Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Suggest improvements for this active load?

Status
Not open for further replies.
U

uoficowboy

Full Member level 3
Joined
Apr 4, 2009
Messages
169
Helped
6
Reputation
12
Reaction score
5
Trophy points
1,298
Location
Seattle, Wa, USA
Activity points
2,964
Hi - I'm working on a high power active load. I'm planning on dissipating most of my power in a MOSFET, and then I'll be using a resistor with kelvin connections to monitor the current in it. I've attached a picture of the schematic for it below.

The FET and current sensor will be attached to a heatsink (and cooled with a fan), whereas the rest of the circuitry will be on a PCB that is about 20 cm away and connected via a cable. No high current will be flowing to the PCB.

Does this schematic look OK? The value of the shunt resistor is not yet chosen (thus the gain on the difference amplifier may need to change) and I'm not sure of what FET I'll be using, but otherwise this is pretty much what I plan on building.

My biggest worry is noise. I added in a voltage source to give the sense resistor some common mode noise. The circuit seems to handle it pretty well. What should I be doing to get rid of noise? I had thought that putting a small capacitor across the inputs of U2 would help get rid of common mode noise, but that didn't seem to help. I was also thinking about putting capacitors from U2's inputs to ground to get rid of noise that is not common to both inputs.

Further, I'm wondering if I should be limiting the bandwidth of the difference amplifier circuit. I'm thinking that I should limit it, but still have its bandwidth be several orders of magnitude greater than the FET driver's bandwidth. Does that make sense? I think the best way to limit it would be a capacitor in parallel with R7. Does that seem right?

Thanks so much for your help!!
 

Where do hook up the DUT (device under test, which your circuit provides load for)? Does V3 simulate DUT in your schematic?

C1 in its current location is a good idea, because it helps prevent oscillation of the current sink.

- Nick
 

kender said:
Where do hook up the DUT (device under test, which your circuit provides load for)? Does V3 simulate DUT in your schematic?

C1 in its current location is a good idea, because it helps prevent oscillation of the current sink.

- Nick
Click to expand...
Hi Nick - I don't have a good model for the DUT. I'm testing high power lithium-ion cells. Maybe a small series resistor and a capacitor in parallel with that resistor would be a reasonable first order approximation of the DUT. I'm not sure.
 

uoficowboy said:
Hi Nick - I don't have a good model for the DUT. I'm testing high power lithium-ion cells. Maybe a small series resistor and a capacitor in parallel with that resistor would be a reasonable first order approximation of the DUT. I'm not sure.
Click to expand...
Making a model for a battery is a nontrivial problem. Still, if DUT were a black box without any model, where would it connect to your circuit? Is it between the drain of the MOSFET and ground?
 

kender said:
uoficowboy said:
Hi Nick - I don't have a good model for the DUT. I'm testing high power lithium-ion cells. Maybe a small series resistor and a capacitor in parallel with that resistor would be a reasonable first order approximation of the DUT. I'm not sure.
Click to expand...
Making a model for a battery is a nontrivial problem. Still, if DUT were a black box without any model, where would it connect to your circuit?
Click to expand...
Sorry I missed your question. Yes V3 is the DUT.
 

I see why you sense current through differential OpAmp - long wires between power elements and control loop. For better accuracy, I would use instrumentation amp (e.g. AD620), because it has internal resistor networks that are matched.

Another approach, of course, is to place the feedback loop next to the MOSFET, sense resistor and the whole feedback loop together and use one OpAmp. On the other hand, if your control signal V4 comes from external device (e.g. from a DAC card connected to PC), you still have a problem with moving the control signal onto "shaky" power ground. That issue can be addressed with this circuit:
sgnd_pgnd_const_current_rev02.PNG


- Nick

P.S. I consult. If this is a commercial project, you are in a crunch, and you need professional help, let me know.
 

kender said:
I see why you sense current through differential OpAmp. For better accuracy, I would use instrumentation amp (e.g. AD620), because it has internal resistor networks that are matched.

Another approach, of course, is to place the feedback loop next to keep the MOSFET, sense resistor and the whole feedback loop together and use one OpAmp. On the other hand, if your control signal V4 comes from external device (e.g. from a DAC card connected to PC), you still have a problem with moving the control signal onto "shaky" power ground. That issue can be addressed with this circuit:
http://www.prolifictec.com/filehost/sgnd_pgnd_const_current_rev02.PNG[img]

- Nick

P.S. I consult. If this is a commercial project, you are in a crunch, and you need professional help, let me know.[/quote]
Hi Nick - Thanks for your input! That's an interesting circuit you posted! So you're wrapping any ground error into the control signal to the FET driver. You could even modify it slightly to enable kelvin current sensing. I worry about getting it stable while also maintaining high bandwidth (which is my goal). The feedback loop is split up in a weird way... but my controls theory is a bit rusty so I'd have to sit down with some paper and a pencil!

As for using an instrumentation amplifier - my understanding of the major advantages/disadvantages between the two would be that ins amps have higher input impedance, lower bandwidth, and lower parts count than diff amps. Since I don't really care about input impedance, I figured a diff amp would make sense. You bring up a great point that the resistors need to be matched - I had forgotten about that. I can use high precision resistors to battle this - but when I have more time I'll check to see how bad the error will be. Do you think this is a bad choice?

To answer your question - the signal will be coming from a DAC that is on the same PCB. No ground problems there.

This is for a one off, personal project. So my budget is somewhat tight! Since my volume is low (one) - I don't mind using spendy parts though.

Thanks!

-Michael
Click to expand...
 

uoficowboy said:
As for using an instrumentation amplifier - my understanding of the major advantages/disadvantages between the two would be that ins amps have higher input impedance, lower bandwidth, and lower parts count than diff amps.
Click to expand...
Probably, the biggest advantage of InAmps is high common mode rejection. Yes, InAmps tend to have lower BW. How much BW do you require? What will be the frequency of your control signal (V4)? What's the sample rate of the D/A? Out of curiosity, what type of load will you be simulating?

Here's my experience with an InAmp. I've fed a 110kHz differential signal (20kHz BW centered at 100kHz, Loran signal) into AD620 InAmp with gain of 5 to 10. I didn't get BW limitation from the InAmp (checked with spectrum analyzer). Datasheet for AD620 specifies up to 120kHz with gain of 100.

- Nick
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top