This sounds like great progress Bob, well done.
Happy holidays Chris
On Tue, Dec 19, 2023 at 10:24 PM Bob Stricklin via pacsat-dev < [email protected]> wrote:
*Here is an update on my progress checking out the QuadRxOneTx board.….*
I got the power supply working properly with the chips Zack sent. Took some time to repair the things I did to debug the issue and get the small parts soldered on the board.
When i powered up both supplies were working properly. I did add clamping Zener diodes across the outputs to provide some modest protection.
Then I had to reinsert the programming resistors on the current limiting MAX4995 which I removed to isolate sections of the board.
When I started testing one of the 3.3V supply sections was not powering up and the supplies were dropping out in a constant pattern. Found the issue was the Watch Dog was not dibbled and this was turning things off and back on in time. Disabled these. I also found I have the control logic which turns the power on and off reversed. This is another case of picking the correct logic on the watch dog chip by using the proper suffix. The one I specified is AL which I assume is Active Low and I needed active high. I disowned this and hard wire them all to be on for now. Will need to work on this more later. The watchdog chip can also be configured in different ways so all this needs to be worked when CPU is working.
With all the board functional (minus the TMS570) still had a short on one 3.3v section. After some investigation I determine the short was in one or more memory chip. The data sheet for these indicated they are particularly sensitive. So I removed all four of them. After removal and some component testing I found only one of the four was Actually shorted. However with these off the board the power supplies were all functioning properly throughout the board.
One troubling thing i noticed was with the short on the 3.3V supply if you jumped by the current limiter which I did to find the issue and removed the jumper that supply would jump up to about 3.8 volts and take a bunch of seconds to settle back down to 3.3V. This is a issue with a DC to DC regulator which is plus width modulated and adjusting to hold the output at a fixed level. Typically we will not have a short but if a temporary short were to happen in space the supply could over react and zap some other parts. Could put the clamping zener on for space use if we can find a good voltage to use but then these could become a failure point.
None of the other chips on the board seem to have issues from over voltage now that the supplies are up and working but will need to test individually to be sure. The AX5043 have never been turned on yet.
At this point with no CPU and no memory and the AX5043 all turned off the board is pulling about 25Ma of current. I should also manually turn the AX5043s on and check current tomorrow.
Tomorrow I will solder the three memory chips I took off which are not shorted back on power up to be sure things are still working.
Next, I will solder a TMS570 I have on hand and check power and function. If the board still functions I should be able to talk to the TMS570 via JTAG and load a program into flash.
While waiting for the chips from Zach I built up an interface cable to connect the 10 pin JATAG to the 20 pin DSC100 connector with the correct pins wired together. This took me about 3 hours because the wires are so small and working out the details. Due to this I laid out a PCB to support using two off-the-shelf ribbon cables from Mouser. I will post the connection info to the GitHub. I will also order enough ribbon cables to support the four board sets but need to verify what is needed for each programmer people will be using.
I have about 20 of these on order at JLCPCB which should be in my hands in about a week +- holidays.
The board is layer out so one can cut and rejumper if there is an error or another connection is need for a different programmer.
*Looking to the future…*
The boot loader and all initial code will have to run from FLASH. I have not looked at the current memory map to see if this is the case now. Jim if your blinky program will run from flash you can send it along.
Need to review the memory test program and hopefully it also runs from flash. I will want to test the three of four memory chips to see if they are working and not damaged from voltage or soldering and removing and desoldering. I will need to purchase one more MR25H40CDE memory chip to have a full set again. These chips are in the $28 dollar range from mouser so I hope the three are working. I am assuming I will be able to solder them back on the board. I believe I can but I want be reflowing the solder on the pad initially. If they get hot in use I may try to do this later with hot air system.
This is all a typical bring-up process for a board this complex. It looks like if we put replacement voltage regulators on the other boards and tune the logic on the watch dog or replace the current control devices ($5 each) which would be easy to do the other three boards should come up working.
i have also ordered 12 more of the voltage regulators from China which I hope will be working and equivalent. Should have those in a week also.
Lots more to test and checkout but if the TMS570 will work it should go faster.
Bob N5BRG
pacsat-dev mailing list -- [email protected] View archives of this mailing list at https://mailman.amsat.org/hyperkitty/list/[email protected] To unsubscribe send an email to [email protected] Manage all of your AMSAT-NA mailing list preferences at https://mailman.amsat.org
