negative voltage reference problem

[Jwillis]

Hello I'm trying to get a -10 voltage for the reference and was wondering if I can pull that from the center tap and ground and negative regulate it or should I draw it of the 20 volt side of the transformer and invert it .I know I should have use the transformer differently but its all I have and I need a minimum of 25 volts at the final output.The rest of the circuit is over built but again its all I have.Any ideas or suggestions would be helpful.

[danadak]

Cleanest way would be use 20V side, you will have to ground one side
of bridge to set ground ref thruout design, or use a charge pump IC
off the positive supply.


Regards, Dana.

[StillTrying]

Is the 10K variable voltage pot wired correctly, it seems odd to have the min. output voltage not at the track end.

[danadak]

I am thinking you lift the ground end of pot and tie that to wiper.

That will provide offset when pot is 0 ohms to get ~ 0 Vout, and other
pot settings gain up its Vout.


Regards, Dana.

[Kleinstein]

As the negative supply currents seems to be low, one could use a charge pump from the 50/60Hz line frequency, from the same transformer.

Using a 35 V split coil transformer for an output of up to  25 V is a little odd. It is just to low that one can use the lower voltage from the center tap.  One should consider using some kind regulation already before the large filter caps. As the voltage is rather high, there is no real need to use extra large caps. Already a simple (though not small) filter inductor for passive PFC could save quite some power loss. Even just a suitable resistor can help to reduce the power loss.

If adjustable and also used below 20 V I would consider provisions to use the center tap of the transformer too. This can reduce the worst case power loss to something like 60% - so less heat sink at the costs of a few extra diodes and lower power transistors.

The worst case DC voltage can be slightly above 50 V - this might be too much for the regulator.

[schmitt trigger]

Hmmmmmmmmmmmmmm.

Since the emitter followers are outside of the feedback loop, wouldn't this affect regulation as the load changes?

But to your original question, The charge pump method outlined above should be the best approach for a negative reference.

[Jwillis]

Is the 10K variable voltage pot wired correctly, it seems odd to have the min. output voltage not at the track end.

Ya my bad on the Variable pot .That should be wired as a rheostat.Thankyou

As the negative supply currents seems to be low, one could use a charge pump from the 50/60Hz line frequency, from the same transformer.

Using a 35 V split coil transformer for an output of up to  25 V is a little odd. It is just to low that one can use the lower voltage from the center tap.  One should consider using some kind regulation already before the large filter caps. As the voltage is rather high, there is no real need to use extra large caps. Already a simple (though not small) filter inductor for passive PFC could save quite some power loss. Even just a suitable resistor can help to reduce the power loss.

If adjustable and also used below 20 V I would consider provisions to use the center tap of the transformer too. This can reduce the worst case power loss to something like 60% - so less heat sink at the costs of a few extra diodes and lower power transistors.

The worst case DC voltage can be slightly above 50 V - this might be too much for the regulator.

My Kicad experience is low
The output VAC was written to high.Sorry for the confusion .The current bench build is giving 27.5VAC before the rectifier 37.5 VDC after the rectifier and  0.6 to 26.1 VDC at the final output.Currently theirs only 2 10000uf cap installed but figured 4 might be better for high loads.

I was thinking of a 7910 off the negative side and center tap but was thinking that the negative side was to high.I would prefer to save the 20V side for supplying  the  digital indicators.

[Kleinstein]

The last circuit drawing does not work. One can not just use the center tap to get a more negative supply. At best you get something in the center. It is only with a series capacitor and extra diodes to make it a kind of voltage doubler / charge pump that one gets a negative voltage from this.

The output voltage will not be very stable, as there is no feedback from the output.

Having the caps not too large prevent the power factor from getting really bad - so larger caps would be only needed if the ripply gets to high. With a large extra margin in voltage, they don't need to be that large. The margin could even get larger if 2 separate diodes and a small filter cap is used for the regulator part.

The current ratings are also a little confusing the large caps and 30 A fuse suggest a high current, but with a 37 V DC input, the 4 transistors would be good only about 2-2.5 A each, if mounted on a really large heat sink.

[Jwillis]

The last circuit drawing does not work. One can not just use the center tap to get a more negative supply. At best you get something in the center. It is only with a series capacitor and extra diodes to make it a kind of voltage doubler / charge pump that one gets a negative voltage from this.

The output voltage will not be very stable, as there is no feedback from the output.

Having the caps not too large prevent the power factor from getting really bad - so larger caps would be only needed if the ripply gets to high. With a large extra margin in voltage, they don't need to be that large. The margin could even get larger if 2 separate diodes and a small filter cap is used for the regulator part.

The current ratings are also a little confusing the large caps and 30 A fuse suggest a high current, but with a 37 V DC input, the 4 transistors would be good only about 2-2.5 A each, if mounted on a really large heat sink.

OK cool that would save me some smoke.At a later date I intend on replacing the transformer to something more suitable.I just have limited resources where I live so the idea is to have a custom E core built at a local Rewind.I haven't done a controlled load test but a uncontrolled accident does blow the 30amp fuse without ill effect to the rest of the circuit.Weird.
I wonder if running  a second positive regulator  then invert for the negative reference or would that make it unstable as well.
Good thing I'm having fun experimenting.
 
 

[Kleinstein]

There is no need to go the path of a stable positive supply and than a charge pump chip to make a negative supply. One can better use the line frequency AC to drive a charge pump made from a cap and a few diodes. Stabilization can than be with a zener diode / shunt regulator like the LM113 shown.

For a short time, like a sudden short the transistors can deliver more current and might survive a short. However longer time (e.g. a few seconds) overload (e.g. more than 8-10A total at a low output voltage) might destroy the transistors. This can be faster than the heat sink heating up, just from exceeding the transistors FBSOA. Even a 10 A fuse might not provide enough protection.

There should definitely be a fuse between the transformer and the rectifier. The additional resistance of the fuse even helps.

[Jwillis]

Confused again.What does FBSOA mean and are we talking about the same transistor.I was under the impression that these https://www.onsemi.com/pub/Collateral/MJ15003-D.PDF could take the abuse.
Do you have a simple schematic of that AC driven charge pump ? I'm not that advanced.

[not1xor1]

Confused again.What does FBSOA mean and are we talking about the same transistor.I was under the impression that these https://www.onsemi.com/pub/Collateral/MJ15003-D.PDF could take the abuse.
Do you have a simple schematic of that AC driven charge pump ? I'm not that advanced.

this is an example of how you can get multiple DC outputs from a transformer with just one secondary winding:

[Damianos]

Why is there a need for a negative (and stabilized!) voltage? As is the circuit, just put an additional diode in series with the D7... By the way, what is the primary role of this diode and resistor R5?
This is a variable voltage power supply without regulation, so there is no sense to try to adjust it with precision, at any value, especially at the lowest range. Imagine connecting to its output a 5V or 3V device and expecting to survive!
The circuit seems to have "designed" with the random felling of parts on the table!!!
... ...
A simple way to implement this is to remove all electronic components of the circuit, keeping the transformer, rectifier and filter capacitors ... place a variac to the mains' side. Doing so, you will have better results in terms of stability of voltage and much more, the efficiency.
Otherwise, consider a switching power supply.

I you insist to design a linear power supply, start studying from what is each component, its properties and how it can be used with other components in a circuit. An about 1kW circuit is not a good start for a newcomer to electronics.

[Jwillis]

Why is there a need for a negative (and stabilized!) voltage? As is the circuit, just put an additional diode in series with the D7... By the way, what is the primary role of this diode and resistor R5?
This is a variable voltage power supply without regulation, so there is no sense to try to adjust it with precision, at any value, especially at the lowest range. Imagine connecting to its output a 5V or 3V device and expecting to survive!
The circuit seems to have "designed" with the random felling of parts on the table!!!
... ...
A simple way to implement this is to remove all electronic components of the circuit, keeping the transformer, rectifier and filter capacitors ... place a variac to the mains' side. Doing so, you will have better results in terms of stability of voltage and much more, the efficiency.
Otherwise, consider a switching power supply.

I you insist to design a linear power supply, start studying from what is each component, its properties and how it can be used with other components in a circuit. An about 1kW circuit is not a good start for a newcomer to electronics.

The need for a negative rail is to pull down the variable regulators reference below 1.25 volts to achieve a 0 volt output.
The schematic  from not1xor1 is a means to achieve multiple rails "pre-regulator" ,not a complete PSU.
I don't care whether or not its the most efficient means .I'm learning new ideas and multiple methods to achieve a goal.
The supply I'm building  for multiple applications in one box.
I don't want a switching supply because they tend to be noisy and leaky if not done correctly.Linear supplies provide a reasonable amount of isolation for my requirements and low radio interference.
Sure I could buy one ,but aside from a quick shot of happy hormones ,where's the fun it that.Besides I already build the case.

this is an example of how you can get multiple DC outputs from a transformer with just one secondary winding:

Thank you very much not1xor1 ,I appreciate the help.

[Damianos]

Why is there a need for a negative (and stabilized!) voltage? As is the circuit, just put an additional diode in series with the D7... By the way, what is the primary role of this diode and resistor R5?
This is a variable voltage power supply without regulation, so there is no sense to try to adjust it with precision, at any value, especially at the lowest range. Imagine connecting to its output a 5V or 3V device and expecting to survive!
The circuit seems to have "designed" with the random felling of parts on the table!!!
... ...
A simple way to implement this is to remove all electronic components of the circuit, keeping the transformer, rectifier and filter capacitors ... place a variac to the mains' side. Doing so, you will have better results in terms of stability of voltage and much more, the efficiency.
Otherwise, consider a switching power supply.

I you insist to design a linear power supply, start studying from what is each component, its properties and how it can be used with other components in a circuit. An about 1kW circuit is not a good start for a newcomer to electronics.

The need for a negative rail is to pull down the variable regulators reference below 1.25 volts to achieve a 0 volt output.

I proposed a way to do that in the existing schematic, based on the measurements that you presented.

The schematic  from not1xor1 is a means to achieve multiple rails "pre-regulator" ,not a complete PSU.

No, it is not that. It can provide only some low current bias...

I don't care whether or not its the most efficient means .I'm learning new ideas and multiple methods to achieve a goal.

I doubt that anyone can learn without studying...

The supply I'm building  for multiple applications in one box.

I am curious of what specifications you are expecting from this, you did not refer them.

I don't want a switching supply because they tend to be noisy and leaky if not done correctly.Linear supplies provide a reasonable amount of isolation for my requirements and low radio interference.
Sure I could buy one ,but aside from a quick shot of happy hormones ,where's the fun it that.Besides I already build the case.

If you think it will be fun to play with it, I wish you enjoy it!

...

You can also have a look at the datasheets and the relative application notes from NS/TI, for the three terminal variable voltage regulators (such as the series of LM117, LM150, LM138). There you will find not only "dry schematics" but also analytic descriptions, calculations, advisements etc.

[rdl]

You can also have a look at the datasheets and the relative application notes from NS/TI, for the three terminal variable voltage regulators (such as the series of LM117, LM150, LM138). There you will find not only "dry schematics" but also analytic descriptions, calculations, advisements etc.

The basic regulator circuit, including the negative reference to obtain zero volts output, is basically straight from the NSC data sheet. I know because I built almost the same thing a few years ago.

[Damianos]

I refer to the entire circuit and how designing a regulated power supply.