Simulated DC input signal

[Teun]

Hi,

I need to simulate a DC signal on a product we designed. It needs to replicate the voltage drop of a car battery when starting the engine.

The added image shows the DC level I need to simulate. How can I do this best?
(Programmable Input?, prog. load?, or better ideas?)

The other image shows the schematic with my product as DUT.

[sokoloff]

What do you mean by simulate?

Run it on a desktop/laptop in Spice or equivalent (electrical simulation)?

Run it on a lab bench in 3D with real parts? For this, I'd use a programmable DC power supply (GPIB, serial, or network control).

[Berni]

At work i was involved in building a device that does this.

Most regular PSUs are not fast enough to reproduce the waveform so we used a MCU and DAC to generate the waveform and then boost it up with a power opamp output driver to get it to a max output of 20V 5A.

It turned out to be a pretty useful unit for these tests. All you do is fill up a SD card with a collection of different voltage profiles and it executes the selected profile at the push of a button.

[Teun]

I mean run it on a lab bench with real parts

[BravoV]

It needs to replicate the voltage drop of a car battery when starting the engine.

Not an expert though, why not use your company's DSO , hook-up to the car, of course assuming people at your company already know how NOT to toast the scope like probing technique at car's voltage, and then do DSO single shot at the car's battery or the voltage rail to get an idea what will happened during engine start.

Car's engine starting period up to its running usually happen in quite short period, I guess less than 3 seconds maybe ?

Once captured, use the details info like the voltage spike/drop and etc and apply/review the designed product whether it can live with that.

cmiiw

[Teun]

That is what the first image represents. The voltage of the battery when the starter engine is switched on.

t7 = 15ms, t8 = 50ms, t9 =5s, t11 = 5ms

the total will be used for product testing. 

[Berni]

You can't just hook the device under test across the battery of a random car and test it.

These test waveforms are usually worst case scenarios. In most cars the voltage is not actually that horrible, most cars have small and easy to turn petrol engines that are cranked by a lead acid battery that is in resonably good condition and under normal ambient temperature. If the device can run from 6V then it likely won't even notice the cranking.

Then you get a car with a big diesel engine and a old battery and the vehicle has been sitting outside in -10°C weather. The engines run glow plugs in the cylinders when its cold and that can take a good 100A for a few seconds and after that it has to crank that engine that takes a lot of oomph to turn since diesels have high compression ratios and the thick oil in the cold doesn't really like to flow very well. All that load and the tired old battery will have its voltage drop like a rock.

We built our engine cranking test box because we make electronics for airplanes and they tend to have pretty big engines (old tech and need for a lot of horsepower) while they like to save on mass and use a small battery(They even skimp on wire gauge since copper is also heavy). So to fix this i was made to design in a tiny "supercap UPS" that keeps it running trough all of that. So a magical box was needed to repeatevly reproduce that test waveform for use in quality control in order to make sure that all the devices off the production line survive the cranking.

[ConKbot]

If you can't get a programmable power supply to keep up, you could look at SMUs if you can find one with enough current output, or some older HP power supplies with analog control inputs. Or for the big guns, a Kepco power supply/op-amp can generally do ~10KHz frequency response, if you don't want to roll your own high current buffer.

[Berni]

Oh and if you want off the shelf test gear to do it then you can pick up a Agilent 66332A.

It looks a lot like a system power supply, but has a internal buffer that can be filled up and used to update the output voltage 100k times per second. It also has a switch on the back that makes it use faster loop compensation so you get a faster responding output at the penalty of not being able to drive large capacitive loads or use long sense wires. Another important feature is also that unlike most PSUs this one can also sink current, this makes it able to produce that sharp downward edge in output voltage since it can pull the power back out of any capacitors on the power rail.

I have one of them but i mainly just use it as a 100W electronic load due to its sinking capability.