What should be possible with programmable power supply?

[prasimix]

I'm compiling a list of special cases for which a dual programmable power supply could be used. Until now I have the following:

1. Battery charging in various modes (pulsing, CC, CC+CV) with and without temperature sensor monitoring (if one exists on battery)
2. Zener voltage measurements for Zener diodes
3. Simple "arbitrary waveform generator" (variable voltage and current output)
4. Data logging (voltage, current, power)
5. "Tracking" or "coupling" when one channel output follows changes on another channel (voltage, current, power). In this mode of operation activation of any protection (OVP, OCP, OTP, OPP) on one channel will also switch off other channel's output.
6. Display output value as YT graph (similar to scope, the time base is of course limited with speed of display scrolling and ADC sampling rate)

Please let me know if you have anything to add. Thanks.

Your suggestions on one place:

7. [Kleinstein] entering stand-by when any of protection tripped
8. [Kleinstein] software PID-regulator using load temp sensor input (10KOhm)
9. [Kleinstein] output sequencing
10. [chickenHeadKnob] battery simulator

[Kleinstein]

Sometimes a kind of electronic fuse mode can be nice. So not only limit the current to a hard value, but turn off the supply if the current is higher than a set limit for a longer time (e.g. 1-1000 ms range).

If there is an external input anyway (e.g. temperature sense for charging), there could be the option for a slow (e.g software controlled) PID regulator.

A special case for the slow waveform function could be a defined power on sequence of the two channels.

[bitslice]

Simulating noise on one or more channels would be handy for checking RS232 immunity

As would simulating a variable brownout for checking a microprocessors reset system

[C]

Acting like any power supply.
For you working on a circuit, you need specific settings. To computer test a circuit you need to test with all possible settings to see if there is a fail.

If you can computer test a power supply and record the results. Then you have a chance of testing the circuit using your supply as if it was connected to that power supply.

[Cyberdragon]

You are making me jealous. Mine has corrupted memory and I can't seem to get it to understand commands, just says "syntax error". I need a proper GPIB controller instead of an arduino.

To add insult, the power line filter exploded.

[chickenHeadKnob]

Perhaps battery discharge curve simulation, for various battery chemistries, although this could be tricky. I suspect most people developing battery run portable products use actual intended batteries however I could see the value of having a standard (simulated) battery to test against. I don't know how you would best implement it as it would seem to require coulomb counting.

[Berni]

I have a supply that can do some of that stuff if you write some software.

Its a 66332A Dynamic Measurement DC Source. Its a cheap 2 quadrant SMU with a fast DAC. It can playback a waveform on the output with a sample rate of 100KHz and can do reasonably fast measurement of the output voltage and current. Since it can sink and source current you can use it as a "virtual battery" when testing battery chargers.

But the fan in it is loud even tho its speed regulated because they used a tiny heatsink.

[prasimix]

Sometimes a kind of electronic fuse mode can be nice. So not only limit the current to a hard value, but turn off the supply if the current is higher than a set limit for a longer time (e.g. 1-1000 ms range).

Makes sense. The shutdown of complete PSU is already added when e.g. any of POWERGOOD signal failed. This one can be added instantly. Accompanying SCPI command could be SYSTem:POWer:PROTection:TRIP <bool>

If there is an external input anyway (e.g. temperature sense for charging), there could be the option for a slow (e.g software controlled) PID regulator.

OK

A special case for the slow waveform function could be a defined power on sequence of the two channels.

Good. Instead of implementing that as a special case of what is mentioned under #3 in initial post, that could be achieved with software triggering with delays between turning on different channels outputs.

[prasimix]

Simulating noise on one or more channels would be handy for checking RS232 immunity

As would simulating a variable brownout for checking a microprocessors reset system

Don't know how to simulate noise. The power supply is not that fast. But simulating variable brownout could be accomplished with #3. Thanks.

[prasimix]

Acting like any power supply.
For you working on a circuit, you need specific settings. To computer test a circuit you need to test with all possible settings to see if there is a fail.

If you can computer test a power supply and record the results. Then you have a chance of testing the circuit using your supply as if it was connected to that power supply.

If I understand you correctly that looks like great combination of #3 and #4.

[prasimix]

Perhaps battery discharge curve simulation, for various battery chemistries, although this could be tricky. I suspect most people developing battery run portable products use actual intended batteries however I could see the value of having a standard (simulated) battery to test against. I don't know how you would best implement it as it would seem to require coulomb counting.

This is a good one! If we can measure output voltage and current, than we can measure to good extent also power and energy e.g. in Ah. Hopefully usable battery discharge curve simulation exists somewhere on the Net that can be used for this.

[prasimix]

You are making me jealous. Mine has corrupted memory and I can't seem to get it to understand commands, just says "syntax error". I need a proper GPIB controller instead of an arduino.

To add insult, the power line filter exploded.

Hi Cyberdragon, didn't got what's happened to you and which power supply you have. Whatever is the case I think that for reliable programming you also need proper software not just interface. What arduino has with that?

[prasimix]

I have a supply that can do some of that stuff if you write some software.

Its a 66332A Dynamic Measurement DC Source. Its a cheap 2 quadrant SMU with a fast DAC. It can playback a waveform on the output with a sample rate of 100KHz and can do reasonably fast measurement of the output voltage and current. Since it can sink and source current you can use it as a "virtual battery" when testing battery chargers.

But the fan in it is loud even tho its speed regulated because they used a tiny heatsink.

Thanks Berni, this is a piece of equipment. In this moment I'd like to limit programming scenarios to the first quadrant only despite the fact that with down-programmer circuit we are entering second one. 2 or even 4 quadrant PSU is a good candidate for the next project 

[C]

Acting like any power supply.
For you working on a circuit, you need specific settings. To computer test a circuit you need to test with all possible settings to see if there is a fail.

If you can computer test a power supply and record the results. Then you have a chance of testing the circuit using your supply as if it was connected to that power supply.

If I understand you correctly that looks like great combination of #3 and #4.

Not quite.
#4 from your power supply would let you replay a problem the connected load had to see if problem is curred.
If #4 was from a load connected to my very poor power supply would let you test to see if your load would function on my power supply. Much harder to do as my power supply gets better.
My very poor power supply could be a cap & diode with the cap too small or way to large at low end.  Could be better then this. Easy to record, harder to create a computer program to create same info.

Note that #3 would need more than one output so that your power supply output could switch from CV to CC for example.

new
Your positive output acts like a LM317 or LM334.
Easy to see a LM334 as a current load or a current regulator.
Little harder to see a LM317 used as a voltage load.
Might look at changes needed for your power supply to act as a load. With two you could then test one against the other.

[jeroen79]

Add inputs and outputs (digital and analog) so that users can add their own things to it.

Or give it a user programmable microcontroller that talks to the main controller.
Then users can write whatever they want for it.
Kinda like a PSUduino.

[prasimix]

Acting like any power supply.
For you working on a circuit, you need specific settings. To computer test a circuit you need to test with all possible settings to see if there is a fail.

If you can computer test a power supply and record the results. Then you have a chance of testing the circuit using your supply as if it was connected to that power supply.

If I understand you correctly that looks like great combination of #3 and #4.

Not quite.
#4 from your power supply would let you replay a problem the connected load had to see if problem is curred.
If #4 was from a load connected to my very poor power supply would let you test to see if your load would function on my power supply. Much harder to do as my power supply gets better.
My very poor power supply could be a cap & diode with the cap too small or way to large at low end.  Could be better then this. Easy to record, harder to create a computer program to create same info.

Note that #3 would need more than one output so that your power supply output could switch from CV to CC for example.

new
Your positive output acts like a LM317 or LM334.
Easy to see a LM334 as a current load or a current regulator.
Little harder to see a LM317 used as a voltage load.
Might look at changes needed for your power supply to act as a load. With two you could then test one against the other.

I'm now even more confused . Any chance to make simple diagram with all elements of such test environment?

[prasimix]

Add inputs and outputs (digital and analog) so that users can add their own things to it.

Or give it a user programmable microcontroller that talks to the main controller.
Then users can write whatever they want for it.
Kinda like a PSUduino.

Ok, maybe a plug-in module with MCU that has properly isolated/protected inputs and outputs?

Instead of PSUduino maybe is better to think about open-lab-duino, a modular design that can be maintained using software framework that also include many ready-to-use modules for accomplishing various T&M (test & measurements) tasks?

[C]

Your power supply's purpose is to help you design, build and test circuits.
What Is the circuit that is connected to your power supply going to have for it's finished power source?

You design and test in LTSpice, but this is not real world. From LTSpice, do you jump strait to 9v battery power or do you have your power supply act like a 9v battery where you have more control.
Think of the steps, LTSpice, your power supply, battery.
Would be very handy to be able to control your power supply so that is looks more and more like a battery.

If that 9v battery is to last a year, you have a long test to run. Do you pick just a few points and test. You could have missed a test and not know it. With your power supply acting as a battery you could speed up time and still check the full range. 

Power Supply Acts like a battery. Very simple statement and very very hard to do.
By acting like a ______, you gain control so you can test the connected circuit for proper operation with _______ as power source. 

One specification in a data sheet is power supply rejection ratio or PSRR.
The noise PSRR is trying to remove changes by power supply & type.
A 0 noise power supply does not let you check/test PSRR.
Again the best noise to add back in is the noise that could or will happen with ______ as a final power supply.