Hello everybody.
After Maximax77 request, I want to share with you, a dual rail PS that I have used in all my class A builds.
The idea had come to me by an elektor negative power supply for printers.
So I designed a dual rail +- power supply using the famous 2N6287 / 2N6286 or 7 darlington transistors and used it after my diyaudio/firstwatt CRC board for all my class A amplifiers. The power supply is adjustable via a 2K5 multiturn trimmer and using a 0R1 / 5w resistor at the output, it can draw up to 10A / 50V. The 2N6284/86 transistors require a large heat sink whether have to draw more than 1A continiously. At the image below, the positive rail is only appeared( I don't have a circuit editor yet ) as the negative rail is identical but the transistors ( 2N6287 / BD140 / BC556 ) and the zener diode that is reversed for negative circuit ( The output cap must be flipped vertically too).
The zener diode is used for limiting the lower voltage regulation point. For example, if you want a 28VDC at the output, you will use a 24V zener so that when you adjust the trimmer for the desired DC out, the DC measurement will never go below 24VDC. When there is a good DC filtering at the input, the AC ripple is less than 0,2mV.The circuit is very simple and reliable so I don't think there is nothing else for me to add. Have a look and if somebody is interested tobuild it, the layout 6 gerbers are available.
Have a nice day.
Spiros
After Maximax77 request, I want to share with you, a dual rail PS that I have used in all my class A builds.
The idea had come to me by an elektor negative power supply for printers.
So I designed a dual rail +- power supply using the famous 2N6287 / 2N6286 or 7 darlington transistors and used it after my diyaudio/firstwatt CRC board for all my class A amplifiers. The power supply is adjustable via a 2K5 multiturn trimmer and using a 0R1 / 5w resistor at the output, it can draw up to 10A / 50V. The 2N6284/86 transistors require a large heat sink whether have to draw more than 1A continiously. At the image below, the positive rail is only appeared( I don't have a circuit editor yet ) as the negative rail is identical but the transistors ( 2N6287 / BD140 / BC556 ) and the zener diode that is reversed for negative circuit ( The output cap must be flipped vertically too).
Have a nice day.
Spiros
Last edited:
Thanks a lot Spiros. How much up-rating does the transformer need to cover losses in your circuit?
It's 3V . If your transformer after rectification is at 50 Volts, you can't get more than 47VDC. You can use that PS for a very accurate bench PS too.
Hello Spiros,
I am not able to send you a PM,
I am interested in the Gerbers.
best regards
Helmut
I am not able to send you a PM,
I am interested in the Gerbers.
best regards
Helmut
If the darlingtons get very hot, the transformer also needs a higher VA rating than with a simple CRC PSU. How much extra will that be?
Nice circuit. Any output impedance measuremnents, preferably against frequency?
A tip for a future PCB review: if you rotate the 2N6287 180 degrees, the heavy traces get much shorter.
Better yet, if you place both power transistors horizontal the PCB can be much smaller.
Jan
A tip for a future PCB review: if you rotate the 2N6287 180 degrees, the heavy traces get much shorter.
Better yet, if you place both power transistors horizontal the PCB can be much smaller.
Jan
WOW > at that price I would probably go for a slightly less compact design at much lower cost -
but then again, good class A amps aren't cheap to build in any case.
I have purchased a couple of them from Mouser for 7 or 8 euros each.
Same price on the local market in Greece.
You can play with gerbers and modify the pcb as you want.
I will try to place the transistors horizontally for more space but there must be space for the aluminum L bracket. The circuit works fine with a really low ripple.
Last edited:
I always order a slightly higher VA transformer. If 300VA is recommended, I order 350. My transformers have never got hot using that PS. If the amp needs a 3A drain, 5A is recommended.
I use a similar linear supply for my NUC11 that works as a streamer for my DAC. No need for the Darlingtons... MPSA18, 2SD313 and 2 X MJ15024 (on sizeable heatsink) will support 6-8A of constant current draw (my NUC can burn around 6A when using ASDM7ECv2 modulator to generate a fast DSD stream). That transistor combination worked extremely well. The front end is CLC, with fairly small C-values. Infinitely better, compared to any SMPS I tried (and I did try a lot, with various filters... what a disappointment that was...).
I encourage everyone to give this linear power supply a try...
I encourage everyone to give this linear power supply a try...
MJ15024 is a really a good choice, as many others I guess. Comparing the two, there are similar or identical characteristics exept of the collector current ( 16A vs 20A continuous and 30A vs 40A peak). If you have got the schematic you can share it with us. Class A amps need a powerful and reliable PS to eliminate ripple. An adjustable PS has a great advantage in diy designs. Is your PS adjustable?
That's Darlington review from RF Wireless World
➨It offers very high current gain in comparison to single transistor.
➨It offers very high input impedance.
➨It uses few components and hence can be used for easy circuit designs.
➨It can amplify signal to larger extent.
➨It allows designers to drive more power applications by few mA of current source.
➨It is very sensitive to current.
➨Photo-Darlington pair introduces less noise than photo-transistor with external amplifier.
➨Sziklai Darlington pair which uses cross configuration of transistors produces less heat.
➨Sziklai Darlington pair offers less response time.
➨There is voltage drop across base ("B") and emitter ("E") terminals when Darlington pair is in saturation (i.e. Fully ON) region.
➨Switching speed is lower.
➨Bandwidth is limited.
➨This configuration introduces phase shift at certain frequencies in negative feedback circuit.
➨It offers high power dissipation due to high saturation voltage.
➨Leakage current of first transistor is amplified by the next transistor. Hence overall leakage current of Darlington pair is higher.
Benefits or advantages of Darlington transistor
Following are the benefits or advantages of Darlington transistor:➨It offers very high current gain in comparison to single transistor.
➨It offers very high input impedance.
➨It uses few components and hence can be used for easy circuit designs.
➨It can amplify signal to larger extent.
➨It allows designers to drive more power applications by few mA of current source.
➨It is very sensitive to current.
➨Photo-Darlington pair introduces less noise than photo-transistor with external amplifier.
➨Sziklai Darlington pair which uses cross configuration of transistors produces less heat.
➨Sziklai Darlington pair offers less response time.
Drawbacks or disadvantages of Darlington transistor
Following are the drawbacks or disadvantages of Darlington transistor:➨There is voltage drop across base ("B") and emitter ("E") terminals when Darlington pair is in saturation (i.e. Fully ON) region.
➨Switching speed is lower.
➨Bandwidth is limited.
➨This configuration introduces phase shift at certain frequencies in negative feedback circuit.
➨It offers high power dissipation due to high saturation voltage.
➨Leakage current of first transistor is amplified by the next transistor. Hence overall leakage current of Darlington pair is higher.
General properties of one component in a complex circuit are not what we get out of the whole circuit.
However, there are two of high concern in that list: lower switching speed and bandwidth.
Class A operation has, for most designs, steady current only without signal, and any AC current at the amplifier output is drawn from the supply. So, PS' transient response (reaction to fast load changes) is very important. With 100 uF at the PS output, AC current output relies on regulator performance. Ripple (PSRR) is important but transient response and output impedance are critical IMO.
That’s why it is good to know how some regulator actually behaves. Stable DC voltage is not enough for a regulator measurements.
We are in the same camp as I don't build amplifiers without regulated PS. 🙂
However, there are two of high concern in that list: lower switching speed and bandwidth.
Class A operation has, for most designs, steady current only without signal, and any AC current at the amplifier output is drawn from the supply. So, PS' transient response (reaction to fast load changes) is very important. With 100 uF at the PS output, AC current output relies on regulator performance. Ripple (PSRR) is important but transient response and output impedance are critical IMO.
That’s why it is good to know how some regulator actually behaves. Stable DC voltage is not enough for a regulator measurements.
We are in the same camp as I don't build amplifiers without regulated PS. 🙂
It was just a suggestion. But you really should measure the output impedance. Low ripple is necessary, but if the output varies with load current, that should also get attention. It is a regulator, so it would be nice to know it actually does that and how well it does that.
Jan
With 100 uF at the PS output, AC current output relies on regulator performance.
As you can read from post#1, I use the diyaudio/firstwatt CRC for a 60 or 80mF capacitance. the measured ripple at that PS output is not more than 0.5mV.
As you can read from post#1, I use the diyaudio/firstwatt CRC for a 60 or 80mF capacitance. the measured ripple at that PS output is not more than 0.5mV.
As i can recall from my memory, I had tested the circuit with National Instruments 12.0, using a dummy load at the output. Both V and A were steady. In real world, my 4 class A amps work superbly with that PS.