This unit is used in a portable system to groom multiple incoming T1s coming from a constellation of TSU600e channel banks and Netvanta routers.. Because it is only used 3 times a year for special events, it sits unpowered for long periods of time.
It is the hub of a large deployable radio communications system for special events,
It grooms each incoming T1 from TSU600s at multiple remote sites into 8 E&Ms, 1 FXS, and the balance of the DS0s are aggregated into a data pipe. The pipes leave the Atlas and hit a 4305 router that is a multipoint layer 2 bridge. Each endpoint of the system has a TSU600e and a 1224 router.
I have owned this unit for about 10 years. Never a single peep of trouble.
* Dual 48V DC power supplies which are themselves supplied by a 24V to 48V step up switching power supply.
* Dual system controllers
There are no errors showing,
All module related LEDs on the front panel are green.
The map is empty. It was last powered up in January and February of 2018. It is now August of 2018.
We think that the step-up power supply overheated and shut down. There was an alarm showing on the front panel. Not sure if something that occurred in this shutdown might have scrambled the map.
It appears to be OK after allowing the step up supply to cool off, but it is as blank as if it were new from the box.
The configuration and time-of-day clock are stored in a Dallas/Mostek real-time-clock chip. At least on the 550, this chip is not socketed. It looks like an extra-tall 24-pin DIP. I'm not sure about the 890 motherboard but it's U4 on the 550. Instructions on replacement are here: Re: Atlas 550 Loses Config
It can be replaced relatively easily if you're comfortable with through-hole component level repair. I strongly recommend replacing the old chip with a socket so that a few years down the road it isn't as much of an issue.
Thanks very much for the info on the clock chip. I can do board level. I
will try to get the newer TI chip. Socket sounds like a good idea.
Once replaced, do you know what the lifespan of the chip is in terms of
retaining config and time? Will it fail faster if the 890 is unpowered for
long periods of time?
Besides the 890 I own a bunch of TSU600e, a 550, a bunch of 1224s, a 4305,
about a dozen Tracers, and a DSL (T1) HTU-R/HTU-C pair.
I love Adtran.
3rd St. R & D Production Services
Special Event and Disaster Communications Services
Technical Event Production Services
Alcons Audio North American Warranty Depot
417-336-4045 24 hrs
Surefoot Sound, LLC
Alcons Audio Sales and System Design
Surefoot Microphone Support Products
Comm/Logs - Missouri Disaster Response System, Inc.
Safety Officer 307, Communications Officer
Central Taney County Fire Protection District
I think the specification is ten years, but that's assuming that the device is powered on most of the time. Another factor is that these chips aren't in high demand so it's likely that any that you buy have been sitting on the shelf for a while. The clock is shut down from the factory in order to save battery. When you first power it up and set the clock, that "kick-starts" the chip and the battery goes active. Practically I'd expect about four or five years on the low side and ten if you're really lucky.
I recommend installing a socket so that the next time it's substantially easier.
Jayh - I told you this on the phone, but wanted to post it here for others. On the upper middle of the controller modules there is a yellow lithium battery that snaps into the chip under it. I assume this to be the clock and the backup supply for programming memory. These batteries are available from Mouser. I bought 8 to have as spares. The yellow module contains a crystal for the clock and a lithium battery.
Before replacing these batteries on both controller modules of both of my 890s, they could not remember time/date programming after a power cycle -- now they do. I am assuming that the config memory will also be protected after I reprogram the unit.
I have not had a chance to rebuild the config, but I will soon, and we'll know for sure.
Thanks very much for your assistance!
That sounds like a much better design than the original Mostek "Timekeeper" chips. Those were the same basic layout with the battery on top of the chip, but the whole thing was potted in epoxy so it looked like an extra thick IC. When the battery died there was no way to replace it except for replacing the chip. Some of the Sun Microsystems folks keeping old Sparc systems running would carefully carve out the epoxy with a Dremel tool and replace the battery when replacement chips weren't available.
To the best of my knowledge, most or all of the Adtran Atlas original chips were of the potted design, and NOT in a socket. So either someone before you modified it or there was a factory-available Timekeeper-style chip made from the factory with a replaceable piggyback battery. In either case, you got lucky.
I've got 2 Atlas 890s now (The original primary and one for parts and offline programming testing) and both of them have this style chip and snap-on battery.in both of their Controller modules.
And, I've got spare batteries in case these age out on me.
Before I discovered the difference, I had ordered 5 of the TI replacement chips and I'll hang on to them. (BQ4287MT-SB2)
I've got the config rebuilt on the primary and the system is up and running, but I have not tried to power cycle it to see if it remembers, but I will have to do that soon.
I am having problems getting the TFTP server running and being able to archive the config. I'll work on that some more. I did store screen grabs of the config so that I don't have to learn the 890 again from scratch.
It's a great piece of gear, even if clunky to program.
Thanks again for your help.