Yes it is and will work fine. There were others that wanted/discussed a more elaborate shunt regulator. In short it consists of a constant current generator and a power zener formed by the zener diode and shunt transistor.
Zeners wil determine the operating voltage while the current generator will set the operating current. Always set the current >20% than what you need so the zener can regulate the voltage.
110lbs each, of which some 65lbs is the chassis, machined from a solid 265lbs lump.
Also Swiss made, and twice the cost of a set or Ayre MX-R amps (the latter sound rather jolly, imo).
Sand casting would be much cheaper, but has the risk of surface defects due to porosity and small particle inclusions.
Means either tossing 1 out of 5, or touch-up MIG welding.
(not lit. based experience, i did sand casting 40 hrs a week for a year, including +250lbs items. Great for biceps, bad for back)
Nico, as you have pointed out many things have come out in the wash. I don't have time, but I will respond later. As the aim right now is not to try "improving" the original, and to just address the issues that raise eyebrows, so to say, I don't see the CCS as a problem. It's contribution to the performance in simulation is small. If we later try to improve the original design, I'm sure this will be addressed.
As far as stability, Jam has indicated there is something to be desired here. Goldmund didn't seem to be able to fix these issues with PCB design alone, so is it realistic that we should tune the circuit's own behavior?
Also, when checking pulse behavior I usually plot the time derivative of the output square wave. This is very revealing about the behavior of the amp.
I would like to try modeling the the oscillation models of the FETs (described in a Cordell article I think?) in simulation to more realistically determine which type of compensation is best.
Dacuycoy, your schematic is terminally flawed. Look at where the MPSA42/92 are. The CCS for both polarities is wrong.
- keantoken
As far as stability, Jam has indicated there is something to be desired here. Goldmund didn't seem to be able to fix these issues with PCB design alone, so is it realistic that we should tune the circuit's own behavior?
Also, when checking pulse behavior I usually plot the time derivative of the output square wave. This is very revealing about the behavior of the amp.
I would like to try modeling the the oscillation models of the FETs (described in a Cordell article I think?) in simulation to more realistically determine which type of compensation is best.
Dacuycoy, your schematic is terminally flawed. Look at where the MPSA42/92 are. The CCS for both polarities is wrong.
- keantoken
Dacuycoy,
Some thought of improving the circuit
a) Replace the 4.7k resistots with currenr doides
b) Add small cap across zeners
c) Split the first filter cap into two caps
d) Replacing the shunt element with a Darlington?
Question, do we need a cap after the current source?
Any thoughts guys, I have used this basic regulator before and can vouch that it sounds very good.
Jam
Some thought of improving the circuit
a) Replace the 4.7k resistots with currenr doides
b) Add small cap across zeners
c) Split the first filter cap into two caps
d) Replacing the shunt element with a Darlington?
Question, do we need a cap after the current source?
Any thoughts guys, I have used this basic regulator before and can vouch that it sounds very good.
Jam
Last edited:
Once the front end power supply is finished and TESTED, would it make any sense to anybody here to modularize the Front-end including the VAS to plug into a socket similar to what the other thread has from Goldmund. It, the module would facilitate trying out different topologies and or configurations with different components.
Have the power supply split with space in the middle to plug in the MODULE. The back end could be built with either a MOSfet or bipolar configuration with as many or less number of pairs as wanted. I am thinking of what I have seen OS do. Then when all is said and done and agreed on build a final board that everyone can do the group buy in.
Is there any chance of their being a thermal feedback for bias control? I have an old idea for a possible good one, IF that could be considered part of the topology of this project.
Just some thoughts that came into my head.
Take out your cutlasses and make shards of them at your leisure.
Have the power supply split with space in the middle to plug in the MODULE. The back end could be built with either a MOSfet or bipolar configuration with as many or less number of pairs as wanted. I am thinking of what I have seen OS do. Then when all is said and done and agreed on build a final board that everyone can do the group buy in.
Is there any chance of their being a thermal feedback for bias control? I have an old idea for a possible good one, IF that could be considered part of the topology of this project.
Just some thoughts that came into my head.
Take out your cutlasses and make shards of them at your leisure.
Without replace the fets.......
Adding the zener protection to the base of the drivers current limits the drive to the mosfets and allows the integral protection to cope with sub 1 ohm loads.
Without the current limiting the robust drive circuit will destroy the gate protection.
I reading this from Hi-Performance amplifiers by Ben Duncan.
Attached a part of the Rauch DVT Manual showing the zeners & mosfet drive.
Fairly relevant as very similar to the Goldmund drive circuit.
Rauch DVT 250s Dual Mono Lateral Mosfet,
per channel:-
4 x Hitachi 2sk1058 sj162 pairs,
1KVA transformer,
2 X 15,000uf caps
87 - 0 - 87 Volt rails.
300W @ 8R / 550W @ 4R.
Without the current limiting the robust drive circuit will destroy the gate protection.
I reading this from Hi-Performance amplifiers by Ben Duncan.
Many high output lateral mosfet amplifiers have the zener protection circuit. Without it a high power amp blows up fairly spectacularly!
Attached a part of the Rauch DVT Manual showing the zeners & mosfet drive.
Fairly relevant as very similar to the Goldmund drive circuit.
Rauch DVT 250s Dual Mono Lateral Mosfet,
per channel:-
4 x Hitachi 2sk1058 sj162 pairs,
1KVA transformer,
2 X 15,000uf caps
87 - 0 - 87 Volt rails.
300W @ 8R / 550W @ 4R.
Attachments
i had simulated nico's circuit with proteus profesional and here are the result. first picture 4k7 load. 2nd picture 1k5 load and 3rd picture 1k load
Attachments
Hi Dacuycoy,
When using a shunt regulator, you first need to decide what current you need and design your CCS around this value plus some spare current to operate your shunt and allow for variations that may occur.
The second criteria is to decide what the operating voltage of the system is and that determines the zener.
You also need to consider the power dissipation of the series pass transistor at the full current with relation to the volt drop across it.
Finally the power dissipation of the shunt transistor at full load (when there is no external load connected).
So you can choose values that suit the application quite easily. Keep in mind that the pos and neg rails will never be exactly equal as the PNP and NPN transistors have a difference in Vbe, so this will be the difference in rails.
I suppose that you can omit the series pass transistor with just a resistor, but I like having a CCS which allows for short circuit protection, in other words current limiting.
When using a shunt regulator, you first need to decide what current you need and design your CCS around this value plus some spare current to operate your shunt and allow for variations that may occur.
The second criteria is to decide what the operating voltage of the system is and that determines the zener.
You also need to consider the power dissipation of the series pass transistor at the full current with relation to the volt drop across it.
Finally the power dissipation of the shunt transistor at full load (when there is no external load connected).
So you can choose values that suit the application quite easily. Keep in mind that the pos and neg rails will never be exactly equal as the PNP and NPN transistors have a difference in Vbe, so this will be the difference in rails.
I suppose that you can omit the series pass transistor with just a resistor, but I like having a CCS which allows for short circuit protection, in other words current limiting.
Nico
The L Mosfet gate resistors on the Rauch circuit I posted are 680R.
Hope that helps with your practical test!
I tend to use different values for the N & P fets increasing the N value to 1K to compensate for the input capacitance.
May I point you to this post by Antony Holton on using dual die fets.
http://www.diyaudio.com/forums/solid-state/114953-vbias-semelab-double-die-lmosfet-3.html.
I have found this information very useful when trying to tame oscillating lateral mosfet amplifiers.
The L Mosfet gate resistors on the Rauch circuit I posted are 680R.
Hope that helps with your practical test!
I tend to use different values for the N & P fets increasing the N value to 1K to compensate for the input capacitance.
May I point you to this post by Antony Holton on using dual die fets.
http://www.diyaudio.com/forums/solid-state/114953-vbias-semelab-double-die-lmosfet-3.html.
I have found this information very useful when trying to tame oscillating lateral mosfet amplifiers.
Hi Xoc1,
I used a 22 000 uF cap charged to 50V as power supply and wired the MOSFET and driver with and without zener & diode on the driver. With a short i.e. capacitor connected across the drain and source I destroyed two MOSFETs blown open circuit. The fuses were 3.5 amp (fast blow) and remained in tact when the zener & diode was connected.
I then replaced the short with a 1 ohm load (source resistor) and the third MOSFET survived with and without the zener & diode on the driver but the fuse blew each time.
MOSFETS were Hitachi 2SK1058
I am not sure if this is useful information or proves anything really.
If I may suggest would it not be of some significance to use a source resistor instead of nothing (Goldmund) as there would be a voltage developed across the source resistor which may bring the zener protection into play.
These are expensive fets so I guess many tests would be out of the question.
I used a 22 000 uF cap charged to 50V as power supply and wired the MOSFET and driver with and without zener & diode on the driver. With a short i.e. capacitor connected across the drain and source I destroyed two MOSFETs blown open circuit. The fuses were 3.5 amp (fast blow) and remained in tact when the zener & diode was connected.
I then replaced the short with a 1 ohm load (source resistor) and the third MOSFET survived with and without the zener & diode on the driver but the fuse blew each time.
MOSFETS were Hitachi 2SK1058
I am not sure if this is useful information or proves anything really.
If I may suggest would it not be of some significance to use a source resistor instead of nothing (Goldmund) as there would be a voltage developed across the source resistor which may bring the zener protection into play.
These are expensive fets so I guess many tests would be out of the question.
Last edited: