Switching the fan stored preferred fan direction for JG552A fans

[Spork Schivago]

Hello,

My HPE 5900AF-48G-4XG-2QSFP+ switch can have two types of fans (excluding normal and high performance):

Ones that blow the air from port to power, ones that blow the air from power to port.

I have two of the high performance JG552A port-to-power fans. The ones I want are JG553A, which blow from power-to-port.

The physical brushless fans inside are the exact same, but the wiring is switched around a bit. So I flipped the fans around and switched the wiring so it matched the JG553A's, and sure enough, now the fans blow the way I want, from power to port.

However, there appears to be some place on the circuit board for these fans that store the information on how the fans are supposed to blow, and I'm constantly getting an error message saying something along the lines:

DEV/1/FAN_DIRECTION_NOT_PREFERRED: Fan 1 airflow direction is not preferred on slot 1, please check it. 

DEV/1/FAN_DIRECTION_NOT_PREFERRED: Fan 2 airflow direction is not preferred on slot 1, please check it.

I display the fan configuration, and sure enough, it does show that I have them set to power-to-port, but they're preferred direction is port-to-power.

I was going to upload some pictures of the circuit board to see if anyone saw a chip that might contain the data, where I could just switch it around. A kinda cheap hack. Has anyone ever done anything like this before?

I couldn't find any mods on the net. The proper fans will cost me over 600$, so it'd be nice if I could save that money and just change something on an eeprom or something.

Thanks.

[Spork Schivago]

Re: Switching the fan stored preferred fan direction for JG552A fans

Found a Microchip 24LC128 128Kb I2C compatible 2-wire Serial EEPROM on the board.

I believe this is where it'd be stored. Looking at the features though,

Single Supply with Operation down to 1.7V for24AA128/24FC128 devices, 2.5V for 24LC128devices
Low-Power CMOS Technology:- Write current 3 mA, typical- Standby current 100 nA, typical
2-Wire Serial Interface, I2C™ Compatible
Cascadable up to Eight Devices
Schmitt Trigger Inputs for Noise Suppression
Output Slope Control to Eliminate Ground Bounce
100 kHz and 400 kHz Clock Compatibility
1 MHz Clock for FC Versions
Page Write Time 5 ms, typical
Self-Timed Erase/Write Cycle
64-Byte Page Write Buffer
Hardware Write-Protect
ESD Protection >4000V
More than 1 Million Erase/Write Cycles
Data Retention > 200 years
Factory Programming Available

That hardware write-protect has me a bit worried that it might be write-protected and that I could not modify it. Finding where the actual data is stored is going to be another problem.

I highly doubt I can find someone who has a dump of their JG553A EEPROM for comparisons.

[Spork Schivago]

Re: Switching the fan stored preferred fan direction for JG552A fans

If anyone is actually interested, I found two new (other) (to me, that usually means used!) power-to-port high performance fans for a buck a pop. I bought two. I can dump the firmware, see if it's write-protected, and maybe find where the serial info is, and hopefully where the bit is (or data) that says port-to-power / power-to-port. Might just be a model number thing. if(model==strcasecmp(JG553A)) { /* ... configure power-to-port ... */ } else { /* ... configure port-to-power ... */ }; kinda thing, you know?

[diif]

Re: Switching the fan stored preferred fan direction for JG552A fans

Why the need to have the fans the other way ?
Have you seen this ? https://www.amazon.com/HP-Network-De.../dp/B00BNR1NI2

[Spork Schivago]

Re: Switching the fan stored preferred fan direction for JG552A fans

Why the need to have the fans the other way ?
Have you seen this ? https://www.amazon.com/HP-Network-De.../dp/B00BNR1NI2

I bought two of those same units from another source, new, and only paid 99$ total for both, (50$ a pop).

It's because of how the rack is now laid out. The rack moved to a new position. It's actually better where it is now anyways for many reasons. Anyway, it made more sense to have the ports on the rack facing the back of rack, and the switched installed in the back, seeing how that's where the entrance for the cables come into the rack. Didn't want hot air blowing against the other devices in the front. Now the front devices blow the hot air out the back, the back devices blow it out their front (which is the back). air flow all going the same way.

Still gotta work on how to remove all the heat once I get the grounding kit for the rack. Still need to find sides. They've disappeared and my main VAR cannot get them. Got the 36 unit rack, not a 42 unit. It's the HPE 11636 1075 mm pallet universal rack (HPE Part Number: H6J77A)

The side kit I need should be HPE part number: BW898A. Gonna give RackMount Solutions a call and see if they have any still.