Hello all,

Before I start I have to say I am an amateur enthusiast, I have an avid interest in electronics for many years but alas I do not have any formal training, but very willing to take advice, please be forgiving.
That out of the way, I have been studying op-amp circuits for a bit, with intentions to build a fairly accurate bench power supply using a non-inverting op-amp circuit. I have achieved this with limited success using a TLE2041CP op-amp. I have soldered it on a board with the power transistors on a large heatsink, also used a beefy transformer.
The problem is a voltage drop on load, 30mV on a 1.75amp draw and 60mV on a 4amp draw.
I have a stable reference at pin 3 of 4.41v on/off load.
Voltage on load at both inputs of op-amp are exactly the same 4.41v, the op-amp seems to be doing its job?.
Why the drop?, am I expecting too much accuracy from this circuit?, any help much appreciated.

 

The problem is a voltage drop on load, 30mV on a 1.75amp draw and 60mV on a 4amp draw.

Sounds like a drop from the internal wire resistances.
Where exactly are you measuring this voltage?

For best load regulation the two 5k adjustment pot connections should be directly at the output plus and minus terminals.
That will eliminate the effect of the internal wire resistances.

 

Sounds like a drop from the internal wire resistances.
Where exactly are you measuring this voltage?

For best load regulation the two 5k adjustment pot connections should be directly at the output plus and minus terminals.
That will eliminate the effect of the internal wire resistances.

Thanks for your advice.
I have constructed this circuit on veroboard and used it into an old broken ham power supply.
I am measuring the voltage output at two large screw terminals at the front .
One 5k pot (first one nearest the output balast resistors) is at the front and used to vary the output voltage from 4 to 16v, the second one is actually on the board.
Do you think moving the second pot off the board closer to the output terminals will help?

 

Heavy duty power supplies (PSU) designed to deliver large currents use thick cables for the supply and return lines connected to the load. Two separate sense wires are connected at the load, not at the output of the PSU, to compensate for the voltage drop along the supply cables.

 

Is it ok to parallel pass transistors like this? I was wondering about one of them going into thermal runaway.

 

Heavy duty power supplies (PSU) designed to deliver large currents use thick cables for the supply and return lines connected to the load. Two separate sense wires are connected at the load, not at the output of the PSU, to compensate for the voltage drop along the supply cables.

Thanks for your interest and help, mr chips.
I have kept the wires used going to the power transistors and from the power resistors both short and thick and short as possible and used a car bulb for testing quite close to each output post.
The only wire that is thin is the wire going from the op-amp output to the tip41?

 

Thanks for your interest and help, mr chips.
I have kept the wires used going to the power transistors and from the power resistors both short and thick and short as possible and used a car bulb for testing quite close to each output post.
The only wire that is thin is the wire going from the op-amp output to the tip41?

I have a fan which is thermostatic controlled which keep the power transistors at 40 deg cent. on a 4amp load

 

You ave exceed the common mode input reference range of the OP-amp. It's about 3V. Your at 4.41. Who knows what it''s going to do.

The HP DC power supply handbook https://archive.org/details/DC_Power_Supply_Handbook_Agilent_Technologies_Application_Note_90B is a book you have to read.

As said, high current supplies have sense inputs. In a lot of casses if left open the circuit regulates - 2 diode drops.

 

You ave exceed the common mode input reference range of the OP-amp. It's about 3V. Your at 4.41. Who knows what it''s going to do.

The HP DC power supply handbook https://archive.org/details/DC_Power_Supply_Handbook_Agilent_Technologies_Application_Note_90B is a book you have to read.

As said, high current supplies have sense inputs. In a lot of casses if left open the circuit regulates - 2 diode drops.

ok a lot of advice in a short amount of time, thanks everybody, I have been kinda grasping in the dark with this for quite a while, changing values at each point, changing power resistor values and also adding an extra transistor before tip41a.
If I understand you correctly the reference voltage is set too high for starters?.
I will read the linked handbook.

Just to ad, I measured the value of the second pot closest to the chip on the board and got a value of 2K, I left the pot disconnected and instead soldered a resistor from the other pot to the ground post to the first pot ( sorry should have labelled them in picture) but still got same results.

 

Is it ok to parallel pass transistors like this? I was wondering about one of them going into thermal runaway.

Normally transistors configured common collector are not subject to thermal runaway.

 

Isn't the supply +23V? Then the minimum input common mode voltage limit is 21V which is well above the opamp's input voltage.

Why is there an emitter-follower transistor and some other parts in between the zener diode and the opamp input? The transistor causes the output voltage to rise when it gets warm and drop when it gets cold. To reduce the 5V from the zener diode to 4.4V then use two resistors as a voltage divider instead.

The output transistors need a resistor to ground at their emitters to cancel their leakage current when they are hot.

 

Isn't the supply +23V? Then the minimum input common mode voltage limit is 21V which is well above the opamp's input voltage.

Why is there an emitter-follower transistor and some other parts in between the zener diode and the opamp input? The transistor causes the output voltage to rise when it gets warm and drop when it gets cold. To reduce the 5V from the zener diode to 4.4V then use two resistors as a voltage divider instead.

The output transistors need a resistor to ground at their emitters to cancel their leakage current when they are hot.

Sorry I'm getting lost "minimum input common mode voltage limit is 21V", my supply is too high yes?

I tried using the zener directly to op-amp input but its stability suffered under load, I added the transistor to achieve some stability

Could you recommend a value for the leakage resistor at the power transistor

 

My guess is that the feedback network isn't connected to exactly the correct point, and that the 30mV is voltage drop due to the output current.
The two presets should be connected to the output and ground, so that there can be no voltage drop between the feedback connections and the output, and between the 0V feedback connection and the real ground.

 

My guess is that the feedback network isn't connected to exactly the correct point, and that the 30mV is voltage drop due to the output current.
The two presets should be connected to the output and ground, so that there can be no voltage drop between the feedback connections and the output, and between the 0V feedback connection and the real ground.

I am running this op-amp on a single rail + and - no ground, could that be part of the problem?
The return pot closest to chip has been moved off the board to connect with neg output post.

 

The return pot closest to chip has been moved off the board to connect with neg output post.

In which case the negative end of your reference must go to the same point.

 

Sorry I'm getting lost "minimum input common mode voltage limit is 21V", my supply is too high yes?

Yes your input is too high which makes your output transistors VERY hot. for the output to ne 14.8V at 4A then the output transistor bases must be at least 16.2V and the base of the TIP41 must be 17V. The opamp needs a supply that is at least 18.3v. So your 23V is cause extra heating of (23V - 18.3V) x 4A= 18.8W which is a lot of heat. Use a 14V or 15V supply instead.

I tried using the zener directly to op-amp input but its stability suffered under load, I added the transistor to achieve some stability

The datasheet says the zener is stable. The opamp and output transistors are unstable. Add a compensation capacitor parallel with the feedback variable resistor, try 100pF and another 100pF between pin 1 and pin 2 of the opamp.

Could you recommend a value for the leakage resistor at the power transistor

A 4.7k/1/4W resistor will be fine.

 

Yes your input is too high which makes your output transistors VERY hot. for the output to ne 14.8V at 4A then the output transistor bases must be at least 16.2V and the base of the TIP41 must be 17V. The opamp needs a supply that is at least 18.3v. So your 23V is cause extra heating of (23V - 18.3V) x 4A= 18.8W which is a lot of heat. Use a 14V or 15V supply instead.


The datasheet says the zener is stable. The opamp and output transistors are unstable. Add a compensation capacitor parallel with the feedback variable resistor, try 100pF and another 100pF between pin 1 and pin 2 of the opamp.


A 4.7k/1/4W resistor will be fine.

Thank you, I will try your suggestions, just one point, I am using a single rail supply of 23v,

 

The TLE2141 opamp has inputs that work all the way down to 0V in your circuit so it does not need a negative supply.

 

I’ve simulated in spice and the output voltage sags with increased loads.

Voltage is not absolute, the supply Voltage to the opamp is fine, it specifies +/-22v which means you can supply single supply up to 44v. -22 to +22 of dual supply is exactly the same as a single supply 44v.

 

I’ve simulated in spice and the output voltage sags with increased loads.

Voltage is not absolute, the supply Voltage to the opamp is fine, it specifies +/-22v which means you can supply single supply up to 44v. -22 to +22 of dual supply is exactly the same as a single supply 44v.

The single polarity 14.8V power supply circuit does not need an input of 44V (it will make too much heat) and it does not need a negative supply.