723 psu issue

[m3vuv]

Hi all,im putting together a psu using a 723,im using this schematic,the iss use a lower value sense resistorue i have is,i want the current feed back to work from about 100ma upwards,the issue i have is to get 600mv from the current sense resistor,it has to be a high value to produce 600mv across it for 723,this lowers the output to much overall,is it possible to amplify  the sense voltage to feed the 723 cl and cl pins,therefor use a lower value sense resistor,ive thought of using a sg3532 but they are expensive!,cheers m3vuv.

[stj]

why parallel drive?

i have stuff i repair that can do atleast 7A that just uses a single 2n6057 darlington.

[m3vuv]

what has that got to do with what im asking??

[stj]

it's relevent to the circuit being over-complex.

[duak]

Back in the 70's I built a 723 based lab supply that still works to this day.  The downside is that there's no elegant way of having a wide current limit range without switching.  If you have a variable sense resistor, it has to have a large enough value for the low current setting while being able to carry the supply's maximum current.

m3vuv's comment made me think that it should be possible to amplify the developed current sense voltage to feed the internal current limiter transistor.  It seems like an obvious solution, but I've never seen it.  A 1X to 70X variable gain amplifier should be a piece of cake to design.  One of the nice things about the 723 current limiter is that it's very fast in that there is only one transistor's delay in reducing the current.  An amplifier will add some delay.  During high current transients (short circuits, charging capacitors, etc.)  I could see the amplifier winding up and applying too much drive current to the current limiter transistor and then possibly damaging it.  I would add a few parts to limit the output span and drive of the amplifier.

I could also see using the existing current limiter to limit the maximum current, while adding an amplifier and second current limiting transistor in parallel to implement a precision variable current limiter.

Hope this makes sense,

Cheers,

[Kleinstein]

It is not very practical to divide down the shunt voltage to adjust the current. The more practical way is to add some voltage to lower the 600 mV threshold. This could be part of transistor BE junction, e.g. the power transistor.

To lower the voltage lost, one can also use the emitter resistors that are there for current sharing anyway.

Using separate parallel power transistors is a good solution, nothing bad with this. From some point on it gets difficult to use a single large transistor. Much more than about 100 W from a high power TO3 case (e.g. 2N3773) is rather difficult. It is much easier to use several smaller ones at maybe 60 W each.

If a low drop out is required, one might consider a combination of NPN and PNP, e.g. in a kind of semi low drop configuration. Already a separate filter cap for the LM723 supply could help.

P.s.: at 30 V maximum output and thus some 35 V raw voltage already 2.5 A is a little on the high side for an 2N3055. It would be a good idea to add another 2 or 3 .  0.1 Ohms are also rather low for the emitter resistor.

[m3vuv]

the first part lost me there klienstien,im not sure what you mean by the be junction etc,im using 6x 2n 3055 pass transistors,im using the drop across the current sharring resistors as it is,there is one on each emitter leg of the pass transistors  (6x 0.1 ohm) i still need 5+ amps draw to get it to go into current limiting,i want it to start limiting at 200-300ma up to 30A.

[iMo]

..i still need 5+ amps draw to get it to go into current limiting,i want it to start limiting at 200-300ma up to 30A.

you need a "logarithmic" conversion then,for example an opamp with a diode in the feedback needs to be added into the sense circuitry..

[xavier60]

This should be able to made to work with some experimenting.
Q2 and associated circuitry is an adjustable current source which sets a voltage drop across R5 which determines the current limit threshold.
One unknown is how much Base current that the 723's current limiter transistor needs.

[xavier60]

The V-I curve will have a soft knee at the lower current settings.

[not1xor1]

why parallel drive?

i have stuff i repair that can do atleast 7A that just uses a single 2n6057 darlington.

It looks like you do not know the meaning of SOA, so please do not spread potentially dangerous misinformation.
A 2N6057 cannot withstand more than 20Vce at 7A and that just in ideal conditions (i.e. infinite heatsink) while the schematic provided by the original poster mentions an input voltage of 40VDC.

So a single 2N6057 would work with 7A just for a low voltage PSU.
Low voltage, even in a fixed voltage PSU, because you must not forget to take into account Vce in case of shortcircuit.

[not1xor1]

Hi all,im putting together a psu using a 723,im using this schematic,the iss use a lower value sense resistorue i have is,i want the current feed back to work from about 100ma upwards,the issue i have is to get 600mv from the current sense resistor,it has to be a high value to produce 600mv across it for 723,this lowers the output to much overall,is it possible to amplify  the sense voltage to feed the 723 cl and cl pins,therefor use a lower value sense resistor,ive thought of using a sg3532 but they are expensive!,cheers m3vuv.

If you want to build a variable voltage/current PSU (i.e. lab PSU) then it is better to use a 2 opamps design rather than the 723.
Any current regulation with the 723 would always be quite coarse and tend to drift unless you add another opamp and at that point IMHO it just does not make sense to use a 723.

[xavier60]

Using a Base Emitter junction for sensing the shunt voltage is such a blunt instrument, only good for about five to one control range.
I built a floating type, two op-amp bench supply that is happy to regulate current over a range of 20ma to 20a with a 10 milliohm shunt.
I paralleled ten  0.1 ohm resistors.

[m3vuv]

interesting,would you have the schematic?

[xavier60]

Yes, it can be found in another thread. I show only the regulator section.
I  explain the reason for it having a negative reference voltage and some extra circuitry that fixes a problem sometimes mentioned where op-amp current control loops are usually slow to respond to sudden overloads.
https://www.eevblog.com/forum/projects/lab-power-supply-the-lost-current/100/

[xavier60]

I must mention also that it will not be capable of safely supplying 20 amps without  pre-regulation.
And I don't think that MOSFETs are suitable for paralleling. I haven't really looked into it.

[xavier60]

This should be able to made to work with some experimenting.
Q2 and associated circuitry is an adjustable current source which sets a voltage drop across R5 which determines the current limit threshold.
One unknown is how much Base current that the 723's current limiter transistor needs.

This idea had already been thought of in post #27, https://www.eevblog.com/forum/projects/linear-lab-power-supply/25/

[m3vuv]

hi xaviar60 any idea what sort ov size pot v3 should be in the schematic you sent?,cheers Paul m3vuv.

[m3vuv]

i have some of these(op05cp intrument amprifier ic's),could i use one somehow to amplify the voltage generated across the sense resistor to feed the cl/cs pins on the 723 somehow?.

[xavier60]

hi xaviar60 any idea what sort ov size pot v3 should be in the schematic you sent?,cheers Paul m3vuv.

The usual 10k linear pot. It will be ok with a lower value so long as R1 is of a value that will pass enough current to keep D2 forward biased at times when the unregulated rail drops in voltage.

[xavier60]

i have some of these(op05cp intrument amprifier ic's),could i use one somehow to amplify the voltage generated across the sense resistor to feed the cl/cs pins on the 723 somehow?.

Problem is that even if a single rail type op-amp is used, it will need a + supply rail referenced to the PSU's + output terminal. When the PSU is working near its full output voltage, the + rail for the op-amp will approach or even need to exceed the  unregulated rail voltage.
There are a few options to make this work. It is starting to evolve into a floating type regulator design.

Extra: It is difficult to ignore the fact that the LM723 design looks so uncomplicated.

[xavier60]

Does anyone know what the dropout voltage is for a typical LM723 design like in the first post?

[xavier60]

An op-amp could be powered from the full unregulated rail voltage if one can be found that is rated for what the maximum unregulated voltage is and its input common mode range has to include its - supply voltage to within about 2V of its + supply voltage.

[perieanuo]

why parallel drive?

i have stuff i repair that can do atleast 7A that just uses a single 2n6057 darlington.

With 150w transistor you stabilise 5V from over 30Vdc and output 7 amps?



Envoyé de mon iPad en utilisant Tapatalk

[David Hess]

Xavier60 has the right idea for a simple variable current limit; the current through the resistors in series with the base subtracts from the base-emitter voltage lowering the current limit.  The same trick works for 317 based current limit circuits.

Xavier60 also identifies the problem with Vbe current limiting; the curve is soft and this only gets worse as the added offset is increased limiting control range.

If a higher performance current limit is desired, then instead of an amplifier, I would consider using the error amplifier from a second 723 to duplicate how the MC1466L works.