Op Amp - ramp generator problem

[abuhafss]

Hi

I have been using this circuit to test Automotive Voltage Regulators (AVR).

When the start button is switched on, the upper IC LM358 generates a ramp voltage from 5V to 31V. This ramp voltage is fed into the AVR which trips at a preset voltage. Hence the output from the AVR is also a ramp from (5V - 0.70V) to the preset value. As soon as the output reaches the preset value, it drops to about 750mV. This causes the output of the lower IC LM741 to go high, which turns off the PNP transistor thereby, disconnecting the 3V supply to the ramp generating circuit hence, the ramp is paused. That paused value of ramp voltage minus 0.70V is the tripping voltage of the AVR.

The circuit has been performing perfectly for 12V AVRs. But when a 24V AVR is connected, the ramp output would stop at about 20V though the transistor is still on.

I have checked the ramp generator, it is working perfectly ----- 5 to 31V. I checked the AVR by applying variable voltage (manually) at the input, it is working perfectly. The circuit of a 24V AVR is identical to that of 12V AVR, except that 470 Ohms resistor is replaced with 220 Ohms and 12V zener is replaced with a 24V zener.

Why the ramp would stop at 20V?

 

[Harald Kapp]

The datasheet states on page 5: Voh=26V for V+=30V, Rl=2kOhm.
Although your circuit is not identical to the test conditions for the datasheet, this suggests that the LM358 can deliver at most V+ - 4V at its output. If you power it from 24V -> Voutmax=20V.

Try the circuit with a 30V supply.

 

[abuhafss]

I have tried with 30V and 32V, it's not working.

Shall I replace LM358 with some other op-amp? Which one do you recommend, please?

 

[Harald Kapp]

If it's not working at 30V, I doubt the OpAmp is the culprit.

Another possibility:
Look at the 741. The max. input voltage is +-15V or positive supply, whichever is smaller (datasheet page 2). Once the input voltage rises above that value, the protection circuit (diodes) will become active, drawing current through the input pin. This currrent is supplied by the LM358 and is therefore no longer available for charging the feedback capacitor.
reduce the output of the AVR by a voltage divider (e.g. 1:2). Change the reference divider at the 741's "+" input to 10k/470Ohm (for 1/2 reference voltage, too).

 

[abuhafss]

The op-amp is not dude, I have checked 6 of them......all have same result.

I have halved the output from the AVR and the reference voltage of LM741. But no improvements.

One thing I should mention here, the tripping voltage for a 12V AVR is 14.4V and for the 24V AVR it is 28.8V. The LM741 works perfectly for input 14.4V which is higher than the supply +12V.

You can see the behavior in LTSpice. Change the extension txt to asc. I don't have spice models for LM741 and LM358 so I have been using LT1490A.

 

[Harald Kapp]

Sorry, I can't simulate this. I do have LTSPICE, but I don't have the LTSPICE models for TIP127, !N4007, BZX384B24, BC557.
While I could find them on the internet and put them into the simulation, you've obviously already donde so. can you supply the necessary libraries/symbols? Or modify the circuit so it uses only LTSPICE standard models?

One more test to isolate the problem: How does the ramp generator behave when the AVR is connected but not the 741? What does the output of the AVR look in that case? This is just to exclude any influence the 741 may have on the ramp circuit.

 

[abuhafss]

1N4007, BZX38424, BC557 are included in LTSpice standard library, otherwise you can use any diode, any 24 zener and any small signal PNP transistor like 2N3906 etc. Also for TIP127 you can use 2N3906+TIP32 or TIP42. Anyway for your future use here are necessary files for TIP127 and TIP121. You might also find the following links useful:

http://ltwiki.org/?title=Components_Library_and_Circuits
https://groups.yahoo.com/neo/groups/LTspice/info

To disconnect 741, I have two options

1) to disconnect the 741's (-)ve input
2) to disconnect power supply to 741

In either case, same old result. So, 741 has no affects on the ramp.

 

[Arouse1973]

The zener is not a perfect switch and will start to conduct before it fully avalanches say . This could be enough current to turn on the first transistor and put supply onto the 1K. this current could be over 20mA which is the max the chip can handle but the chip could have a current rating as low as 10mA. I think you are overloading the LM358. put an emitter follower after the chip to try it.
Adam

 

[Harald Kapp]

Adam is on the right track. The simulation shows that the zener becomes conducting at a ramp voltage of ~14V.
The zener voltage (24V) is given for a current of 5mA - the transistors have long before become conducting.
The diode will start to conduct at ~0.-6*Vf , see datasheet figure 2. That is where the 14.4V come from.
For much bettter accuracy use e.g. an LM431. You will need to adjust your circuit similar to figure 19 in the datasheet.

 

[abuhafss]

Adam

The 12V AVR draws 16mA and the 24V AVR draws 25mA.

If the emitter-follower is introduced between LM358 and the AVR then both (12V and 24V AVR) would draw 16mA but, the ramp would not impove. When 24V-AVR is connected, the ramp would stop much earlier at 16V.

 

[abuhafss]

Adam is on the right track. The simulation shows that the zener becomes conducting at a ramp voltage of ~14V.
The zener voltage (24V) is given for a current of 5mA - the transistors have long before become conducting.
The diode will start to conduct at ~0.-6*Vf , see datasheet figure 2. That is where the 14.4V come from.
For much bettter accuracy use e.g. an LM431. You will need to adjust your circuit similar to figure 19 in the datasheet.

The AVRs are built with zeners for decades. All of them have similar circuit with or without slight modifications. If the zener is culprit, why would it work when variable power supply is connected manually?

 

[Harald Kapp]

Becasue the variable power supply can supply more current than the OpAmp?

 

[abuhafss]

In that case, the best solution would be to change the op-amp which could provide more current. So, which one is recommended?

 

[Harald Kapp]

Google "power operational amplifier" to get a list of possible OpAmps.

 

[abuhafss]

I think you are overloading the LM358. put an emitter follower after the chip to try it.
Adam

Earlier I introduced the NPN transistor the other way i.e. collector to the output of op-amp and emitter to the AVR.

Now, I connected the collector to 32V supply, emitter to the AVR and the base to the output of the op-amp thru a 1k resistor. It is working perfectly, now.

Thanks both of you, for your time.

 

[Arouse1973]

Great News
Thanks for your support Harald, what a nice end to the week both helping someone we have never met and working together to help, makes me smile

Adam

 

[Harald Kapp]

Glad it works.
Have a nice weekend, too.