Without knowledge of the UPS how can we tell? Have you asked the manufacturer?

I did just now. This was also a good idea.

 

Without knowledge of the UPS how can we tell? Have you asked the manufacturer?

A manufacturer representative just answered me the following:

Unfortunately we can not say if the power reduction is the problem. We advise against such changes, please use the original fans.

I kind of expected that, that's why asking the manufacturer didn't appear to me as viable option from the beginning. Anyway, I tried.

 

The resistors arrived. The error remained. Thank you for your help.

 

Hello,

It looks like there is a sense for the amount of airflow.
Use fans with the same or more airflow as the original fans.

Bertus

 

Like I mentioned in the beginning, the UPS works fine with two new fans in the case replacing the old fans and one old fan outside of the case and not cooling anything. (All fans are connected in parallel.) Hence I believe that the error gets triggered purely electrically and only from the fans themselves and not from any other sensors.

 

Does the fan have a yellow wire?

 

Does the fan have a yellow wire?

The original fans have no yellow wires.

 

For now I put the external fan in a Tupperware box and placed the box far away from my working space.There is a cable between the fan box and the UPS. The UPS doesn't complain anymore. A Stone Age solution for lack of a better one.

 

For now I put the external fan in a Tupperware box and placed the box far away from my working space.There is a cable between the fan box and the UPS. The UPS doesn't complain anymore. A Stone Age solution for lack of a better one.

If it works, great.

Maybe the UPS is measuring brushed fan motor back-EMF during a dead-time slot. A resistive load won't work if this is the case.

https://www.precisionmicrodrives.com/content/ab-021-measuring-rpm-from-back-emf/

 

Very interesting way of getting feedback from motors without additional sensors. Thank you.

 

@nsaspook Can we chat via conversations? You have to initiate the conversation. One initiated, I have that as a medium (same conversatio forever) untill all participants pull out. I know we have similar interests. I worked in a semiconductor research lab. Think about it. I'll clobber this if a get a PM or late Saturday night if I remember. If you tell me not more PM's. then no more PM's. Thanks!

 

To be quite honest, I think you're going to need a o-scope on it, and you can use a clamp-on ammeter to get current easily. The waveform without the 3rd (original) fan, and then with it, comparing the two is probably the easiest way to determine what's going on.

 

To be quite honest, I think you're going to need a o-scope on it, and you can use a clamp-on ammeter to get current easily. The waveform without the 3rd (original) fan, and then with it, comparing the two is probably the easiest way to determine what's going on.

Oscilloscope settings (amateur configuration; tried until the image was relatively stable and until it fit into the scope screen in a meaningful way):



Without original fan (details):


Without original fan (overview):
With the original fan (shot 1):

With the original fan (shot 2):

When the original fan is connected, the lines below the x-axis move while the ones above are relatively stable. That's why I took two shots.

 

As suspected you are seeing what looks like EMF spikes from the original fan during dead-time and none without it.

 

As suspected you are seeing what looks like EMF spikes from the original fan during dead-time and none without it.

Do you know of any silent 8 cm fan that works like this besides the original?

 

Cut the blades off an OEM model fan. There wont be any airflow. For fun and giggles, PWM modulate the speed and see how low you can go.

 

Cut the blades off an OEM model fan. There wont be any airflow. For fun and giggles, PWM modulate the speed and see how low you can go.

I did cut the blades. The fan started spinning faster. This caused the UPS to beep again.

 

I did cut the blades. The fan started spinning faster. This caused the UPS to beep again.

Which models the behavior of an old school DC fan motor without electronics.

I've had this problem before with older devices and obsolete fans. (electronics paleontologist is one hat I wear at work ) One solution I used was to find and isolate the fan 'GOOD' signal on the original PCB and then use a microcontroller to convert the new three wire fan signal(s) into a pass/fail dry contact to provide the needed emulation function as needed.

Prototype board
https://github.com/nsaspook/fanmon_xc8

 

Possibly by using the power of the rotor I could cut them in the same way a lathe works. The blades are made of hard plastic, though. This would require a precision construction in order to avoid the rotor becoming off-center. I am wondering if this is worth it.

Which models the behavior of an old school DC fan motor without electronics.

I've had this problem before with older devices and obsolete fans. (electronics paleontologist is one hat I wear at work ) One solution I used was to find and isolate the fan 'GOOD' signal on the original PCB and then use a microcontroller to convert the new three wire fan signal(s) into a pass/fail dry contact to provide the needed emulation function as needed.

Prototype board
https://github.com/nsaspook/fanmon_xc8

Thank you. Very interesting.