Building an electronic load simulator

Ok, I have browsed for one on the internet that is higher current (15A or more), and I found THIS. now, I have some 2N3055 power transistors, and to drive them, I was thinking I could use a Darlington optocoupeler (4N33, because I have some) to provide the load control. I was planning on regulating the current with PWM from an Arduino. and using an “AttoPilot Voltage and Current Sense Breakout - 180A” from Sparkfun to sense current and voltage.
now, using PWM will chop the current draw, and this will reflect on the current and voltage sense, I assume. can I use a low-pass RC filter between the optocoupeler, and the 2N3055 base to remove the ripple?

My project idea is to use this to load test various sized batteries.

In reference to the attached Fritzing…
I couldn’t find all the right parts in Fritzing.
so U1 is supposed to be a 4N33.
Part 2 is supposed to be the Sparkfun 180A current and voltage sensor.
R4 and R5 is supposed to be the voltage divider in the 180A sensor.
R6 is the load resistor, and will be sourced accordingly. may just use automotive headlight bulbs for the draw.
The RC low pass filter (R2 and C1) values will be determined when I build the prototype.
R3 is just a pull-down resistor.
VCC1 battery will be any battery I need to test, up to 18V.

I don’t have any code yet as it is still just a brainstorm.

I am just wondering if it will work? I am not fully skilled in much analog design. mostly just digital.


I’d suggest just getting some power resistors. Measuring/monitoring the voltage across the (known value) resistor should tell you everything you need to know. If necessary, you can use different resistors for different batteries.

I was planning on regulating the current with PWM from an Arduino. and using an “AttoPilot Voltage and Current Sense Breakout - 180A” from Sparkfun to sense current and voltage.
now, using PWM will chop the current draw, and this will reflect on the current and voltage sense, I assume. can I use a low-pass RC filter between the optocoupeler, and the 2N3055 base to remove the ripple?

I don’t see any reason to “chop” the current for battery testing.

which will it be then ?
the constant adjustable current supply of 1.8A in your diagram?
The 60A with load resistors?
Or the 180A with the Sparkfun sensor?

Forgive me for being short, but it does not seem like you know what you want to do (exactly). I spent 2+hrs in Multisim refining your circuit diagram, it will not work beyond 2A, so if you want to measure like 60A now, guess i wasted 2hrs.. :confused:

The 4n33 as you spec'd has an ABSOLUTE Collector rating of 100mA which would suffice provided you do not change the 18V OR 10 ohm load resistor. The 2n3055 will handle ~10A without much trouble (about 200-400mA base current for 10A collector current. After that, gain will drop like a rock and base current will shoot up to ove 2A, defeating the purpose of said 2n3055.

Going that route of your 180A sensor, you'd need:
20x 2n3055's = loaded @10A ea = ~250:500mA ea= 5-10A total base current drive needed
20x 0.22 ohm ballast resistors - dont forget these, bad juju will ensue otherwise (video tape it if you leave these out please) XD
1xMJ2955 PNP to control all those base's
and for good measure a Darlington to control that PNP's base, connect that darlington's base to your low pass filter of 1uf and a series ~400 ohm resistor to your arduino pin directly. Good to go!

An aside - 180A@18V you'd be dissipating 3.24KW!, so you might want that heat sink for those 20 2n3055's made in the water cooled variety and STOUT water pump! Could easily be a hot tub water heater, lol.

Google - carbon pile battery tester
also a copy of The Art of Electronics 3rd edition Must have BIBLE if you're going to tinker.

Worth Watching: EEVblog episode 102

@Travis - apologies for being short the other day was trying hard to get a sketch to fetch data from the eeprom in a custom struct array, but I failed to notice two bugs in my for loop. So as you replied to my post above, yes, you're pretty much on the money with those presumptions. When you're ready and if you like some assistance on the rough layout of a transistor based test setup for variable constant current load I may be of some help. I'm in the middle of my automotive LED headlight replacements and am using custom designed Variable constant current drives for the 4000+lm@4A (ea) headlights. These drives also run a small current through the main LED to sample the Vf as a function of temperature to control the water loop.

My drives are based on a few mosfets for PWM'ing the bases of TIP42's PNP 4A 10A pulsed transistors in a 10 channel configuration.
Yes - The ballast resistors will also reduce by half and the 2n3055 base drive transistor can be half the rating, just keep in mind the current range of the base drive and choose an appropriate transistor approx 2x the rating as a safety factor.

You'll also notice in the datasheets of transistors the gain values, basically the gain will drop off as base current increases past a "knee" point as the transistor transitions into its saturation region where maximum collector current is reached. the typical gain of 50 only occurs when the base is at about 0.5A, after that the gain will reduce rapidly to about gain 5, when you'll have to dump nearly 2A into each bse to get the full 14A from each transistor.
Generally, to prevent thermal runaway and to make your design more reliable, you want to select components that can handle 2x the power you intend to power them with. This can be circumvented somewhat in your case if an adequate heat sink is employed.
When paralleling transistors like the 2n3055 keep in mind to isolate the case of the transistor electrically from the heat sink, & other transistors, as the case is connected to the collector.

A custom copper waterblock heat sink and PC 120mm radiator cpu water cooling kit with pump could easily cool off 6-10 2n3055's

You might want to browse mouser dot com and narrow your search by selecting transistors with the desired specifications. NPN Darlington power transistors would be a good choice as your output devices, their gains are typically in the 1000-2000 range. this means for a 15A Darlington with a gain of 1000 (very conservative) you only need a base current of ~ 15mA (0.015A). This is much better base current when compared to the 2n3055 of gain 50 and base current of 0.4-0.5A.

You will need serious heat sinking, as in large expensive finned chunks of aluminum with dedicated fans blowing down them, in order to dissipate 3.5kW. This is the same amount of power as running 3 hair dryers on max heat. That also means learning how to attach the transistors to the heat sinks.

In fact, you might want to open your windows and turn on the fan as well, or wait for the dead of winter...

Its a good eevblog by Dave.

If you are going to have to parallel up many load control devices to get you current spec, you will have to consider wiring layout and heatsinking when it comes to current sharing through the devices (BJT, MOSFET etc.).

Tom... :slight_smile:

And depending on the battery, PWM != CW. I would go for analog control.

PWM is fine for motor control, but not testing.

"The art of Electronics - Third edition"

wow, expensive book! I have added it to my shopping list for a future purchase.

Thank you for all your help.

I have the older second edition and would certainly add that to my list. A cheaper book that is also quite good (not not nearly the detail "The Art of Electronics" has) is "Practical Electronics for Inventors. Some time ago I posted (they moved it to another section of the forum) an interview with Horowitz (one of the authors of "The Art of Electronics") - I thought it was an interesting interview with the man. Certainly knows his stuff. Here is the link to the interview with Lady Ada that I had posted back in March:
Lady Ada interviews Paul Horowitz

Pwillard's Youtube link in Reply#5 is a good tutorial on power dissipation and heatsinks, that explains the math for calculating the heat rise in a heatsink based on power dissipated.

If I may make a suggestion, if I were you I would post a your Design Criteria and then do the math for the heatsink . You will be probably conclude you need a fairly decent size heatsink with the fan built into in. You'll have to drill and tap the power transistor mounting holes but that's really not a big deal. Just get some cutting oil to lube the tap. Look up the drill size online get yourself a drill if you don't have one. I would also recommend using thermal paste when you mount the power transistors.
TO3 mounting hardware

TO3 mounting kits


TO3 mounting kits


What do you have for a supply that is putting out 60A?