EXTECH Quad Power Supply Upgrade


October 23rd, 2016, My Dad got a few power supplies at a battery manufacturer bankruptcy auction, this EXTECH Quad Output DC Power Supply (Model 382270) had some problems so he gave it to me.


November 13th, 2016, Finally got some free time to get the power supply on the work bench.
All the screws on the power supply are heavily rusted/corroded. This is most likely from being exposed to corrosive vapors during battery manufacturing.

Testing the power supply found that the channel 4 voltage pot and channel 2 current pot where intermittent.
Also, this video points out the main problem with this power supply, you can't adjust the voltages before enabling the outputs:

Next, I opened up the power supply and put the scope on the output from the pot to verify it was the problem and not something else...
And it fixed its self!
The wiping action of the pot cleaned off the corrosion and everything started to work properly. I even got it on video:


November 19th - 27th, 2016, Now that the power supply was working properly...
I began my quest to upgrade the power supply so you can adjust the voltages before turning on the outputs.

My design requirements for this project:
1. Adjust the voltages and have them displayed, while the outputs are disabled.
2. Not effect any other features of the power supply, series/parallel/isolation/current limit, etc...
3. Independent control of all four power supply outputs.
4. Simultaneous all ON/OFF.
5. Store and recall last state.
6. Programmable ON/OFF cycle times.

Disassembly of the front panel, and installation of the momentary push buttons.

       

Originally I wanted to leverage the fact that there is already an "Output ON/OFF" button...
But after a bit of reverse engineering I found that this would not work, since it just controls the AC side transformer tap relays into the four linear regulators.


Using parts that I have on hand, I started to put together a design. Including using one of my RTMCS2 boards as the power supply controller:

While I had the front panel off, I traced out the power path to the terminal posts...
The output terminals are bolted to a copper pour that and has multiple traces going off to the feedback, mode relays (series/parallel/independent), and voltage sense circuits on both sides of the PCB.
It did not look like it would be easy to modify the circuit board by cutting the correct traces to isolate the output terminal post...
So I did the next best thing and modified the output terminal posts to isolate them from the PCB.

Removing the shoulder and part of the threads from the output terminal on the lathe.

   

An out of focus (I could not get my cell phone to focus when the lathe was running) video showing the modification process:

This picture shows the stack up assembly of the modified terminal:

With the shoulder removed and the heat shrink tubing over the removed threads the terminal is no longer in contact with the PCB.
After inserting the terminal into the power supply, a ring terminal is installed, then an insulating nylon washer, then a second ring terminal, and finally the nut.


The first ring terminal contacts the PCB and the second ring terminal contacts the output terminal.
This lets the relay switch the output on and off.

This video might explain it better:

Wired up the ring terminals (using wire from a old computer power supply) and installed the relays:

       

I needed to find power for my microcontroller and the relays...
A bit of probing with the volt meter found that the using the blue wire as the common, I get 21.5V and 10V (half rectified with no load DC).
The 21.5V is enough to activate my 24V relays.
And then 10V supply will give me enough over head into my 3.3V regulator to be able to detect a power loss and save the state of the outputs to a byte in flash.

   

Next I sized a piece of protoboard that looked like it would fit all the parts and built up the power controller.

       

Video explanation and overview:

After mounting the power controller to the back wall of the power supply, I wired up everything.

       

Then I wrote some code to test out all the hardware... and it does not work.
My microcontroller can only sink and source 12mA max, and it was not enough to drive the transistors in the optoisolators into saturation (needed around 20mA).

Video showing the problem:


December 3rd, 2016, I disassembled the power controller, and removed the optoisolators (CNY17F-3)...

   

I decided to use some logic level surface mount MOSFETs (IRLL014N) that I had, in stead of the huge overkill MOSFETs (STP75NF75) I mentioned in the previous video (I want to save the power MOSFETs for a motor controller).

I then converted the surface mount MOSFETs to through hole by soldering on some wire wrap pins:

     

Rebuilt power controller now MOSFET powered!

   

Video showing the new power controller:

Power controller now installed in the power supply.

And a video showing it working:


December 9th, 2016, I finished the code.

When the ADC detects that the input voltage to the 3.3V regulator drops below 6.0V, it will cause an interrupt which saves the state of the outputs to a byte in flash.
Then on start up it reads that byte to restore the outputs to the state before the power supply was turned off.

6.0V was chosen as the trip point, since I measured the power supply dropping to around 7-ish volts, when going from no load to full load, while using that transformer tap.

I also added code to allow power cycling at arbitrary ON and OFF times.
To enable the cycle feature you press and hold the desired supply enable button (It will start flashing).
Then when you let go of the button the supply will enable... (The power controller is now counting how many milliseconds...)
This programs the ON time.
Pressing the button again turns off the output... (The power controller is now counting how many milliseconds...)
This programs the OFF time.
Now pressing the button again causes the power controller to replay the ON/OFF cycle, until you press the button again.

Max ON and OFF times = 2^32 / 1000ms / 60s / 60m / 24h = 49.71 Days. That should be more than enough. ;)

Video showing the new features in action:

Finished modified EXTECH Quad Output DC Power Supply (Model 382270):

Perhaps I will write some code to control it over USB in the future... ;)


More EV Projects