Power Supplies Part 1 and 2

[EEVblog]

What about the  LT1963AEQPBF? http://www.alldatasheet.net/view.jsp?Searchword=LT1963AEQPBF
It is about the same price, easier to find and has almost the same caracteristics.

Might as well use the LM317, unless you need low dropout, as the LT1963 only goes down to 1.2V.

Dave.

[shebu18]

I want a regulated power supply for my projects, the one thing is that i will need around 5 amps at 7,4V for my rc-car tests. I have made a power supply with a lm371 and a power transistor 9as shown in the schematics). The thing is that i can't control the amps.
I like the design you made because i can also control the supply via a uC and the current protection.

Offtopic, what was the forumula for the lm371 for calculating the output voltage? I found one in the schematics but i don't know what Iadj is and where to find it.
Thanks. You do a great job!!


@ejeffrey: Thank You.

[ejeffrey]

Iadj is about 50 microamp, so it should provide a negligible offset in most circumstances.  It is listed in the data sheet.  For the standard configuration where you use a voltage divider with a fixed 240 ohm feedback resistor, it will give approximately a 1% error term.  If you make it programmable by driving the adjust pin from an op-amp, the adjustment current has no effect.  If you use larger feedback resistors, the adjustment pin current can cause a more significant offset.

[RJSC]

If you feed a negative voltage to the opamp's negative rail instead of ground, you can force the LM317 to Zero, by driving the ADJ pin to -1.25 V. With a pot, you can connect its low side to -1.25V.

Given that the ADJ pin current is so low, you can generate the negative supply with a dedicated charge pump IC or with an oscillator (555 for example) to avoid the need of a second small transformer or a big transformer with 2 windings.

You can read something about it here: http://www.stephenhobley.com/blog/2011/03/28/creating-a-negative-voltage-from-a-positive-one/

Maxim has a lot of charge pump IC's that require only external capacitors, but I don't know of any that should be available at every electronics store like the LM317.

[McMonster]

Given that the ADJ pin current is so low, you can generate the negative supply with a dedicated charge pump IC or with an oscillator (555 for example) to avoid the need of a second small transformer or a big transformer with 2 windings.

I've seen a few circuits where forward biased diodes are used in series from some point in the circuit to ground to get a small reference voltage equal to voltage drop across diodes below ground, for example in LM35DZ's datasheet. Is this applicable here? Where else it would work and where not?

[RJSC]

I've seen a few circuits where forward biased diodes are used in series from some point in the circuit to ground to get a small reference voltage equal to voltage drop across diodes below ground, for example in LM35DZ's datasheet. Is this applicable here? Where else it would work and where not?

It is applicable but is has some problems.
The voltage you are dropping on the diodes times the current will be dissipated as heat, and you have to provide an higher voltage transformer to account for that drop or have a lower output with the same transformer.
Also, a diode does not switch on cleanly at 0.7V, meaning that at a light load they will drop less than 0.7V and with high loads (1 A +) they can drop as much as 1V. Diodes also exhibit something called a "dynamic resistance" which you can understand as a small value resistor in series with the diode.

As you can see on this graph, the voltage drop on the diode varies with the current passing trough it (although by a small value):


If you want some precision on your supply you shouldn't mess with the ground path, thats why Dave put the current sense resitor on the high side with a differential amplifier instead of putting it on the ground return.

[ejeffrey]

Given that the ADJ pin current is so low, you can generate the negative supply with a dedicated charge pump IC or with an oscillator (555 for example) to avoid the need of a second small transformer or a big transformer with 2 windings.

If you are going to do this, see if you can make the negative supply able to sink 5 milliamps.  That will let you meet the load current requirement of the LM317.

[RJSC]

If you are going to do this, see if you can make the negative supply able to sink 5 milliamps.  That will let you meet the load current requirement of the LM317.

Good point! We need to sink 5 mA from the LM317 for it to be stable or we will be on the same situation of Dave's circuit that needs a load connected to be able to drop to 0V (it's not able to be a reference for an high impedance circuit).

I hope Dave doesn't get mad that we are messing his whole design... 

Anyway, even using the LT3080(3) a negative supply capable of sinking the minimum load current would be neat, as it would allow the current source to keep drawing current down to 0v output into the negative supply.

[alm]

The commercial power supplies circuit using a discrete pass transistor and opamp require compensation (filters, usually low-pass RC) to eliminate the output oscillation with capacitive or inductive loads. Almost any circuit you connect to it will have capacitors, so, it is critical, and the math involved in calculating the compensation network is intimidating (systems and control theory).

No argument here, but that doesn't make the availability of the LT3080 (in some parts of the world) any better. The single-sourced part with no close replacements would also be risky in a commercial design, although I guess Linear is better than some of their competitors in delivery schedules and discontinuing parts.

[shebu18]

Search for LT1963.

[alm]

To repeat Dave, how us that better than any random adjustable three terminal regulator like the LM317? The special feature of the LT3080 is that it supplies a constant current to be used as reference voltage, which allows the voltage to go down to 0V.

[shebu18]

i found it as a replacement with near specs to lt3080. what can we use instead of the lt?

[Bored@Work]

what can we use instead of the lt?

Almost everything or nothing.

What you haven't got is that you can use whatever fulfills your needs and adapt the circuit if necessary. Dave's schematic is rather generic, isn't rocket science and can be adapted to many adjustable regulators. Including, if necessary, a negative offset to drive a non-zero Volt regulator down to 0V, or a transistor to sink more than your OpAmp can sink.

Of course, if you don't want to adapt the circuit then you may end up with no suitable replacement at all.

All this depends on your needs. Do you need output adjustment down to 0V? Down to 1.2V, to 2.5V?

[FreeThinker]

As Dave has said prior in the thread this is not the finished article, he still has a few tricks up his sleeve  and his is promising to release it as a kit so supply problems are moot. Just hang in there and wait for the final release and then decide if it is for you or needs to be modded..... Remember Santa is coming SOON!