0-30V Stabilized Power Supply

[redwire]

denci,  

This is a picture of a MJ11016 (upper left hand corner of picture) I was using to test my new smd board
http://www.electronics-lab.com/forum/index.php?action=dlattach;topic=7317.0;attach=19559
At this time I was using a brake light bulb  (about 2.5 A).   The previous 5A test was done on a cold day in my garage and the heat sink was not in a project case.   These are not the worst case conditions.   When you package things up and contain the heat you are going to need at least 2-3 transistors. large heatsinks and a small fan if you want  5A and rock solid durability. Then your output will likely be determined by what size transformer you can afford.    Digikey and Mouser have a nice 6.25 A, 28 V transformer by Triad for about $32.
Note that the worst conditions are at low voltage and high current (or a short).   When the voltage was at 1 V using a 0.27 Ohm, 10W resistor just like the one on the board, you could feel the heat pouring off the MJ11016.  The same happened when I set the current for 2.5A and shorted the output.  I didn't hold the shorted conditions for an extended period because  heatsink was getting quite.  The bd139 without a heat sink was not hot.  

 

[sonycman]

Hello, audioguru!
Trying to adapt your schematics for microcontroller control and indication.
Thinking about changing 11.2 volt reference voltage to 5.0 volt and using digital potentiometers.

Which purpose C5 serves for? Should I connect it phisycally close to Q2, as on the picture, or better close to U2?
Can it be ceramic type, or better polyester?

 

[audioguru2]

Thinking about changing 11.2 volt reference voltage to 5.0 volt and using digital potentiometers.

The 5.6V zener diode has very good voltage regulation (lower and higher voltage zener diodes are worse) and has zero voltage change when the temperature changes unlike other zener diode voltages. The opamp U1 that feeds this zener diode is a current source for even better voltage regulation. Another member said to use a 78L05 regulator instead.

Which purpose C5 serves for? Should I connect it phisycally close to Q2, as on the picture, or better close to U2? Can it be ceramic type, or better polyester?

C5 is a high frequency filter. Mount it close to opamp U2. It should be a polyester film type.

 

[sonycman]

zTher 5.6V zener diode has very good voltage regulation (lower and higher voltage zener diodes are worse0 and has zero voltage change when the temperature changes unlike other zener diode voltages. The opamp U1 that feeds this zener diode is a current source for even better voltage regulation.
another member said to use a 78L05 regulator instead.

What about TL431 precise voltage reference? Looks even simpler than opamp and its discrete elements.
Though TL431 might need buffering too...

C5 is a high frequency filter. Mount it close to opamp U2. It should be a polyester film type.

So it is shunt for U2. Okay, understood!

Thanks for help!
Sorry for bad english...  :

 

[audioguru2]

Yes, you can use a TL431 voltage reference instead.

C5 is a high frequency filter for the supply voltage for the output amplifier that helps to prevent oscillations at high frequencies.

 

[redwire]

sonycman,  U3 requires a voltage between 0 and 1.5V (for 3A output) and U2 requires a voltage between 0 and 11.2V to adjust the output voltage.  If you are using a microcontroller, couldn't  you delete U1, and U3 and feed a DAC signal to U2 of 0-5 V to control voltage (Yes you will need to adjust the gain for U2 by changing R12 to about  70K and R1 to about 10K so you get a gain of about 6 so that the 0-4.5V signal will provide the same amplification as the original project).  The microcontroller would also control the current by measuring  the voltage drop across R7 and  replace  U3.  If the voltage drop across R7 was greater than your set voltage, then the microcontroller could automatically drop the voltage to U2 until it matched or was lower than the set voltage.  You could also incorporate a temperature sensor that could also reduce current if a certain value is achieved.    Make sure you pick a processer that has enough pins to support these features and a LCD display.  Not sure if a digital potentiometer is necessary. 

 

[sonycman]

redwire,
Yes, using DAC instead digital potentiometer is another way to go, will see about it.
It must be the DAC with the same precision and external reference voltage as digital POT.

As to remove U3 and forcing to limit current in software by MCU - did`not agreed in that.
Current limiting must be done in hardware only by high speed opamps for reliability reasons.
MCU work is to set up references for opamps and to read voltage and current values via ADC for displaying it.

Thanks for helping!

 

[redwire]

sonycman,    Good luck.    Did you check out http://www.electronics-lab.com/projects/test/007/index.html ?  It will give you a starting point for a MCU output display.  The hex file provided  is for a Atmega8 chip so don't purchase a Atmega88 chip. 

 

[sonycman]

redwire,
thanks again! I have some experience with MCUs, they are no problem 

 

[electroguy]

Unfortunately, both the op-amps MC34071 and TLE2141 is not available in my country but there is MC34074 available.
Are you using the single op-amps or the duals ?
Is it possible to replace the three MC34071s with one MC34074 ? but the problem remains in U2 for the pins 1 & 5

 

[audioguru2]

Unfortunately, both the op-amps MC34071 and TLE2141 is not available in my country but there is MC34074 available.
Are you using the single op-amps or the duals ?
Is it possible to replace the three MC34071s with one MC34074 ? but the problem remains in U2 for the pins 1 & 5

An MC34074 will probably get too hot.
You do not say what is your country. Go to www.farnell.com and click on the flag of your country. They have warehouses all over the world called Element14.

 

[electroguy]

I am from Bangladesh,it is a country in South Asia.
Unfortunately, again,  I don't see the name of my country in Element14.
If I use MC34074 with a big heat sink and a computer fan and did you use the offset null for U2 ?
On the other hand, MC34074 do not have any offset null.
Save me!!!

 

[audioguru2]

I do not know why you cannot buy ordinary parts in your country.

An MC34071 quad opamp is in an epoxy package that cannot have a heatsink. You can glue on a heatsink but the epoxy will not conduct heat to it well.

The input offset adjustment on opamp U2 sets the output to exactly 0V when the voltage-setting pot is at zero.
Without adjusting the offset then the output could be plus or minus 20mV.

 

[electroguy]

Should I try this old modified one proposed by Ivan on 02/2010.
Oh, I forget to mention that I already bought the old parts last year

 

[audioguru2]

Should I try this old modified one proposed by Ivan on 02/2010.
Oh, I forget to mention that I already bought the old parts last year

The circuit you found uses a 7812 voltage regulator so that the original opamps do not have a supply voltage too high. Then opamp U1 will not work unless its circuit is changed to reduce its output voltage.
The maximum output from opamp U2 is about +9.5V and the maximum output voltage from the project is about +6V. If the original 24VAC transformer is used then Q2 will get extremely hot and might fail and the output transistors will get very hot.

 

[redwire]

Unfortunately, both the op-amps MC34071 and TLE2141 is not available in my country but there is MC34074 available.
Are you using the single op-amps or the duals ?
Is it possible to replace the three MC34071s with one MC34074 ? but the problem remains in U2 for the pins 1 & 5

Audioguru wrote:  An MC34074 will probably get too hot.

U1 and U3 need to provide very little output current.  U2 is the real driver in this circuit.  If electroguy utilized a darlington power transistor instead of the 2N3055 then the BD 139, and subsequently the MC3074 would need very little current to drive the BD139.  I think the quad opamp could handle this load.  I would suggest the MJ11016.  Yes, you don't have the input offset pins, so setting the voltage to exactly zero may be impossible but a few millivolts should be a problem?  I think this would be more preferable than building the old circuit.

 

[electroguy]

Instead of redesigning the circuit for the quad one, what if I just place all the pins: 2;3;6 of U1,U2,U3 at the input and output of the quad one. will it work ?
The design is shown below.
The three red/ blue/ green dots represents the input and output of the quad one.
As I also bought the old parts, is it going to be wasted?
The transformer is 24V 3A, if I buy MC34074 and use it for 24V, less than 30V will not be a big problem for me.
@redwire: a few millivolts is not a problem for me

 

[denci]

I finished with working on power supply and made a test yesterday but just one think is not working properly, i can adjust the current only up to 3A then current blocking turning on.
I use elements for 5A project (Redwire) as i mentoined couple of times above but i have not idea where is the problem, i hope that i didn't overlook some element value, i will make a review of all elements in circuit once again.

 

[MAntunes]

Hello!
Would a 24VAC transformer be good for this power supply?
I am building the 3A one.
Regards

 

[deepak.roy.alchemist]

Hello! MAntunes
then you won't get 30V DC @ 3A at the output.