A voltage regulator like my 21st Century Maida Regulator offers the lowest output impedance as you can get. In other words, it is as close to an ideal voltage source as practically possible. It also offers the highest ripple rejection possible. Of the two regulator types, it also dissipates the lowest amount of power, thus, making it much easier to integrate into a design.
A shunt regulator works well in applications that draw low current. Voltage references for example. For higher current applications - including many preamps - the shunt regulator is not an optimal choice. The ripple rejection is limited by the voltage divider formed by the series resistor and the dynamic impedance of the shunt and the output impedance is limited by the parallel combination of the two.
The bottom line is that I don't see any way a shunt regulator can outperform a series regulator when used in an audio amp.
I prefer to listen to the music and the amp. The power supply should be sonically transparent. Others are free to disagree...
~Tom
C1, C2, and C4 are needed for stability, hence, it is important that they are mounted directly on the board (i.e. not tucked away in a far corner of the chassis and connected by wires).
The value of C1 is not critical, just make it 1.0 uF or higher.
C2 needs to be 2.2 uF or higher for stability. It also helps protecting the regulator against high di/dt pulses from poor connections (think loose tube sockets and the like).
C4 needs to be 1.0 uF.
C2 and C4 need to be film types (low ESR). C1 can be electrolytic or film.
~Tom
The value of C1 is not critical, just make it 1.0 uF or higher.
C2 needs to be 2.2 uF or higher for stability. It also helps protecting the regulator against high di/dt pulses from poor connections (think loose tube sockets and the like).
C4 needs to be 1.0 uF.
C2 and C4 need to be film types (low ESR). C1 can be electrolytic or film.
~Tom
Last edited:
Any suggestions on how one could turn off the regulators output if the output tubes -ve bias supply passes a predetermined set point?
My preference is to stay with an unregulated -ve bias supply for my KT88 UL PP amp but my concern is that the output tubes could be damaged if the mains voltage should drop too low when an unregulated bias supply is used.
My preference is to stay with an unregulated -ve bias supply for my KT88 UL PP amp but my concern is that the output tubes could be damaged if the mains voltage should drop too low when an unregulated bias supply is used.
My solution would be to use a zener regulator (zener stack with a source follower) such as the one I use in my 300B design. That guarantees a stable bias voltage even if the input voltage does vary.
However, if you insist on an unregulated bias supply with a regulated B+ and want to implement a fail-safe, I suggest basing the fail trigger off of the bias current in the output tubes rather than just the bias voltage. I would use a relay to open the B+ supply. Watch out for the voltage rating of the relay. As with any failure detection circuit, you have to ensure that it engages when you want it to engage and does not engage when you don't want it to engage. Ensuring reliable start-up with those circuits can be a challenge.
Given that tubes are pretty rugged, all you may need is an over-current indicator. A small sense resistor with an LM311 comparator and a zener reference will do that. Use it to light an LED.
~Tom
However, if you insist on an unregulated bias supply with a regulated B+ and want to implement a fail-safe, I suggest basing the fail trigger off of the bias current in the output tubes rather than just the bias voltage. I would use a relay to open the B+ supply. Watch out for the voltage rating of the relay. As with any failure detection circuit, you have to ensure that it engages when you want it to engage and does not engage when you don't want it to engage. Ensuring reliable start-up with those circuits can be a challenge.
Given that tubes are pretty rugged, all you may need is an over-current indicator. A small sense resistor with an LM311 comparator and a zener reference will do that. Use it to light an LED.
~Tom
Or you could just put a nice Simpson milliammeter in the amp front panel and glance at it occasionally. You'll know if your current drifts too high, you'll be able to watch it over time to see just how stable it really is....plus, it looks kinda cool!
I put one in my 833 amp because I REALLY didn't want that running away!
I put one in my 833 amp because I REALLY didn't want that running away!
Tom, could you please explain why Maida regulators have problems with capacitive load connected to its output?
According to LM317 datasheet as quoted below, it says that by adding a 1.0u F tantalum or 25uF aluminum electrolytic capacitor on the output will ensure stability to the regulator. Thus, wouldn't doing this will fixed the capacitive loading problem?
Although the LM317 is stable with no output capacitance,
like any feedback circuit, certain values of external
capacitance can cause excessive ringing. An output
capacitance (CO) in the form of a 1.0u F tantalum or 25uF
aluminum electrolytic capacitor on the output swamps this
effect and insures stability.
A large capacitor on the regulator output means that the regulator starts up into, basically, a short circuit. I have seen several cascode devices in Mike Maida's original circuit explode under these circumstances. Now, there may be ways around that, but I chose to take the bull by the horns and design an updated circuit that offers several key advantages over Maida's original circuit:
- Much lower output impedance.
- Much better ripple rejection.
- Much lower keep-alive current needed in the regulator --> less need for bulky, expensive, and harder to find power resistors.
- Soft start.
- Etc.
~Tom
- Much lower output impedance.
- Much better ripple rejection.
- Much lower keep-alive current needed in the regulator --> less need for bulky, expensive, and harder to find power resistors.
- Soft start.
- Etc.
~Tom
Diodes
Thanks, Tom. That makes a lot of sense now.
It is a great design you did there. I have been reading through half of the posts in this thread (16 pages!! and still reading), and it seemed that a lot of people are satisfy with the 21st Century Maida Regulator.
The other thing that I wanted to ask you is regarding to the diodes in the circuit.
1) Why did you put D4 zener diode across the MOS's gate and source? It seems to me D4 doesn't do anything other than to ensure Vgs doesn't goes above 10V. I wonder if such condition could ever occur?
2) Is Zener Diode D3 there to provide inrush current by pass for the LT3080?
3) Why D1 and D5? How do they play their part in this design?
Thanks, Tom. That makes a lot of sense now.
It is a great design you did there. I have been reading through half of the posts in this thread (16 pages!! and still reading), and it seemed that a lot of people are satisfy with the 21st Century Maida Regulator.
The other thing that I wanted to ask you is regarding to the diodes in the circuit.
1) Why did you put D4 zener diode across the MOS's gate and source? It seems to me D4 doesn't do anything other than to ensure Vgs doesn't goes above 10V. I wonder if such condition could ever occur?
2) Is Zener Diode D3 there to provide inrush current by pass for the LT3080?
3) Why D1 and D5? How do they play their part in this design?
Last edited:
D1 is there to prevent the regulator from blowing if the output voltage is ever higher than the input voltage. This could happen on power-down if the main reservoir cap discharges faster than the caps in the circuit.
D2 biases the cascode, Q1.
D3 protects the IC against voltage spikes. It's probably not necessary, but I figured it provided good peace of mind for an entire 10 cents worth of silicon.
D4 protects the gate oxide of Q1. How the various voltages behave in the event of a momentary short circuit or loose tube socket is a bit of an unknown. D4 ensures that Q1 does not blow on those occasions. The STW12NK95Z actually has this diode built-in, so D4 is not populated if the STW... is used. Fitting a zener across the G-S junction on a MOSFET is common industry practice.
D5 protects the regulator against reverse voltage. This can happen in amps where the B+ supply discharges faster than the bias supply.
D6 allows for safe discharge of C5 on power-down.
In my day job, I design circuits that are mass produced in the hundreds of thousands if not millions, so I tend to take reliability seriously...
~Tom
D2 biases the cascode, Q1.
D3 protects the IC against voltage spikes. It's probably not necessary, but I figured it provided good peace of mind for an entire 10 cents worth of silicon.
D4 protects the gate oxide of Q1. How the various voltages behave in the event of a momentary short circuit or loose tube socket is a bit of an unknown. D4 ensures that Q1 does not blow on those occasions. The STW12NK95Z actually has this diode built-in, so D4 is not populated if the STW... is used. Fitting a zener across the G-S junction on a MOSFET is common industry practice.
D5 protects the regulator against reverse voltage. This can happen in amps where the B+ supply discharges faster than the bias supply.
D6 allows for safe discharge of C5 on power-down.
In my day job, I design circuits that are mass produced in the hundreds of thousands if not millions, so I tend to take reliability seriously...
~Tom
Last edited: