Programmable Load Build

RE: Programmable Load Build

Jan Cumps — Tue, 02 Oct 2018 17:31:20 GMT

I've done measurements with the feedback omamp controlled by 1K / 6K8 resistors.

I get the same result as you.

[View:/resized-image/__size/924x520/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-89778195-e1af-404c-a74a-5e301e05bfd3/contentimage_5F00_190676.png:924:520]

Here's the data at the knee of the graph. Column A is the DAC setting, B is the current pulled by the load.

Input 4.5 V

1020,001591030,001591040,001791050,001891060,001991070,002101080,0022101090,0023101100,0024101110,0025111120,0026111130,0028111140,002810

The difference with 100K / 680K (as expected with this low bias current opamp) is less than half a bit.

RE: Programmable Load Build

Jan Cumps — Sat, 29 Sep 2018 14:47:55 GMT

The horizontal scale is 4s (yes, 4 seconds!) per division, so it's very low
frequency noise. It's in the region where flicker noise would dominate a conventional op amp's noise, so is that what I'm
looking at here?

I will try to replicate ...

RE: Programmable Load Build

fmilburn — Sat, 15 Sep 2018 15:21:28 GMT

Hi Jon,

Thanks for describing in detail how you analyzed the circuit to discover the source of the offset current and then modified it - it was very educational for me.

Frank

RE: Programmable Load Build

Jan Cumps — Fri, 14 Sep 2018 13:51:49 GMT

[mention:4f57fc9d538949ad9eb336ddb9469bb8:e9ed411860ed4f2ba0265705b8793d05], there is a Windows GUI to play with the load, made by [mention:6e474abc33b64a82aa78c9813503c962:e9ed411860ed4f2ba0265705b8793d05]. It's available on the main blog post (link in the GUI section).

It has some handy functions:

- easy way to set DAC 1.

- it can poll all ADCs each second, either as pure value or converted to voltage (ignore ADC 4 because I don't sample that in the firmware - it's not connected and ignoring it allows me to sample the 3 other ADCs faster)

- you can shoot SCPI commands easier than with a terminal

- it shows error state of the SCPI engine.

RE: Programmable Load Build

Jan Cumps — Mon, 10 Sep 2018 13:47:11 GMT

I have an offset on the current, but it's much larger - it's around 11mA.

I'd guess it's the different opamp. I also had a higher offset when I used our first opamp choice with higher input currents.

RE: Programmable Load Build

Jan Cumps — Fri, 07 Sep 2018 17:17:36 GMT

(The pin 1 indicator is a line of silkscreen running down the side of the pin but it's so close it looks like it was cropped back by the board manufacturer.

Yes. There are inconsistencies in how pin 1 is indicated in the KiCAD libraries. I don't know if other PCB packages have more consistent libraries. I haven't tried out others.

[View:/resized-image/__size/563x613/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-89778195-e1af-404c-a74a-5e301e05bfd3/contentimage_5F00_190659.png:563:613]

From what I read on EEVBlog, most professional PCB designers make all footprints themselves.

RE: Programmable Load Build

Jan Cumps — Sun, 02 Sep 2018 10:17:36 GMT

DEVE:DAC1 0 results in a voltage of 0.391mV at the output pins of the DAC-ADC board (that's the low end of our control range - ideally it would be zero)

This is a real problem. Mine has > 0.5 mV unconnected , and also with a 10 K Resistor to ground. It prohibits that the load can be controlled in its lowest area of operation (below 1-2 mA).

Maybe we can forge an extra piece of circuit in to offset that in the control circuit. We have 3 DACs free to provide some offset, and because we're stacking the boards, there may be a possibility to plug an offset regulator board in between DAC and control circuit?.

RE: Programmable Load Build

Jan Cumps — Sun, 02 Sep 2018 08:58:40 GMT

[View:/resized-image/__size/600x420/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-89778195-e1af-404c-a74a-5e301e05bfd3/contentimage_5F00_190654.jpg:600:420]

Oops. Yes, the footprint I used for U2 is not good, and I placed the caps too close to U2.

When building up my own board, I put a warning in the original blog on how to mitigate that:

Solder advice: in hindsight I should have used larger IC packages. The board isn't easy to solder at home. But it's doable with proper care, hand stability of a snooker player and eyesight of an eagle.
It's best to leave capacitor C5 off and solder it separately when the other components (in particular U2) are mounted. Even then it may be good to move it as far as you can from U2's pins. I didn't provide enough clearance in my PCB design.
The VSSOP footprint of U2 is narrow on the V1_2 board. If you are brave, you can use pliers to bend the pins a bit down. I cut of the pins 5, 7 and 8 with a hobby knife.

My result looks Meh too:

[View:/resized-image/__size/676x381/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-89778195-e1af-404c-a74a-5e301e05bfd3/contentimage_5F00_190655.png:676:381]

RE: Programmable Load Build

shabaz — Sat, 01 Sep 2018 14:00:05 GMT

Hi Jon,

Great to see the success with CCS and the code compilation.

I too picked an Ethernet based board for a project recently, mainly for the isolation. Looking forward to seeing the progress!

RE: Programmable Load Build

Jan Cumps — Sat, 01 Sep 2018 10:40:55 GMT

If you want the latest latest version of the source, that can always be retrieved from GitHub. I commit often.

https://github.com/jancumps/msp432/archive/master.zip

The link above gives a zip archive. If you unzip that, then copy the MSP432_SCPI_ElectronicLoad folder to your workspace root, it'll update your eload source base.

It will not impact any project or CCS settings. It's sources only.

RE: Programmable Load Build

Jan Cumps — Sat, 01 Sep 2018 08:23:55 GMT

On understanding the software:

It is organised in two big chunks:

the instrument part: all things related to being an eload. It is built as an API, with all functions listed in the header file in the eload_api folder.
It has ADC, DAC, I2C, Input enable, temp control, control loop and calibrate blocks. All of them have their separate folder with an API in the header file and the implementation in a .C file.
the user interface: all things for external use. These blocks only use the eload_api functions. They know nothing else about the instrument than the functions listed in the eload API.
This part has the display code, and the UART USB interface for SCPI, the SCPI configuration and the SCPI lib source.

[View:/resized-image/__size/568x682/__key/commentfiles/f7d226abd59f475c9d224a79e3f0ec07-89778195-e1af-404c-a74a-5e301e05bfd3/contentimage_5F00_190653.png:568:682]

All is built upon RTOS. The sources are in the root directory (generated by the RTOS wizard) The only file worth looking at is main_tirtos.c. This is where we start all the RTOS tasks for the instrument.

The schedules for the tasks themselves are kept in a single file (I prefer to do that so that I have an overview of what runs how often, and what the priorities are). They are stored in the rtos schedules folder;

Because each module is contained in a single folder, and they are fairly isolated, you can start to understand the software piece by piece.

A basic understanding of RTOS and RTOS tasks/messages/events helps.

Keep the controls sources for latest. That's the most complex part of the code because it allows to plug in constant voltage, current, resistance, power, .... control strategies.

We only support constant current at the moment and because that's implemented in hardware, there's not a lot to see here.

RE: Programmable Load Build

Jan Cumps — Sat, 01 Sep 2018 07:34:08 GMT

pro tip

Next thing was to download Jan's zip file, with the software project in it, and unzip it to a folder under the CCS
workspace directory.

you can directly import a zip file in CCS by using the Project -> Import Project menu, then select the Archive radio button. this will automatically extract the project into your workspace and add it to the CCS project view.

RE: Programmable Load Build

jw0752 — Sat, 01 Sep 2018 05:35:55 GMT

Hi Jon,

I look forward to following your progress.

John

RE: Programmable Load Build

fmilburn — Sat, 01 Sep 2018 01:18:34 GMT

Glad that you are building this and blogging as I have it on my list of things I want to do also.

RE: Programmable Load Build

genebren — Sat, 01 Sep 2018 00:53:59 GMT

Nice blog detailing your progress. This a very interesting project, one that I keep thinking about taking on, so I will be following along to see how it goes.

Keep up the good work.

Gene