Nice Video! I have used both of those current sources many times. The Op-amp mosfet current source can be prone to oscillation depending on what components you use. (I found out when my wife said that the TV was getting interference (back in the analog TV days). It was oscillating up in the 100MHz range. I switched to a slower Op Amp and that cured it. Adding some RC compensation would work as well.
Great comment, appreciate you sharing some personal experience with current source circuits! Are you recommending the RC network between the op amp output and MOSFET gate to form a low pass filter?
Thanks ! My question is What is the maximum précision that you can get with both circuit ? I have tried LM317 already but with thermal shifting (even with heat sink) the accuracy is less than 2%. Since I want to build a power micr ohmeter ( one amp and less), I need a very precise current source without drift.
The stability of the LM317 is typically around 10mV for junction temperatures from 0°C to 125°C according to the datasheets. You can't do much to improve that. For the opamp circuit, the principal source of variation (the mosfet) is inside the feedback loop so the effect from that is virtually eliminated. You're left with controlling the input offset voltage drift and the resistance drift in the sense resistor. You might use a TL072 with a typical drift of 150μV over the same range. Using a low temperature coefficient sense resistor will also help to avoid drift. Common metal film resistors have temperature coefficients up to around 100ppm/°C, so contribute a variation of up to 1%, but you can get resistors with 1/10 of that.
Hi, I was looking for a method to drive a high current (1.5A) using a microcontroller and fortunately found this video. I still don't quite understand how transistor circuits work, for example if we lower the DAC output voltage which changes the current, the voltage on the sense resistor and load will go down but Vcc remains constant. I thought that many transistors have a fixed voltage drop between terminals but here it is variable, to act as a variable resistor? I hardly know where to begin to find the components I will need to make this circuit to my specifications.
Hey thanks for the comment. Look at a data sheet for a transistor and it should have some type of IV curve based on the gate control signal. In the case of a mosfet that gate control signal is voltage control. You'll see that they have a nonlinear region in that IV curve and that's the region it's being controlled in to act like a variable resistor. You can always set up a mosfet test circuit yourself too and observe this behavior with a power supply and a voltage to source to control it.
@@ForceTronicsThank you for your reply. If the op-amp feedback makes it so that the + and - terminals have the same voltage, won't that mean there is a negligible voltage on the gate and therefore only a small current?
Max current is a tradeoff between the sense resistor value you use and the LM317 current rating. The goal of the video combined with the LM317 datasheet should equip you to answer this question yourself. If something in the video is not clear about that, let me know.
Great video. Is the MOSFET circuit more efficient? Ignoring the advantages of the Arduino control, is there an advantage to one over the other? Thanks!
Great question and maybe I should have done a pros and cons comparison in the video. The first method is easier to implement because it has less components and the linear regulator has built in protection features. Also, the linear regulator is responsible for handling the control loop to regulate it's output. For the second method you as the designer are responsible for properly spec'ing the op amp and MSOFET, if you look at the previous comments you need to watch out for oscillations in the control loop of the second circuit. The advantage of the second method is you the designer have tighter control over the specs and performance of the design. Also the second method is more efficient than using the LM317 and is most likely more efficient compared to other higher performance linear regulator on the market.
Some other advantages of the MOSFET version are possiblities of much higher operating voltages and current due to availability of power MOSFETs which can handle voltages in 100's of volts and 10's of amps. Also the current adjustment range can go down to almost zero amps while the upper value is only limitted by the MOSFET power rating, set by the product of actual drain to source voltage and drain current. So with the choice of right device and decent heatsinking you can have quite a powerful result. Be careful to not become a load for it yourself if using higher voltages as the circuit will try and adjust the voltage across the "load" in order to try and maintain the target current which means that it can be quite lethal even with few milli-amps of output current. Also technically a minor point, the circuit shown here is actually a current sink and not a source. Great video though. Oh, one other thing which can be easily overlooked, you can have multiple current sources with different values of output current all in parallel and output will be the sum of all individual currents. So technically you could build a bunch of identical circuits powered by the same supply and have them all pumping more current than what a single device will do.
Nice Video! I have used both of those current sources many times. The Op-amp mosfet current source can be prone to oscillation depending on what components you use. (I found out when my wife said that the TV was getting interference (back in the analog TV days). It was oscillating up in the 100MHz range. I switched to a slower Op Amp and that cured it. Adding some RC compensation would work as well.
Great comment, appreciate you sharing some personal experience with current source circuits! Are you recommending the RC network between the op amp output and MOSFET gate to form a low pass filter?
Try placing a feedback capacitor (e.g. 200pf) in parallel with the feedback path to the inverting input to mitigate the oscillation.
Yes - you want to reduce the high frequency gain of the op-amp with an RC network negative terminal to output terminal.@@Enigma758
Thank You
Thanks ! My question is What is the maximum précision that you can get with both circuit ?
I have tried LM317 already but with thermal shifting (even with heat sink) the accuracy is less than 2%.
Since I want to build a power micr ohmeter ( one amp and less), I need a very precise current source without drift.
The stability of the LM317 is typically around 10mV for junction temperatures from 0°C to 125°C according to the datasheets. You can't do much to improve that. For the opamp circuit, the principal source of variation (the mosfet) is inside the feedback loop so the effect from that is virtually eliminated. You're left with controlling the input offset voltage drift and the resistance drift in the sense resistor. You might use a TL072 with a typical drift of 150μV over the same range. Using a low temperature coefficient sense resistor will also help to avoid drift. Common metal film resistors have temperature coefficients up to around 100ppm/°C, so contribute a variation of up to 1%, but you can get resistors with 1/10 of that.
@@RexxSchneider ThANK YOU. I will try it.
Hi, I was looking for a method to drive a high current (1.5A) using a microcontroller and fortunately found this video. I still don't quite understand how transistor circuits work, for example if we lower the DAC output voltage which changes the current, the voltage on the sense resistor and load will go down but Vcc remains constant. I thought that many transistors have a fixed voltage drop between terminals but here it is variable, to act as a variable resistor? I hardly know where to begin to find the components I will need to make this circuit to my specifications.
Hey thanks for the comment. Look at a data sheet for a transistor and it should have some type of IV curve based on the gate control signal. In the case of a mosfet that gate control signal is voltage control. You'll see that they have a nonlinear region in that IV curve and that's the region it's being controlled in to act like a variable resistor. You can always set up a mosfet test circuit yourself too and observe this behavior with a power supply and a voltage to source to control it.
@@ForceTronicsThank you for your reply. If the op-amp feedback makes it so that the + and - terminals have the same voltage, won't that mean there is a negligible voltage on the gate and therefore only a small current?
What is the max current source that you could make with the LM317?
Max current is a tradeoff between the sense resistor value you use and the LM317 current rating. The goal of the video combined with the LM317 datasheet should equip you to answer this question yourself. If something in the video is not clear about that, let me know.
Great video. Is the MOSFET circuit more efficient? Ignoring the advantages of the Arduino control, is there an advantage to one over the other?
Thanks!
Great question and maybe I should have done a pros and cons comparison in the video. The first method is easier to implement because it has less components and the linear regulator has built in protection features. Also, the linear regulator is responsible for handling the control loop to regulate it's output. For the second method you as the designer are responsible for properly spec'ing the op amp and MSOFET, if you look at the previous comments you need to watch out for oscillations in the control loop of the second circuit. The advantage of the second method is you the designer have tighter control over the specs and performance of the design. Also the second method is more efficient than using the LM317 and is most likely more efficient compared to other higher performance linear regulator on the market.
Some other advantages of the MOSFET version are possiblities of much higher operating voltages and current due to availability of power MOSFETs which can handle voltages in 100's of volts and 10's of amps.
Also the current adjustment range can go down to almost zero amps while the upper value is only limitted by the MOSFET power rating, set by the product of actual drain to source voltage and drain current.
So with the choice of right device and decent heatsinking you can have quite a powerful result. Be careful to not become a load for it yourself if using higher voltages as the circuit will try and adjust the voltage across the "load" in order to try and maintain the target current which means that it can be quite lethal even with few milli-amps of output current. Also technically a minor point, the circuit shown here is actually a current sink and not a source.
Great video though.
Oh, one other thing which can be easily overlooked, you can have multiple current sources with different values of output current all in parallel and output will be the sum of all individual currents. So technically you could build a bunch of identical circuits powered by the same supply and have them all pumping more current than what a single device will do.