[CB_HK]

UPDATE

I have solved my issue with the Dallas RTC inside my Crimson by disconnecting the internal lithium batteries and attaching a new coin cell battery holder with leads. In total I spent around one hour working on this project including setup and tear down. I am happy to report that it worked perfectly and I have now solved the NVRAM error that was popping up during POST.

Below is a quick walk-through of how I modified the Dallas chip to function with an external battery. As a disclaimer, I did not create this process. I know many people have already had experience with this issue and with that in mind I spent a few days reading up on the best way to solve this problem. In the end I referenced a YouTube video by a gentlemen who was dealing with the exact same issue (Link).

The Dallas RTC inside the Crimson is a bit different from the more common models. This one is a DS1216 and is designed to hold the SRAM chip with environment parameters on top. A variant of this chip used in other computers is for ROM instead of SRAM.



Here you can see the SRAM removed from the RTC, with the DS1216 IC in view. The pins on the bottom of the RTC package are fragile and easily bent.



The first modification to make involves using a dremel with a fine tip engraving bit to bore small holes near pins 8 and 14. The goal is to sever the thin connection bars that link the lithium batteries that are potted into the chip so that when you wire up the coin cell battery you don't run the risk of damage.



Once the batteries are disconnected, flip the unit over and use the same engraving bit to cut away a small amount of the plastic frame covering pins 4 and 8 (left to right). Pin 4 will be the positive hookup and pin 8 will be the ground. I recommend using the smallest point on your soldering iron in order to then tin the pads.



Next I connected the two leads to pins 4 and 8.



While there are other, smaller battery holders, the IO3B has a nice empty spot near the front where this Adafruit battery holder can sit.



Once the connections are complete for the IC and the battery holder, carefully reinstall the RTC and SRAM. I used Kapton tape in a few areas to ensure wires were held in place or that any exposed solder joints were insulated.



Below you can see where I mounted the battery holder and how I tacked down the wire. There's not a lot of room between the IO3B and the IP17 boards when mounted in the chassis so I mounted the battery holder flush to the PCB.



Everything reinstalled and voila! No more NVRAM error AFTER I reset the environment parameters. The first time I booted it gave me a parameter error (which was expected).  Next up: solve my SCSI controller issue!



Original Post

I decided to set up a serial console with my Crimson to poke around a bit and see how things were going. The real goal was to make sure my console setup was working correctly before I attempted to hook up to a VGXT board set and make sure it's operating correctly. While the VGXT passed all of its power-on tests, I noticed that the Crimson was outputting an NVRAM error to the serial console but not to the main display.

Any thoughts on what the issue might be? I attempted to fix the issue with the Command Monitor but no luck yet.

Code:

Non-volatile RAM checksum is incorrect:
Initializing the non-volatile RAM parameters.
From the Command Monitor, check with printenv and fix with 'setenv ENV_VAR'
where ENV-VAR is a non-volatile environment variable such as bootmode.