EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: bogdant on May 26, 2020, 11:51:45 am
-
I am tring to setup a current sense high side with LF411. But in datasheet does not specify if could be used in single supply configuration.
https://www.ti.com/lit/ds/slos011c/slos011c.pdf?ts=1590493120118 (https://www.ti.com/lit/ds/slos011c/slos011c.pdf?ts=1590493120118)
What do you think, is it possible to be used power from 0-32V? Do I have a chance to build a differential amplifier with this opamp?
The resistor sense will be 0.3ohm, and the factor will be 3. I intend to setup the resistor 100k and 330k.
-
I am tring to setup a current sense high side with LF411. But in datasheet does not specify if could be used in single supply configuration.
https://www.ti.com/lit/ds/slos011c/slos011c.pdf?ts=1590493120118 (https://www.ti.com/lit/ds/slos011c/slos011c.pdf?ts=1590493120118)
What do you think, is it possible to be used power from 0-32V? Do I have a chance to build a differential amplifier with this opamp?
The resistor sense will be 0.3ohm, and the factor will be 3. I intend to setup the resistor 100k and 330k.
I think it will work except for when the PSU's output voltage is low. The way I interpret the data sheet, the input common mode range doesn't include either supply rail voltage.
JFET opamps usually work with the inputs at one rail or the other depending the polarity of input JFET used.
The output needs to be able to go to 0V also if no negative rail is used.
-
Even if an opamp has the ability to swing its output down to 0V, it may not actually be able to completely do so because of the pull-up current from the input resistors.
-
An LM358 would be a reasonable starting point for experimenting.
A very simple current sink could be added.
I did a quick experiment with an upside down BC548. With 100µA Base current, the Emitter to Collector voltage drop was 10mV at 100µA.
It started coming out of saturation at about 250µA.
-
yesterday I tried with LF411, but the output is always 2V with no load, connecting load does not do anything. If adding load does not do anything. 100k /330k and 0.3ohm sense resitor, there is a link here with the schematic.
The reason why I do this because the I have a power supply here:
https://github.com/bogdan-tirzioru/Powersupply/blob/master/factory/sampleB/Job2.PDF
Implemented with lm324, i think it is similar to lm358, but with the supply voltage of 32V with no load I have 0.20V on the output of the diferential amplifier. Connecting the next stage the voltage goes to 0.5V with no load. :-//
I did try yestarday also the OPA134 and LT1490 but had the same results as lm324:
I = 0A -> Umeas=0.21V (no load)
I = 0.27A -> Umeas = 0.43V (75 ohm load)
I = 0.72A -> Umeas = 0.92V (35 ohm load)
I = 2.2A -> Umeas = 2.36V (10 ohm, voltage drops to 24V)
This with the output of differential not connected to any stage, only to my multimeter.
The output offset is to large.
A previous thread suggested to some pulldown to a negative rail, but it is not available on this project so I can not tried. Any idea what to change to make it work ?
-
yesterday I tried with LF411, but the output is always 2V with no load, connecting load does not do anything. If adding load does not do anything. 100k /330k and 0.3ohm sense resitor, there is a link here with the schematic.
The reason why I do this because the I have a power supply here:
https://github.com/bogdan-tirzioru/Powersupply/blob/master/factory/sampleB/Job2.PDF
Implemented with lm324, i think it is similar to lm358, but with the supply voltage of 32V with no load I have 0.20V on the output of the diferential amplifier. Connecting the next stage the voltage goes to 0.5V with no load. :-//
I did try yestarday also the OPA134 and LT1490 but had the same results as lm324:
I = 0A -> Umeas=0.21V (no load)
I = 0.27A -> Umeas = 0.43V (75 ohm load)
I = 0.72A -> Umeas = 0.92V (35 ohm load)
I = 2.2A -> Umeas = 2.36V (10 ohm, voltage drops to 24V)
This with the output of differential not connected to any stage, only to my multimeter.
The output offset is to large.
A previous thread suggested to some pulldown to a negative rail, but it is not available on this project so I can not tried. Any idea what to change to make it work ?
Stay with the LM324 for now.
It's possible to have a combination of separate problems.
One is the inability of the opamp's output to properly sink the pull-up current from the input resistors.
The other possible reason are opamp's input offset and input resistor imbalance.
Even a genuine LM324 should not be expected to properly sink the 60µA from the input resistors causing some uncontrolled voltage.
Try adding the current sink I suggested in my previous post. You should do a bread-board test to confirm that my results are repeatable.
If you get a less than acceptable result when added to the PSU, keep in mind the other possible causes.
-
This is almost interesting,
Inverted operation is often used in analog switch applications. Due to the doping of ordinary transistors, the emitter-collector saturation voltage is almost negligible, a few milivolts vs. 20-40mV. This makes inverted BJTs competitive with JFETs for analog switches, as long as the low hFE can be addressed.
From, https://www.avrfreaks.net/forum/upside-down-bjts (https://www.avrfreaks.net/forum/upside-down-bjts)
-
Inverted bipolar? So colector to GND , emitor to output of opamp and base to where?
-
Inverted bipolar? So colector to GND , emitor to output of opamp and base to where?
That's right. Feed about 100µA into the Base, like a 39K resistor from a 5V source.
-
Looking at your readings in Reply #4, the 0.21V at the opamp's output could be due to resistor imbalance. This would mean that the opamp is operating closed loop. One way to confirm is to see if the opamp's output voltage rises by the proper amount when the PSU is loaded by a small current like 50mA.
What doesn't make a lot of sense is the opamp's voltage rising when the rest of the circuitry is connected which should only be 2 opamp inputs. What voltage is on TP14 and what does it drop to when loaded with a 10K resistor?
Oscillations can cause hard to explain circuit behaviors.
The input pins on U2B appear to be reversed in the schematic. And D4 would never get enough voltage to illuminate.
-
Looking at your readings in Reply #4, the 0.21V at the opamp's output could be due to resistor imbalance. This would mean that the opamp is operating closed loop. One way to confirm is to see if the opamp's output voltage rises by the proper amount when the PSU is loaded by a small current like 50mA.
Does no have any effect until >200mA after this measure correctly(voltage out psu <20V), if comparator not conected.
-
What doesn't make a lot of sense is the opamp's voltage rising when the rest of the circuitry is connected which should only be 2 opamp inputs. What voltage is on TP14 and what does it drop to when loaded with a 10K resistor?
The voltage remais the same 0.2A with no load.
-
What doesn't make a lot of sense is the opamp's voltage rising when the rest of the circuitry is connected which should only be 2 opamp inputs. What voltage is on TP14 and what does it drop to when loaded with a 10K resistor?
The voltage remais the same 0.2A with no load.
I should have explained better.
What does the voltage measure on the rest of the circuitry while disconnected from U2A?
And when loaded with a 10K resistor?
-
Oscillations can cause hard to explain circuit behaviors.
The input pins on U2B appear to be reversed in the schematic. And D4 would never get enough voltage to illuminate.
The oscilation are not present in the Voltage close loop, I checked with Agilent Oscilloscope. The current loop for the moment is open on Tp14, since the measurement for currents <200ms does not work at all. I will try the reverse transz. Is to not better for a Fet current sink ?
I will also reverse pins on U2B.
-
Oscillations can cause hard to explain circuit behaviors.
The input pins on U2B appear to be reversed in the schematic. And D4 would never get enough voltage to illuminate.
The oscilation are not present in the Voltage close loop, I checked with Agilent Oscilloscope. The current loop for the moment is open on Tp14, since the measurement for currents <200ms does not work at all. I will try the reverse transz. Is to not better for a Fet current sink ?
I will also reverse pins on U2B.
I don't how much current the FET will sink. We don't want to overload the opamp. The inverse BJT should work well enough.
There is about 60µA from the input resistors and a few nA from the other opamps.
-
Oscillations can cause hard to explain circuit behaviors.
The input pins on U2B appear to be reversed in the schematic. And D4 would never get enough voltage to illuminate.
The oscilation are not present in the Voltage close loop, I checked with Agilent Oscilloscope. The current loop for the moment is open on Tp14, since the measurement for currents <200mA does not work at all. I will try the reverse transz. Is to not better for a Fet current sink ?
I will also reverse pins on U2B.
I don't how much current the FET will sink. We don't want to overload the opamp. The inverse BJT should work well enough.
There is about 60µA from the input resistors and a few nA from the other opamps.
183mV with BC547 reverse configuration 39k base to 5V.
Conecting a load of 120ohm I measure 0.426.
35ohm -> 0.910V
10ohm-> burn the opamp, 6V, droping , and opamp is heating up. No worry, I have more.
-
I think 32V supply voltage for lm324 it is to much. :horse:
-
Another topology you can use is to make a current mirror with an opamp (or even simple BJT's) I have not checked the lf411, but the inputs of the good old TL072 have a common mode range of upto the power supply rails.
So put one leg of another resistors (For example 1kOhm) also on the power rails and connect it with a darlington (such as BC517) or P-Mosfet to the output of the TL072. The output of the TL072 can not go to the Positive power rails, but the 2 diode drops of a darlington or MOSfet give the output enough headroom to regulate. Connect the opamp in such a way that it compares the voltage over the shunt with the voltage drop over the 1k resistor.
The Collector current of the darlington then goes to GND, via another resistor, which generates a voltage referenced to GND and is easy to measure.
The advantage of this method is that you do not need a voltage divider, and therefore additional amplification which would also amplify the offset voltage of the opamp.
-
Oscillations can cause hard to explain circuit behaviors.
The input pins on U2B appear to be reversed in the schematic. And D4 would never get enough voltage to illuminate.
The oscilation are not present in the Voltage close loop, I checked with Agilent Oscilloscope. The current loop for the moment is open on Tp14, since the measurement for currents <200mA does not work at all. I will try the reverse transz. Is to not better for a Fet current sink ?
I will also reverse pins on U2B.
I don't how much current the FET will sink. We don't want to overload the opamp. The inverse BJT should work well enough.
There is about 60µA from the input resistors and a few nA from the other opamps.
183mV with BC547 reverse configuration 39k base to 5V.
Conecting a load of 120ohm I measure 0.426.
35ohm -> 0.910V
10ohm-> burn the opamp, 6V, droping , and opamp is heating up. No worry, I have more.
At 2.5A load, the opamp's output should have been about 2.7V.
The inverse BC548 I tested draws 0.6mA at 5V Vec, The current increased sharply at about 7V as the Emitter Base junction started to zener.
You should bread-board test your inverse transistor to know how much current it will draw.
The other possible reason is that the opamp is oscillating due to feedback phase lag. Put a 10nF capacitor from its output to Inverting input.
-
Another topology you can use is to make a current mirror with an opamp (or even simple BJT's) I have not checked the lf411, but the inputs of the good old TL072 have a common mode range of upto the power supply rails.
So put one leg of another resistors (For example 1kOhm) also on the power rails and connect it with a darlington (such as BC517) or P-Mosfet to the output of the TL072. The output of the TL072 can not go to the Positive power rails, but the 2 diode drops of a darlington or MOSfet give the output enough headroom to regulate. Connect the opamp in such a way that it compares the voltage over the shunt with the voltage drop over the 1k resistor.
The Collector current of the darlington then goes to GND, via another resistor, which generates a voltage referenced to GND and is easy to measure.
The advantage of this method is that you do not need a voltage divider, and therefore additional amplification which would also amplify the offset voltage of the opamp.
Another member applied that method for high-side current metering.
https://www.eevblog.com/forum/projects/improving-adjustable-dual-voltage-bench-power-supply/msg3028760/#msg3028760 (https://www.eevblog.com/forum/projects/improving-adjustable-dual-voltage-bench-power-supply/msg3028760/#msg3028760)
It would be difficult to apply anywhere other than on the input side of the pass element.
D8 wasn't needed when the opamp was replaced with a TL071.
-
I did a bread-board test with an LM358.
First thing I realized is that there isn't a hope in hell of it working in a useful way without including a balancing trimmer in one of the feedback/input dividers.
At 25V, the opamp's output pin was 70mV. I added the inverse BJT using another BC548 which pulled the output pin down to 4mV.
When I brought the output pin upto 2.5V by mis-adjusting the balancing trimmer, the inverse BJT drew a very safe 0.6mA.
The input offset on the LM358 I used is 1.2mV. Input offset opens another can of worms.
Although I noticed no instability, I would not rule out the possibility.
-
I did try with 10nF but the output remains the same(0.2V), I change the lm324 with a new one. The supply voltage is 32V. The inverse bc547C is giving 1.95mA at 5V.
Do you have in the breadboard current sense high side 100k/330k and 0.3ohm sense rezistor powered to 32V?
-
I did a bread-board test with an LM358.
First thing I realized is that there isn't a hope in hell of it working in a useful way without including a balancing trimmer in one of the feedback/input dividers.
Balancing trimemer? Like a potentiometer connected how?
-
I did try with 10nF but the output remains the same(0.2V), I change the lm324 with a new one. The supply voltage is 32V. The inverse bc547C is giving 1.95mA at 5V.
Do you have in the breadboard current sense high side 100k/330k and 0.3ohm sense rezistor powered to 32V?
I used the same value resistors except 1 ohm for the CS,
The trimpot goes at the junction of a 100K, a 330K and an input pin. The resistors connect each to an outer pin of the trimpot. The wiper connects to the input pin. It can be at either input side, inverting or non-inverting.
A lower value trimpot will give fine adjustment. I used a 10K with a 2K across it.
I used 25V for the opamp and to the CS resistor.
-
Can you please have a look on the picture to check if I understand so far correctly the posts? The Opamo is lm324. I checked the bc547 with my multimetter and it is Ok.
-
What if I supply the op amp at 5V just for the current sense?
-
Can you please have a look on the picture to check if I understand so far correctly the posts? The Opamo is lm324. I checked the bc547 with my multimetter and it is Ok.
Move R13 to the other outer pin of the trimpot.
-
What if I supply the op amp at 5V just for the current sense?
5V supply wouldn't be enough. The minimum supply voltage is 3/4 of the maximum expected PSU output voltage plus a few volts to satisfy the opamp's common mode input range.
If the PSU can go to 28V, then 21 plus another 2 to 3 volts should be enough for the opamp.
-
Can you please have a look on the picture to check if I understand so far correctly the posts? The Opamo is lm324. I checked the bc547 with my multimetter and it is Ok.
Move R13 to the other outer pin of the trimpot.
I have moved the R13 to the other pin to a 1k trimpot and removed 2K resitor. After triming the potentiometer I get this results:
Ix Uoutop Ie Uoutop-no-bjtrev
0A 10mV 2mA 577mV
0.22A 252mV 2mA 570mV
0.69A 725mV 2mA 718mV
2A 2V 2mA 2V
So it seems that works, remains the high current 2mA on the reverse bjt, can be explained, or is to high ?
Is there another better option to change opamp to a better one ? Or a INA solution?
-
Can you please have a look on the picture to check if I understand so far correctly the posts? The Opamo is lm324. I checked the bc547 with my multimetter and it is Ok.
Move R13 to the other outer pin of the trimpot.
I have moved the R13 to the other pin to a 1k trimpot and removed 2K resitor. After triming the potentiometer I get this results:
Ix Uoutop Ie Uoutop-no-bjtrev
0A 10mV 2mA 577mV
0.22A 252mV 2mA 570mV
0.69A 725mV 2mA 718mV
2A 2V 2mA 2V
So it seems that works, remains the high current 2mA on the reverse bjt, can be explained, or is to high ?
Is there another better option to change opamp to a better one ? Or a INA solution?
I have tested 2 separate BC548's and got the same results. It acts like a 100Ω resistor at very low voltages and draws only 0.6mA at 5V,
The Base current can be reduced but the opamp's uncontrolled output voltage will rise.
How are you measuring the 2mA?
There are better options but need a negative rail.
There definitely would be better opamps in other respects such as input offset but not necessarily in solving the uncontrolled output voltage problem.
The 2mA load isn't going to harm the opamp.
More: Are you assuming the 2mA from a simulation?
-
I don't think it gets any better than the OPA180.
It's 36V, very low input offset and drift and rail to rail output.
The inverse BJT is still likely to be needed to get down to a very low uncontrolled output voltage.
The input includes negative rail.
TI have many parts that are rated as rail to rail input which might seem like a good idea but there is a catch.
There can be a zone of increased offset around 13V, like with the OPA192.
-
The 2mA load isn't going to harm the opamp.
More: Are you assuming the 2mA from a simulation?
No, it is measure value by inserting Keyside U1241B on the mA range between emitor and output of opamp.
-
Where did you get the opamps from? Can you take a macro photo of the top of the package?
-
Also what is the lowest opamp output voltage and current by adjusting the trimpot?
-
Where did you get the opamps from? Can you take a macro photo of the top of the package?
I have them from farnell.
-
Also what is the lowest opamp output voltage and current by adjusting the trimpot?
The Iebjt=2mA is the maxim value measured by trimming the pot(Uopamp=220mV). The minum value is 0.07mA (Uopam =8.9mV). This value are with no load.
I think I measured and documented wrong on the previous post, can not reproduce again the value Ie.
-
This is remarkable, 1Kohm can lead to this behavior. I need to pick resistors more carefully, specially in this kind of measurement. Thanks man. :-+
-
Also what is the lowest opamp output voltage and current by adjusting the trimpot?
The Iebjt=2mA is the maxim value measured by trimming the pot(Uopamp=220mV). The minum value is 0.07mA (Uopam =8.9mV). This value are with no load.
I think I measured and documented wrong on the previous post, can not reproduce again the value Ie.
That makes better sense. The 0.07mA matches the current from the feedback resistors. I could get the voltage down to 4mV with an inverse BC548 although I wouldn't expect a BC547 to be much different.
Is the circuit just for the CC or metering as well? It might be tricky to calibrate properly.
-
the circuit is for both: constant current and also for metering, input of 1 gain opamp to the ADC input. But what we did so far is on the testboard.
-
Just a few last things to mention for now.
R2 should be moved to the other side of the CS resistor so that its current won't be measured.
The voltage feedback divider should be made to draw less current.
Did you find the design or made it from the start?
-
It is designed from from scratch by me and the guys from this community. I am sure there are more mistakes in the schematic :popcorn:
You can find all my files on github:
https://github.com/bogdan-tirzioru/Powersupply
-
yesterday I tried with LF411, but the output is always 2V with no load, connecting load does not do anything. If adding load does not do anything. 100k /330k and 0.3ohm sense resitor, there is a link here with the schematic.
The reason why I do this because the I have a power supply here:
https://github.com/bogdan-tirzioru/Powersupply/blob/master/factory/sampleB/Job2.PDF (https://github.com/bogdan-tirzioru/Powersupply/blob/master/factory/sampleB/Job2.PDF)
Implemented with lm324, i think it is similar to lm358, but with the supply voltage of 32V with no load I have 0.20V on the output of the diferential amplifier. Connecting the next stage the voltage goes to 0.5V with no load. :-//
I did try yestarday also the OPA134 and LT1490 but had the same results as lm324:
I = 0A -> Umeas=0.21V (no load)
I = 0.27A -> Umeas = 0.43V (75 ohm load)
I = 0.72A -> Umeas = 0.92V (35 ohm load)
I = 2.2A -> Umeas = 2.36V (10 ohm, voltage drops to 24V)
This with the output of differential not connected to any stage, only to my multimeter.
The output offset is to large.
A previous thread suggested to some pulldown to a negative rail, but it is not available on this project so I can not tried. Any idea what to change to make it work ?
Was this in simulation or real life?
One thing to bear in mind about simulation is models often don't implement the op-amp output stage correctly. In real life the TL072's and probably the LF412's output stage can be pulled up or down to the positive/negative rail, with a current source/sink. See the thread linked below:
https://www.eevblog.com/forum/projects/better-ltspice-tl072-model/msg999066/#msg999066 (https://www.eevblog.com/forum/projects/better-ltspice-tl072-model/msg999066/#msg999066)
-
yesterday I tried with LF411, but the output is always 2V with no load, connecting load does not do anything. If adding load does not do anything. 100k /330k and 0.3ohm sense resitor, there is a link here with the schematic.
The reason why I do this because the I have a power supply here:
https://github.com/bogdan-tirzioru/Powersupply/blob/master/factory/sampleB/Job2.PDF (https://github.com/bogdan-tirzioru/Powersupply/blob/master/factory/sampleB/Job2.PDF)
Implemented with lm324, i think it is similar to lm358, but with the supply voltage of 32V with no load I have 0.20V on the output of the diferential amplifier. Connecting the next stage the voltage goes to 0.5V with no load. :-//
I did try yestarday also the OPA134 and LT1490 but had the same results as lm324:
I = 0A -> Umeas=0.21V (no load)
I = 0.27A -> Umeas = 0.43V (75 ohm load)
I = 0.72A -> Umeas = 0.92V (35 ohm load)
I = 2.2A -> Umeas = 2.36V (10 ohm, voltage drops to 24V)
This with the output of differential not connected to any stage, only to my multimeter.
The output offset is to large.
A previous thread suggested to some pulldown to a negative rail, but it is not available on this project so I can not tried. Any idea what to change to make it work ?
Was this in simulation or real life?
One thing to bear in mind about simulation is models often don't implement the op-amp output stage correctly. In real life the TL072's and probably the LF412's output stage can be pulled up or down to the positive/negative rail, with a current source/sink. See the thread linked below:
https://www.eevblog.com/forum/projects/better-ltspice-tl072-model/msg999066/#msg999066 (https://www.eevblog.com/forum/projects/better-ltspice-tl072-model/msg999066/#msg999066)
LT spice was used to draw the schematic, I could not replicate this behavior in the simulation, they are working perfecly :) until you face reality.
-
This is remarkable, 1Kohm can lead to this behavior. I need to pick resistors more carefully, specially in this kind of measurement. Thanks man. :-+
With the PSU set to output 25V, a 50Ω change in one of the feedback resistors will change the opamps output by about 10mV, that's 10mA of current indication.