Convert 0-30V 3A PSU to 5A or more

[audioguru2]

Hi Again,
I think your output transistors will also quickly melt if your circuits supply 5A at a low voltage. :'(

 

[narog]

Hi, the transistor/Ampops in the schematic are not the final ones. These components are there only for simulation.

The Electronics Workbench has a big library but is a little old.

The power transistor has to be symmetric so I have been search the web and I found:

NPN 2n5631 PNP 2n6031      16A 140V 200W
NPN 2n3772  PNP 2n6609      16A 140 V 150W
NPN 2n3055  PNP MJ2955    15A 90V 114W


For current sense and voltage reference generator ampops you can use tl081 if you build a regulated power supply of 15V 0V -15V, “LM7815 LM7915” this is because the opa445AP cost 8.40euros “Portugal Price”.

For the 2n2369 sorry, my mistake…..

But the opamps sensing both supply's currents have their inputs reversed. Or the opamps they are driving have their inputs reversed.

The positive current sense is well connected, but in the schematic you cant see that. The wires are very close, sorry. For the negative one I need a negative output. If you invert the inputs the ampop will not give you – X.XX V because  -31.74V – (-34V) = 2.26V and not -2.26V.

Thanks audioguru.

 

[audioguru2]

Hi Narog,
Your positive current sensor opamp has positive feedback and therefore won't work.
Your negative current sensor opamp circuit is correct. ;D

The 1k resistors in the current sensor circuits will get mighty hot. Use 10k.

The 0.47 ohm current sensor resistors will also get mighty hot and waste valuable voltage. Use 0.27 ohms or less, then re-do the current setting voltage divider.

The output transistor will dissipate 170W if its circuit has a low (shorted?) output voltage at 5A. With a huge heatsink a power transistor might be able to dissipate only 60W. Therefore at least 3 paralleled output transistors are needed like our proposed circuit.
When the output transistors are paralleled, their individual current is lower, therefore their gain is higher and the driver transistor won't need to supply them with as much current and will be cooler. ;D

View attachment 37558

 

[BigPoppa]

Hello.
I am building this project and have some questions.
-i have 26VAC/150VA transformer?Is this ok?I think i will never use the psu at it`s limits...
-can i change TIP31A with BD241C?
-my R7 is 0,47E/9W.Is this ok?
-C1 will be 10,000uF/50V
-op-s will be OPA445
-bridge rectifier is 35A,type that fixes on the heatsink
-should i use 2 or 3 2N3055`s

Is there anything else that i should be aware of?

Thanks in advance,BigPoppa

 

[audioguru2]

Hello.
I am building this project and have some questions.

Hi BigPoppa,
Welcome to our forum. ;D
This thread is about modifying the original 0-30V/3A project to 5A. Your little transformer won't power this supply to 30V nor to 5A.

-i have 26VAC/150VA transformer?Is this ok?

The project's voltage control's range should be changed to match your transformer since it will allow only about 27V but the control will try to force it higher and its output will become full of ripple and poor regulation.

-can i change TIP31A with BD241C?

Yes, any BD241 is the same or better than a TIP31A.

-my R7 is 0,47E/9W.Is this ok?

It will allow your project to supply 3A like the original modified project will do. Its max dissipation is 4.2W so it won't get too hot.

-C1 will be 10,000uF/50V
-op-s will be OPA445
-bridge rectifier is 35A,type that fixes on the heatsink

Excellent for all.

-should i use 2 or 3 2N3055`s

Use two with emitter resistors. A fairly large finned heatsink should be used.

Is there anything else that i should be aware of?

Use BZX79C5V6 low-current zener diodes and change R4 to 1k.
We suggested increasing the power ratings for many resistors in the other thread.

 

[BigPoppa]

The project's voltage control's range should be changed to match your transformer since it will allow only about 27V but the control will try to force it higher and its output will become full of ripple and poor regulation.

If it will be stable to about 24 volts,than i am happy. ;D I almost never need voltages higher than this,so it is not important to me.

I am using this partlist:

Parts List for 3A Variable Power Supply

*T1 = 30VAC, 4.3A, 130VA stepdown power transformer
*R1 = 2,2 KOhm 2W 
R2 = 82 Ohm 1/4W
R3 = 220 Ohm 1/4W
R4 = 4,7 KOhm 1/4W
R5, R6, R13, R20, R21 = 10 KOhm 1/4W
*R7 = 0,47 Ohm 10W
R8, R11 = 27 KOhm 1/4W
R9, R19 = 2,2 KOhm 1/4W
R10 = 270 KOhm 1/4W
R12, R18 = 56 KOhm 1/4W
*R14 = 1,2KOhm 1/4W
*R15 = 100 Ohm 1/4W
R16 = 1 KOhm 1/4W
R17 = 33 Ohm 1/4W
*R22 = 3,9 KOhm 1W
RV1 = 100K trimmer
P1, P2 = 10KOhm  linear potentiometer
*C1 = 10,000uF/50V electrolytic
C2, C3 = 47uF/50V electrolytic
C4 = 100nF polyester
*C5 = 220nF polyester
C6 = 100pF ceramic
C7 = 10uF/50V electrolytic
C8 = 330pF ceramic
C9 = 100pF ceramic
*D1, D2, D3, D4 = 6-10A/50V diode bridge module
D5, D6 = 1N4148
D7, D8 = 5,6V Zener
D9, D10 = 1N4148
D11 = 1N4001 diode 1A
*Q1 = TIP31A, NPN transistor, no heatsink
*Q2 = TIP31A NPN transistor
Q3 = BC557, PNP transistor or BC327
*Q4 = Two paralleled 2N3055 NPN power transistors with 0.1 ohm, 1W emitter resistors each
U1, U2, U3 = OPA445AP, operational amplifier
D12 = LED

"*" is a change from the original design.

Do i need to change something,since my transformer output voltage is to low?Or is this partlist ok for stable 0-24V and about 2-3 amperes?

Thanks for the zener suggestion,i will change it if i can find some around here. 

Thank you very much for your answers. 8)

 

[audioguru2]

Hi BigPoppa,
I mentioned the zeners and R4 that feeds one, but someone said that R2 gets very hot. Here is the latest parts list:

View attachment 37751

 

[BigPoppa]

Just one more question...What`s the function of R2 and C2?

 

[audioguru2]

What`s the function of R2 and C2?

They are part of the half-wave rectified negative supply.
C2 charges through D5 when the tranformer connection #2 swings positive. R2 limits its charging current so a small capacitor and diode can be used.
C2 discharges through R2, D6, R3 and D7 to make the regulated -5.6V negative supply for some of the opamps, when the transformer connection #2 swings close to ground.
The negative supply is one reason why the opamps must have a very high supply voltage rating.

 

[Kain]

Ok, finally some late update on this project from me. As you already know I assembled the PSU and had some problems with the maximum output voltage. I finally found the problem - there was an unetched copper on the board between 2 traces that looked so good that one can never tell it's not supposed to be there (looked like part of the layout). I could have seen it earlier, but on dual sided boards with the parts already soldered is not exactly easy to look at. Anyways, I fixed this and now the PSU goes all the way up to 34.5V when unloaded. I will be doing ripple testing this week so the results will be here soon. What I tried so far was testing for max output current. With the most recent setup which Audioguru posted the maximum output current is arround 6.85A, and this is at any voltage BUT lower than 2.5. At lower than 2.5V output the current can't reach high values. Audioguru, do you care to explain why?  ;D

 

[Kain]

I just did something simple - I connected the PSU across a huge rheostat set to 6 Ohm and set the PSU to max V and A. First thing to notice - I figured I can cook on the rheostat...

 

[audioguru2]

Hi Kain,
Be careful with your transformer because it is rated for only (!) 210VA. Its peak current is normally about 50A into the rectifier bridge to replenish the charge lost in the main filter cap during high loading but its average power rating is still 210W. The rectifier conducts only near the transformer's voltage peaks of 42.4V, so the max average DC current from your supply is limited by the transformer's power rating to 210/42.4= 4.95A.

I think you are over here with a 60Hz mains. Your 20,000uF for C1 will discharge about 2.8V each half-cycle with a 5A load. The 2.8V ripple should be reduced by the opamps to only about 0.5mV of ripple at the project's output but yours is 800mV when the output transistors aren't saturated. I think that using a proper "star grounding" technique will solve your ripple problem:
1) The 0V wire of R4, R6, U1, the voltage adjust pot P1 and R11 must be connected directly to the output without any of the current that goes into R7. Therefore R7 should have a separate wire connecting it to the output 0V terminal. Voltage regulation will be improved.
2) C1 and the rectifier bridge should be connected together with very heavy short wires to carry 50A current peaks.
3) The positive voltage feed to the circuit must connect directly to C1, not to a terminal strip.
None of the 50A current peaks must reach the circuit.

With an output voltage less than 2.5V, I aren't sure why it doesn't give max current. Check that its negative 5.6V supply is OK.
I couldn't find a schematic of an OPA445 opamp, but its datasheet shows a different connection to its offset voltage adjustment from the one on U2 in our circuit. Try connecting R10 to negative 5.6V instead of the existing output 0V. Then try removing R10 and connect the slider of RV1 to the negative 5.6V. ;D

View attachment 37822

 

[BennoRausch]

Hi everybody!

Audioguru, your're doing an excellent job! Thank you very, very much for your commitment!
I'd also like to built this PSU and understand how it works. As far as I can see there were lots
of changes by the component values but the circuit's topology stayed the same except the
additional resistor Rl. I've read the function description, but I still don't understand half
of this circuit I think.Now here are my questions:

1.This additional resistor Rl forms a load that is needed for correct function of the regulation even if no external load is connected?

2.What is R1 good for? Why is the smoothening-capacitor C1 not enough?

3.Why 15W for R7? With P=R*I^2 and a max. current of 5 amps follows a max. dissipation of 6,75W. Is this because a 15W resistor doesnt get as hot as a 9W resistor? Does this resistor need an extra heat-sink? The lower the value of R7 the better, because less power is dissipated in R7?

4.U1 provides a constant voltage which is amplified by U2. The output voltage is fed back to U2 by R12 (C6 is for stability?). C9 is for stability, but what are R15 (base-resistor for R13?)
and D10 good for?

5. I don't know what D11 and C7 are doing. Same with D10

I hope somebody (I think it will be audioguru ;-)) can help.

Bye
Benno

 

[audioguru2]

Hi Benno,
Welcome to our forum. ;D

1.This additional resistor R1 forms a load that is needed for correct function of the regulation even if no external load is connected?

2.What is R1 good for? Why is the smoothening-capacitor C1 not enough?

R1 isn't additional, the project has had it from its beginning. It simply discharges C1 slowly when the mains is turned off.

3.Why 15W for R7? With P=R*I^2 and a max. current of 5 amps follows a max. dissipation of 6,75W. Is this because a 15W resistor doesnt get as hot as a 9W resistor? Does this resistor need an extra heat-sink? The lower the value of R7 the better, because less power is dissipated in R7?

The circuit is calculated to have about 1.41V across R7 at its max current, 3A uses 0.47 ohms and 5A uses the nearest standard value of 0.27 ohms. People have complained that its high temperature caused their pcb to char. Power resistors operate extremely hot (250 degrees C?) at their rating. Resistors with a higher power rating are larger and are better able to dissipate heat. Therefore at power dissipation levels below their max rating they operate cooler. I've used little power resistors in an aluminum case that are easily cooled by bolting them to an aluminum project box.

4.U1 provides a constant voltage which is amplified by U2. The output voltage is fed back to U2 by R12 (C6 is for stability?). C9 is for stability, but what are R15 (base-resistor for R13?)
and D10 good for?

Correct. C6 and C9 provide lead and lag compensation respectively since the old 2N3055 output transistor is much slower than the opamp.
R15 might not be needed. Its original function was possibly to isolate the output of Q2 from the capacitive load of Q2, to ensure that U2 doesn't oscillate. Its original value of 1k wasted a lot of voltage so I reduced it to 100 ohms that still effectively isolates the capacitance of Q2.
D10 discharges C7 and any other capacitor on the output by Q1 when the mains is turned off.

5. I don't know what D11 and C7 are doing.

D11 conducts to discharge a reverse-charged capacitor that might be connected across the output, I guess. Maybe a bigger one would be better.
C7 is also for compensation and high frequency noise noise reduction. It is also a high frequency low impedance bypass.

I hope somebody (I think it will be audioguru ;-)) can help.

How did you guess? I've been waiting for this site to come back online all day! ;D
"Server not found" :'(

 

[BennoRausch]

typing error:I meant Rl at the output.


Where do you know this from? Designers determination?


But without R15 there would be no current-limiting, when the load is discharged through D10 and Q1?

- Does C5 shorten high-frequency disturbances of the mains?

- Q2 is a emitter-follower, correct? Is this the reason why R16 is needed? And why emitter-follower instead of a common emitter configuration?

- The 3 parallel output-resistors: Why do they have resistors at their emitters?


I know, had to wait to ask my questions.

Bye
Benno

 

[BennoRausch]

Hi Audioguru!

First off all: thanks for your fast reply. I didn't want to be impolite but I messed up my thread above because I didn't know how to quote properly. I hope it will be better next time!

So long
Benno

 

[audioguru2]

typing error:I meant Rl at the output.

I don't know Rl.

Where do you know this from? Designers determination?

Ohm's Law.

Does C5 shorten high-frequency disturbances of the mains?

C5 is the circuit's high frequency bypass cap for stability from oscillating. C1 is big and has inductance which reduces its effectiveness at high frequencies.

Q2 is a emitter-follower, correct? Is this the reason why R16 is needed? And why emitter-follower instead of a common emitter configuration?

No. Q4 gets hot which causes it to leak current. R16 bypasses its leakage current.

The 3 parallel output-resistors: Why do they have resistors at their emitters?

The current gain of transistors having the same part number and manufacturer are different. Without the emitter resistors then the transistor with the most gain will carry the entire load until it fails. The emitter resistors reduce the difference between individual transistors.
Good point! ;D Somebody reported that even with 0.1 ohm emitter resistors one transistor got much hotter than the others. Therefore the emitter resistors should be 0.33 ohm, 2W instead.

 

[Idotry]

But at the moment i dont have the money to buy the heatsink, its gonna be a real fat one, and it is very expensiive.

//Staigen

Im no expert in anything, but would an old CPU heatsink and fan be up to the task of cooling without too much expense?  :-\

 

[audioguru2]

A CPU heatsink is much too small to dissipate 121W.

 

[someshwar_ms]

thank u very much