I work on telecoms and we have a brand of cabinets with built in refrigeration that relays in 2 temperature sensors.
However the manufacturer only gives us a partial admin credential to commissioning the cabinet. And that admin privileges don´t let us near the PWM configuration of the fans (which only goes as high as 70% DC regardless the built in high temperature.)
Our partial solution was to plug to +12V the PWM signal wire which makes the fan go at full speed all the time (a BRUSHLESS fan) but in the long term this causes overheating of the wingdings and eventual failure of the fan. (Tough it has a PTC thermistor in the (+).
So... leaving context behind an as an electronic amateur I was wondering if I can use the 555 chip to "modulate" the original PWM signal of the cabinet. So for example for a 10% original duty cicle, the 555 chip will "multiply it to an output of a 30% duty cicle for example.
I think that I can use the trigger or the reset pin of the 555 to accomplish this task, and first I will need to know at which frequency the cabinet´s PWM signal is operating right?
I would be glad if anyone more experienced or willing to help can throw a bone here
Thanks in advance. And sorry for the language, not native english speaker.
A 555 can produce a PWM signal based on the incoming signal, but its duty cycle can only add to the incoming pulse.
You would basically set it to monostable, and use the PWM pulse of the fridge's signal as trigger for the 555. The 555 output goes high when the incoming pulse goes high, and low a certain time after the incoming pulse goes low.
This is where an Arduino can come in: it can measure the PWM signal of the fridge, and produce a signal with the same frequency but increased duty cycle.
However it's doubtful whether that's a good idea as you said already your fan will overheat when running at 100% duty cycle. There's probably a good reason the manufacturer has limited the fan to 70%.
The problem with the original duty cicle at 70% is that it reaches that when the temperature is already gone haywire, and with 70% DC you can´t bring down the temperature inside fast enough. (Sometimes the NOC (Network operations center) calls us because the temperature is more tha 80° C and we go to the site, open the cabinet for half an hour and then leave the fan at full speed.
The best would be to be able to configure the ELTEK (the brand of the system) I/O interface and lower the value of the temperature in wich the Duty cicle starts incrementing. But as said, that´s not possible.
However newer cabinets for the 4G network from the same ELTEK brand have a built in dip-switch in which you can add 2 more speeds to the fan that is sum up to the existing speed it was already running. (Maybe they corrected the bug in this version?)
My first choice was arduino. But it need to be cost effective. We have 1000 telecom sites around the country and almost all have an average of 3 cabinets. I cannot buy 3000 arduinos to do the task.
I need that for a PWM of let´s say 200 HZ at 10 DC. That the 555 will take this as input, and in the output give 200 Hz 20% DC for example.
Might as well take the 12V DC and use a step-down (DC to DC) converter (buck converter) of suitable power capability to bring it down to 10V or whatever you want...... and then use that to drive the fan.
This is assuming that the "PWM signal wire" that you mentioned is simply the fan's DC voltage supply input.
Fabius_rex:
Our partial solution was to plug to +12V the PWM signal wire
Southpark:
Might as well take the 12V DC and use a step-down (DC to DC) converter (buck converter) of suitable power capability to bring it down to 10V or whatever you want...... and then use that to drive the fan.
This is assuming that the "PWM signal wire" that you mentioned is simply the fan's DC voltage supply input.
It´s not. The input it´s not analog, it´s digital. Already tried a voltage divider and it did nothing.
Fabius_rex:
My first choice was arduino. But it need to be cost effective. We have 1000 telecom sites around the country and almost all have an average of 3 cabinets. I cannot buy 3000 arduinos to do the task.
Just spitballin' :
Would it work to make up a dedicated barebones board with just the ATMEGA328 chip and its oscillator circuit?
Is it possible/feasible to let one Atmel control three cabinets at once? That would cut cost.
Or use an ATtiny. At these numbers you'll be able to roll your own dedicated PCB for under USD 1 a piece. Furthermore, is not the only microcontroller manufacturer around, there are much cheaper ones available (the HC-SR04 sensor has a microcontroller on board, and the whole sensor sells for about USD 0.60-0.70 a piece).
The thing is, a Pro Mini can be bought for about USD 1.5 a piece, making the hardware cost of your project USD 4,500. Small and easy to program. Your cost of custom engineering a solution and having your own PCB made is probably going to cost more than that. Indeed a single one can handily take care of several fans at the same time.
Fabius_rex:
It´s not. The input it´s not analog, it´s digital. Already tried a voltage divider and it did nothing.
Attaching a scope would give you a much clearer picture. At this size of operation you must have those things around (or the budget to buy a USD 300 DSO Quad or similar basic scope).
Fabius_rex:
It´s not. The input it´s not analog, it´s digital. Already tried a voltage divider and it did nothing.
Yeah...... better double-check to see what that particular line/cable is meant to be --- just in case. In any case, if that cable is voltage supply input for the fan, then the voltage being applied (to the fan) should never be the output of a voltage divider - because a voltage divider output voltage changes with output current.
Designing a retrofit hardware device to solve this problem is of course a fun thing to do and there are many alternative solutions to look at, but ultimately if the bought in system is inadequate (in this case cooling capacity is insufficient) then I'd tend to think of that as a problem for the manufacturer to solve.
You have to consider that any modification you make could be rendered invalid if there is a software or other upgrade on the ELTEK systems.
You also have to think of what happens if, say, one of these things catches fire and it is discovered that some unapproved modifications have been made. It might not look too good.
jackrae:
A cooling fan that overheats when running at full speed ? - something not quite right there.
I'm scratching my head too.
I've got a simple solution to double the airflow from a PWM-controlled fan: TWO FANS!
If the system goes out of control with the control system functioning, then there simply isn't enough cooling power available to do the job. If it's some other kind of problem (eg. fans don't re-start automatically after a power outage) then fix that other problem.
6v6gt:
Designing a retrofit hardware device to solve this problem is of course a fun thing to do and there are many alternative solutions to look at, but ultimately if the bought in system is inadequate (in this case cooling capacity is insufficient) then I'd tend to think of that as a problem for the manufacturer to solve.
You have to consider that any modification you make could be rendered invalid if there is a software or other upgrade on the ELTEK systems.
You also have to think of what happens if, say, one of these things catches fire and it is discovered that some unapproved modifications have been made. It might not look too good.
jackrae:
A cooling fan that overheats when running at full speed ? - something not quite right there.
The fan works on -48V system. And has a PTC thermistor to protect it.
Already seen ho-as-hell thermistor and the fan brushless coils.
Even had fans stop working due to overheat
As said before, it would be a fun project, but need to be cost effective.
Those eltek cabinets have +10 years in service and most are for 3G networks wich are slowly being updated to 4G (with newer cabinets with better cooling options)
However there´s still a lot of older cabinets that from time to time get overheated.
Also, it would not be practical to make one IC chip run 3 cabinets, since some 3G frequencys have more load than others and some overheat and some not. Would not want to have a cool cabinet running with more duty cicle because his neighboor is hot
Fabius_rex:
Also, it would not be practical to make one IC chip run 3 cabinets, since some 3G frequencys have more load than others and some overheat and some not. Would not want to have a cool cabinet running with more duty cicle because his neighboor is hot
What I meant was to have three independent PWM channels each with its own thermistor and fan. Don't know if it's practical for you but here's a discussion of multiple PWM.