0
\$\begingroup\$

I want to design a current source that can provide 2 different currents, 100mA and 1A using just a single power amplifier APEX MP38 which has a current limiting feature which is set by a resistor according to the following formula:

$${R_{limit}=\frac{0.7V}{I}}$$

The maximum load that current should drive is an almost-pure resistive load around 50 Ohms.

I will supply the opamp +Vs by 60V and -Vs will be connected to GND.

MP38 As current source

Almost all opamp based current sources that I see, except Howland Current Pump which is too complex for my case (because I only need positive currents) use a PFET at the output of the opamp. Since the opamp that I have is already been purchased I kind of have to use it for this case.

Lets say I choose R3 and R4 in such a way that in parallel they make a sense resistor to limit the current to 1A, and when only R3 is conducting the current will be limited to 100mA.

So the questions are:

  1. Overall, do you think this is going to work?
  2. Is using a switch (e.g. cold-switching a relay) to switch between 2 different sense resistors a wise idea? is it a misuse of the current limiting feature which I beleive is only should be used to define a SOA?
  3. Having the -Vs connected to ground is OK or I need to provide the negative rail of the opamp with some voltage to get out of common mode voltage range?
  4. I want to use a 0-3V3 digital signal to turn on/off the output of the opamp. What sort of gain will I need? Do I need to use another opamp in buffer mode at an earlier stage before the main opamp?
  5. Do I need any capacitors in feedback or other paths (Of course the supply bypass caps are going to be there!)

UPDATE The load that I am talking about is a series of PCB traces that I want to see how do they perform under high-current stress. So it should be theoretically a resistive load with some small/negligible parasitic inductance/capacitance.

\$\endgroup\$
  • \$\begingroup\$ "The maximum load that current should drive is an almost-pure resistive load around 50 Ohms." What are you driving and your intention to build? \$\endgroup\$ – Jason Han Aug 7 '17 at 10:11
  • \$\begingroup\$ Basing the design on something that you feel you have to use because you have already bought it isn't the sensible approach especially when you consider the amount of heat dissipation when the device current limits at 1 amp. Your previous question (electronics.stackexchange.com/questions/321661/…) had answers that advised a different approach so maybe you should consider stating why those approaches (using PWM) are unsuitable for you. \$\endgroup\$ – Andy aka Aug 7 '17 at 11:38
  • \$\begingroup\$ @JasonHan I updated my question with information about the load. \$\endgroup\$ – Sean87 Aug 7 '17 at 13:24
  • \$\begingroup\$ @Andyaka Well I know I should not base it on purchased components but in this case, I kind of have too :( Heat is not the biggest issue, I will use a CPU cooler on the opamp. The load is just a series of PCB traces that I need to stress with that current to see how they live up to it. \$\endgroup\$ – Sean87 Aug 7 '17 at 13:28
1
\$\begingroup\$
  1. You need the output to be a minimum 6.8 volts (-Vs + 6.8), so you will need a resistor in series with R3 on the output as shown. The 50 ohm load only provides 5 volts at 100 ma and your sense resistor adds 0.7 volts. Also keep in mind that the output current limit will vary somewhat as the base-emitter voltage changes with temperature.

  2. Using a switch as you have drawn will be ok. Expect some spikes when you switch.

3 and 4. In addition to comment (1.) you would need the input to be a minimum of 15 volts (-VB + 15), so you will have to bias up the signal on the input op amp.

  1. A capacitor in parallel with R2 will slow down the rise and fall times. enter image description here
\$\endgroup\$
  • \$\begingroup\$ When there's 5V across the 5 ohm load, the Rsense may not be 0.7 as the current is not 100mA. \$\endgroup\$ – Jason Han Aug 7 '17 at 23:44
  • \$\begingroup\$ if the current voltage is less than 0.7 the current is not limited. \$\endgroup\$ – John Birckhead Aug 8 '17 at 1:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.