the novice question but this really hinder me from higher levels of understanding
How much power do you want? At some rail voltage, the design--either the standard emitter follower or the 'stacked' output--may hit limits on the amount of voltage swing the "classic" circuit can deliver to the output stage without some modifications.
I may be experimenting a bit with higher rail voltages soon, using a forced convection heatsink "chimney" with insulated sections for each rail so that TO-3 devices can be mounted with no insulator, and four output pairs for each channel. My bench variac has an isolation transformer with a center-tap winding so I can push rails over 90 VDC. The rest will be whether I can afford to repeatedly replace blown transistors...
A bigger question to me is where to go for significantly improved distortion and noise beyond what the Leach circuit can produce. Hawksford error correction looks interesting, but is far beyond my design skills. I've never done a MOSFET amplifier, either. Leach's elegant design has worked so well that I'm having difficulty deciding where I should go next.
I may be experimenting a bit with higher rail voltages soon, using a forced convection heatsink "chimney" with insulated sections for each rail so that TO-3 devices can be mounted with no insulator, and four output pairs for each channel. My bench variac has an isolation transformer with a center-tap winding so I can push rails over 90 VDC. The rest will be whether I can afford to repeatedly replace blown transistors...
A bigger question to me is where to go for significantly improved distortion and noise beyond what the Leach circuit can produce. Hawksford error correction looks interesting, but is far beyond my design skills. I've never done a MOSFET amplifier, either. Leach's elegant design has worked so well that I'm having difficulty deciding where I should go next.
The existing Leach design uses a cascode in the input stage, with a common base amplifier feeding the VAS stage. The driver and output stages have emitter followers stacked in series and a commercial version of the Super Amp had three sets of output devices; modern output devices in three pairs ought to be really rugged.
Basically, a cascode just divides the voltage the stage sees, but it also improves linearity depending on the design--I don't understand enough to explain the different implementations. But if you read Leach's notes, you see his explanations on why he chose to use--or not--a cascode configuration in each stage of the amplifier. A cascoded VAS might be a logical choice if you understand the design tradeoffs.
Paralleling additional output pairs increases the output current the whole amplifier stage can deliver while remaining conservatively within the SOAR of the individual devices. It also somewhat reduces distortion by staying within the more linear range of a given output device's current. So-called 'perforated emitter' output devices seem to be best in this regard, though their higher Ft >may< encourage oscillation.
On-Semi's "ThermalTrak" devices with in-package tracking diode incorporate perforated emitter design, and I've got a chassis using such devices in construction. I just haven't worked on it in years.
Basically, a cascode just divides the voltage the stage sees, but it also improves linearity depending on the design--I don't understand enough to explain the different implementations. But if you read Leach's notes, you see his explanations on why he chose to use--or not--a cascode configuration in each stage of the amplifier. A cascoded VAS might be a logical choice if you understand the design tradeoffs.
Paralleling additional output pairs increases the output current the whole amplifier stage can deliver while remaining conservatively within the SOAR of the individual devices. It also somewhat reduces distortion by staying within the more linear range of a given output device's current. So-called 'perforated emitter' output devices seem to be best in this regard, though their higher Ft >may< encourage oscillation.
On-Semi's "ThermalTrak" devices with in-package tracking diode incorporate perforated emitter design, and I've got a chassis using such devices in construction. I just haven't worked on it in years.
Damon:
I have sent you a PM with a bit more info.
I want 500w in to 8R. 400 would work too. One criteria is that the amp should be stable in to 2R as the original amp is. I don´t want to make any tradeoffs, thats why my novie thinking says stay close to the original layout and topology as it clearly works. The superamp, seems to me as being a better base as it already produces 300w.
AndrewT mentioned it is no need to use cascode as there exists better devices now that can handle higher voltage.
What about using them in the super amp?
I have sent you a PM with a bit more info.
I want 500w in to 8R. 400 would work too. One criteria is that the amp should be stable in to 2R as the original amp is. I don´t want to make any tradeoffs, thats why my novie thinking says stay close to the original layout and topology as it clearly works. The superamp, seems to me as being a better base as it already produces 300w.
AndrewT mentioned it is no need to use cascode as there exists better devices now that can handle higher voltage.
What about using them in the super amp?
I'd prefer using more modern output devices as they're more linear and more rugged than the originals. MJ21193/94 and 95/96 will work very well; I've used the 93/94s for many years. If you want to use plastic case devices, use the highest power rated ones.
Building a prototype amplifier to test at extended power levels ought to be the first step, to see if it'll survive and behave itself. I've stuck with two pairs of outputs at 60VDC rails because it's more than sufficient power and simpler to implement. If you really want 500 watts, five output pairs is probably best--I'm just concerned as to whether the drivers can deliver the necessary current/slew rate. A super amp ought to use at least three sets in parallel; I don't know of any designs that used four or more.
I really like best the concept of plastic case devices mounted on a single large heatsink with the driver board mounted directly to their leads; whole thing makes a single unit that can be easily tested by itself before final assembly and/or troubleshoot, and has the least possible amount of wiring clutter. It'll require a double-sided board with plated through holes and topside solder pads, and holes to tighten the transistor mounting screws. TO-3 devices would be more rugged, but require a different heat sink with lots of holes; an expensive custom design.
Building a prototype amplifier to test at extended power levels ought to be the first step, to see if it'll survive and behave itself. I've stuck with two pairs of outputs at 60VDC rails because it's more than sufficient power and simpler to implement. If you really want 500 watts, five output pairs is probably best--I'm just concerned as to whether the drivers can deliver the necessary current/slew rate. A super amp ought to use at least three sets in parallel; I don't know of any designs that used four or more.
I really like best the concept of plastic case devices mounted on a single large heatsink with the driver board mounted directly to their leads; whole thing makes a single unit that can be easily tested by itself before final assembly and/or troubleshoot, and has the least possible amount of wiring clutter. It'll require a double-sided board with plated through holes and topside solder pads, and holes to tighten the transistor mounting screws. TO-3 devices would be more rugged, but require a different heat sink with lots of holes; an expensive custom design.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.