Complete project: Constant Current Dummy Load

[johnmx]

Hi fellows,

Since I’m planning to build two bench power supplies, I needed to build a current load to test them. So, I’m going to show you my version of the dummy load.
Please note that this project was built ONLY using existing parts in my personal stock.
I had to be silly to put this project on a wood box, but it was the only box I found around. I don’t recommend you to do it, instead use plastic or metal cases.
I wanted to build something simple but with a minimum protection circuit. Thus I added a control circuit to limit the maximum dissipation power to protect the main transistor.

The main specifications are as follows:
-   Current regulation: 0 to 10A
-   Maximum input voltage: 40V
-   Maximum dissipation power: 100W (limiting circuit with warning LED)
-   Analog control: 10 turn potentiometer
-   Analog display: 10A ammeter
-   External power: 12V/0.3A
-   Enable/disable switch
-   Fuse protected: 15A

If anyone has suggestions or simple improvement ideas, I’ll be glad to hear them. Probably I’m not going to change my circuit because it’s almost finished, but it could be useful to someone that could use this circuit as a start point to build their own.

Let me talk a little bit about the circuit. First I wanted to build a self-powered circuit, but since it is necessary to cool the heat sink with a fan, that was out of question. Thus the circuit is powered by an external 12V supply. A TC962 generates the negative supply (-12V) needed by the opamps. I included a RC filter for this supply to reduce the ripple noise. The voltage reference of +4.096V is generated by MCP1541.

The main transistor (BU941) is a Darlington NPN capable to conduct 15A dissipating up to 155W. The transistor tab is connected directly to the heat sink without electric isolation. This ensures a better heat transfer to the heat sink.
The sense resistor is the one included in the ammeter (~5.2mOhm). The exact value is not important because the circuit can be calibrated later with R28 (see schematic below).
The most important of all opamps is U3 because it must have a low offset. I have chosen the MAX430, which is a chopper-stabilized opamp. The maximum input offset voltage is 10µV.

The load current is set by a 10 turn potentiometer (R23). The voltage applied to R23 is controlled by the enable/disable switch and by the power limiting circuit.
The power limiting circuit is based on a simple analog multiplier (circuit around U8). This circuit has 3 inputs: Vx (Vin), Vy (Iin) and Vw (constant divisor). The output signal is defined by Vpwr = Vx*Vy/Vw. R9 adjusts Vw in order to get Vpwr = 4V @ 100W. R18 is used to set the maximum power. When the power is higher than this value, U4B will decrease its output voltage and the voltage at Iref node will decrease too. Consequently it decreases the load current. Opamp U6 is used as a comparator to detect this condition and to turn on the LED. The value of R14 was chosen in order to decrease the output voltage of U4B but also to protect the inverting input of U6 in case the voltage gets negative.

I need to run some tests in order to check if 100W is not too much for the pass transistor. Probably I will need to reduce this value down to 90W or 80W. In any case I have a new and more powerful fan to replace the existing one.

[EEVblog]

Awesome job, thanks for sharing.
The power limiter is a nice inclusion.

Dave.

[zagg]

Hello !

Thanks for sharing.
Very nice project !!!

Have a nice day,

Davy

[sacherjj]

Thanks for the great schematics.  I'm starting to work through my design for this.  I'm building a current limited variable power supply and it made sense to build a constant current load first to use in testing my power supply design.  (I'm building most of my non critical lab equipment as a learning experience.) 

I'm working on integrating this style of current limiter with a Micro for the same Constant Current use (connected or disconnected to computer), but also measure power and source voltage.  For example, you might want to setup a battery discharge at a fixed current and have the computer measure the voltage of the battery.  At a certain level, you will want to switch the load off for unattended runs without completely killing the battery. 

My idea is to run the positive of the source through a SPST relay (normally open) to enable the microprocessor to turn on the load after initial verification.  However, I will need some sensing constantly connected to the source in front of the relay, to determine if we are good to switch on or not.  My switching current on the relay can be limited, but making sure that they PWM output is setup for low current mode before contacts are switched.  Is there any issue with a floating open circuit existing at the Source or Collector of the transistor, with voltage going to the base.  I couldn't think of a problem, but thought I would ask.

My main issue with with source voltage measurement and input protection.  This isn't really a required piece to make this work, as I can just not be a bone head (sorry, deek head) and not hook up a voltage too high.  However, I thought this would be a great project to learn a little more about accurate measurement with input limiting.

I want a voltage divider and follower much like the center portion of the above schematic with the TL082.  If this is sized to be just under the A/D converter max at the maximum expected incoming voltage, all is good.  However, I'm trying to figure out how I would protect this.  If I have an "accurate" divider/follower that should give an accurate in the expected range, but then have another that is a wider range to be safe for all expected inputs and tell if we are too high of voltage.  Do I just clamp the first with a zener or something similar?  I'm concerned that as we get close to the max, I will start losing accuracy due to leakage current.  Or is there a method of buffering the lower voltage range A/D source that is only enabled by the micro after the sampled "always on" A/D confirms we are good to go?

Anyone have any pointers on a good resource to look at for this?

Thanks!

[johnmx]

Is there any issue with a floating open circuit existing at the Source or Collector of the transistor, with voltage going to the base.  I couldn't think of a problem, but thought I would ask.

There is no problem. You can even add a small relay in series with the main switch (SW1 in my schematic).

Regarding the last part of your question, do you need high accuracy measuring only the higher voltage and not all range? I don’t see the point of that, but you can measure all range with a higher resolution ADC. Or you can design the signal conditioning circuit with any amplifier that has over-voltage protection in the inputs.

[sacherjj]

Is there any issue with a floating open circuit existing at the Source or Collector of the transistor, with voltage going to the base.  I couldn't think of a problem, but thought I would ask.

There is no problem. You can even add a small relay in series with the main switch (SW1 in my schematic).

Regarding the last part of your question, do you need high accuracy measuring only the higher voltage and not all range? I don’t see the point of that, but you can measure all range with a higher resolution ADC. Or you can design the signal conditioning circuit with any amplifier that has over-voltage protection in the inputs.

I guess I wouldn't need precision for the high indicator.  If I detect voltage at the rail, I assume it is too high.  What I'm asking for help on is a signal conditioning circuit that doesn't affect the accuracy.  I'm probably using the wrong search terms, as I've been trying to look at the inputs to DVMs and such to figure out how this is done.  Most of what I find is for protecting against ESD, not regular over voltage.

I'll look at an over voltage protection for amplifiers and see what I come up with.  Seems like I should be able to have an amplifier with 1/4 gain with this protection and do the division and protection in one go.

Edit: This looks like what I'm trying to learn.  Time to do some reading.  http://www.analog.com/static/imported-files/tutorials/MT-069.pdf

[johnmx]

I'll look at an over voltage protection for amplifiers and see what I come up with.  Seems like I should be able to have an amplifier with 1/4 gain with this protection and do the division and protection in one go.

You can easily find precision instrumentation amplifiers with built-in overvoltage protection up to ±40V, e.g.:
FET-input (low bias current): INA121
Low offset voltage: INA128, INA129, INA141 (accurate gain without external resistors)

[sacherjj]

I'll look at an over voltage protection for amplifiers and see what I come up with.  Seems like I should be able to have an amplifier with 1/4 gain with this protection and do the division and protection in one go.

You can easily find precision instrumentation amplifiers with built-in overvoltage protection up to ±40V, e.g.:
FET-input (low bias current): INA121
Low offset voltage: INA128, INA129, INA141 (accurate gain without external resistors)

Thanks for the pointers.

[diff]

thanks for sharing, great job.
I'm looking for constant current dummy load around 50A use for testing of my power supply 50A 2-5 hrs.
currently i'm using heater to replace power resistor but the current always drop when the heater generate heat. i want to build DIY constant current dummy load. if i use this circuit which component should decrease?
any advise highly apperiate.

[Kiriakos-GR]

Nice project ... Congrats   

But with out a picture of the PCB of it , I can not feel that much excited about it. 
I always had in the back of my head the idea that I need one ,
so to test the batteries from used UPS units , the one that they use two batteries in pair ,
and usually just the one fails , and you had to replace them both. 

[johnmx]

I cannot share the PCB design because I didn’t make one. I just used a testboard (prototype board).

The final prototype includes a temperature sensor that controls a small fan. This fan is not in the best place, but I only realize that after opening the hole in the box 

[Kiriakos-GR]

I cannot share the PCB design because I didn’t make one. I just used a testboard (prototype board).

No problem  
Some one with more skills , he will feel jealous about this unit , and he will make one printed PCB about it  

And I will wait in the corner , to get it too ..  

[djcrunkmix]

Bookmarked.

Thanks for sharing. Looking forward to other good projects like this  

[Dan]

Thank you for sharing this, great job!  I am working on building it now.  I have a couple of questions though:

The +V on the load input; does that just link to the +V on U7B?  And I'm not sure about the Rshunt; is that MegaOhm?  Doesn't seem right to have such a high value resistor in line with the load drain.  But I'm only about 1 year into building my own circuits so I may just not understand...  Maybe it stands for MilliOhms or MicroOhms?  Though that seems too small of a resistance to give accurate differentials for the OpAmp.  If you could clarify that would be great!  I'm really excited to get this built!  I'll post pictures if you want to see my completed project.

[johnmx]

Hi Dan,

U7B is used to measure the input voltage.

The value of Rshunt is around 5.2 miliOhm. ‘m’ stands for ‘mili’, ‘u’ stands for micro and ‘M’ or ‘Meg’ stands for Mega.

The MAX430 is a special opamp used to amplify low voltage signals. You can use another, but choose one with very low input offset voltage.
The scale of the potentiometer has a resolution of 20mA and for this value the voltage drop on Rshunt is 100uV which is 10x higher than the opamp maximum offset voltage. The maximum voltage at Rshunt is 10[A]*0.0052[Ohm]=52mV.

You can post your pictures.

[Dan]

Ah!  Gotcha!  Thank you so much for the quick reply!  I understand now, thank you for the clarification!  I'll definatly post pics and maybe video of the project.  I am going to try and add in a variable fan speed dependent on temp. Thanks again for posting your schematics and the quick reply.

[TerminalJack505]

Wow!  Thanks for sharing your project.  I really enjoyed studying the schematic.  The project looks really cool too.  I didn't realize that was a wooden box until I read your description.

I do have one question that hopefully the OP or some other analog guru can answer:  Is it safe to not have any DC feedback on op amps U4A and U4B?  As they are now--to my understanding--they are acting as integrators.  I remember learning that these could be problematic since, over time, they can drift one way or the other.  Is it not a problem with the circuit, as designed? 

I guess I can see how U4A might be okay since it does get feedback through U3 but U4B isn't getting any DC feedback so far as I can tell.

I'm sure I'm showing my ignorance here but I just had to ask.

Edited to add:

Okay.  After taking a closer look, U4B is getting DC feedback in a very round-about way.  It will vary Iref, which will vary Vx and Vy, which affects Vpwr.  So, I think I may have answered my own question.

[steve_w]

Excellent work, I shall steal this immediately.

regards

Steve W

[stef_r]

I hope the OP of this post still is active on this forum and see my reply on his project 
I know it's an older post, but I recently got the interest of building a constant current load for a project which I'm currently working on.
(The project includes a 4-output 'switch' controlled by a PIC 16F1789 microcontroller. Each output can / may deliver a current up to 10A. I now use a couple of big and expensive 100W resistors but since the device is for both 12V AND 24V use, see my problem with the fixed values of the resistor???)

So this design drew my attention!
It's very detailed and I hope I find the time to build one of my own.
I do however need to find a 10A meter to use with it.
Other than that, I also might want to customize the drawing to use with a microcontroller.
With the right firmware, I would like to switch between
1) Constant power
2) Constant current
3) constant resistance (option)
I think these two first would be the most important ones!

One question about the schematics though...
Since I'm still learning each day, what does the en/dis switch enable or disable?
Is it just the warning LED or does is en/disable the complete limiting circuit?

I don't know how much interest is there still in this project?
Because If I would design a PCB circuit of it, I would greatfully share it with others here!

Hope to see a reply on this one soon 

[johnmx]

...what does the en/dis switch enable or disable?
Is it just the warning LED or does is en/disable the complete limiting circuit?

Hi,
The en/dis switch is to turn on/off the dummy load.
When the switch is open (off state) the voltage of the signal Iref is zero. This signal is the user setpoint of the current.
The power limiter is always enabled.

[kind2011]

im planing to build one but with simplified diagram for low cost parts for  5A, i hope a guru knows
the problem is i dont know how to calculate the power dissipation i searched and im confused at juntion max temperature relations with max power dissipation of mosfet here im using irf640 mosfet 18A 200v, i know that at 100 Celsius case Temp,  max drain current downs to 13A.

i tested the circuit of the image with a old heatsink from crt monitor,    20v 600mA and it gets hot.

i read and sounds logic that with the increase of case temperature  the drain source resistance increases too,  i dont know the formula to know at with case temperature the current would be 5A without destroyng itself

[kt315]

 
Search the forum -- there is a lot of designs for dummy loads. Dave had a video on dummy load, where he went thru thermal calculations.

  http://en.wikipedia.org/wiki/Thermal_resistance

The chances are you will need a compensation network to prevent your load from oscillating (again, search the forum).

[pstemari]

Doing some fiddling around with this.  Couple of questions:

• There's a 150 ohm resistor (R4) in series with the negative power rail generated by U5. Why?  I don't see any references to this in the datasheet.
• Switching the output to a linear MOSFET (IXTH40N50L2), and running some LTSpice simulations with the closest MOSFET I could find a model for, it seems to be rather twitchy under changing input voltages and has a tendency to oscillate. Is that a known issues with these designs or something caused by the two-stage feedback amp used here?

LTSpice model is attached.

[pstemari]

Answering  my own question, after breadboarding the key control bits (U2 and U4a in the original schematic, IC3 and IC9 below), response to flipping the input voltage off and on seems to be quite stable.  The traces I captured of the gate voltage on the MOSFET showed only minor overshoot that decayed in a few oscillations, about 2 ms Modified schematics are attached.

[johnmx]

Hi pstemari,

I recommend to put a small resistor between the opamp output (IC9) and the MOSFET gate. The gate capacitance of that transistor is too big. And can cause instability problems to IC9.