I have modified a CCR2004-1G-12S+2XS to run directly on a DC power system, bypassing the original internal AC-DC power supplies. I’ve added a DC jack connected to the internal power headers on the motherboard.
My DC UPS (battery-backed) system operates at a float voltage of 13.5V to 13.8V.
Since this specific CCR model originally comes with 12V internal power supplies, I would like to confirm if the motherboard’s internal regulation stage (DC-DC converters and capacitors) can safely handle a constant input of 13.8V long-term.
From my assessment, most components seem to be rated for at least 16V or 25V, but I want to ensure there are no hidden risks regarding the fans or SFP+ port power rails.
Has anyone else run this specific model on a 13.8V battery system? Any official word on the maximum DC input tolerance for this board?
Components (capacitors) are 16V or 25V because those are the common reference values for them, 2.5/4/5.5/6.3/10/16/25/…
But without checking the circuit, a board intended for 12V should NOT be powered at 13.8, it is 13.8/12=1.15 or 15% more, whilst usually the tolerance on low voltage is +/- 5% or at the most 10%.
Mind you, most circuits nowadays labeled 12V do work nonetheless at 13.8V, but there are exceptions.
There are three possibilities:
the board will run at overvoltage just fine
the board will run at overvoltage BUT some components will heat more AND live less
the board will immediately fry
Since I presume that you already tried it and #3 did not happen (good ) it is 50% #1 or#2.
I would get one of those (intended for automotive use) converters/stabilizers, example:
It’s not just the capacitors, you need to check if every component can handle this voltage.
That means every capacitor, diode, switching regulator etc…
Good thing is that if power converter inside can handle increased voltage it will in fact draw less current from the battery and maybe be a little bit more efficient.
One thing to consider is, battery is a crude, high current source so in the case of some kind of failure your board would be done…
What would I do is I would buy ISOLATED buck/boost converter so you cover all the battery voltage range. (For Lead acid this is from 14.5 to 10.8V)
Buck or step down converter would be good if you have 24V battery system.
Not necessarily. A "random source" says that if step ratio (in DC-DC conversion) increases, then efficiency decreases. And that's even if current (on input side only!) decreases due to higher voltage used.
I agree with those saying that it would be best to actually provide nominal voltage to the board (as suggested, use buck/boost converter) and definitely add an appropriately rated fuse on board input (and possibly another one before converter to prevent from excessive damage in event that converter itself fails in a spectacular way).
That’s why I said maybe, but for that small input voltage change efficiency change is negligible.
On topic, does your UPS regulate output voltage to 13.5-13.8 V or does this voltage change with the battery charge state ?
Maybe it would be a good idea to put another DC/DC converter at the output of the power supply so you have stable 12V at the full voltage range of the battery.
I know a guy who does this for his CCR2116’s (and similar gear) at his 12V two-way radio sites when the router/switch doesn’t support 12V natively. I don’t know of any that have failed in the past 3-5 years that they’ve been deployed.
If you’re concerned, just get a small DC/DC converter that has a wide input (like 7-24V) and narrow output (12V). Easily found on Amazon or via distributors like Mouser and Digikey (or your local favorite), and they’ll fit where the original power supply sits.
What happens if the UPS does the battery charging wrong and puts up 14,4V like it is needed for some lead acid batterys?
You could at leats put a Zenerdiode parallel or a normal diode in Series, that protects against wrong wiring (+ to - and vice versa) and the 0,7V voltage drop of the Diode helps to have the somewhat right voltage.
A wide input DC-DC converter with 12V out, suggested earlier, is the cleanest solution, instead of guesswork about inputs and required passive components for voltage drops.
So far in this thread I have seen 12V, 13.5-13.8V or “what if it puts up 14.4V”. Why guess? Diodes in a saturated state have a fixed voltage drop, not catering to input variations.
DC-DC converters (switching regulators) are a well established and reliable technology giving the desired output from a range of inputs. Further, in this high current situation they will generate less heat, waste less power.
Not only, the kind I suggested are built to be used on RVs/cars/golf-carts, etc, they are enclosed in resin, have an aluminium shell that allows heat dissipation, they can be directly bolted to the chassis without need of washers/insulators/spacers, etc., and they are as robust as they can be both mechanically (resisting to vibrations) and electrically (as on battery powered vehicles there can be spikes when cranking or similar).
If stranded on a deserted island, and with a KBPC5010 available, using two diodes out of it is a good idea, though.
Sure you could MacGyver your way and use only one diodes of the 4 available or two (in parallel) if you MacGyver even more and strap both ~~ pins togheter, but why x.x
You can't use "just two" in series from the 4 available.
Anyway, it was just an option.
I stand corrected, you can "use" only two if you use the bridge rectifier as a .. bridge rectifier, and you feed it DC instead of AC, but that wasn't the way I've mentioned it to be used, I was suggesting to only use it on the positive rail and leave the negative directly connected. That way you still have the same voltage drop but the load is less via every diode since it's split into two paths.
) it is 50% 
This is 100W version, i believe there are lower power ones, 12V 5A is original PSU.