Hey @stud I think something to note here is that there is a pretty big difference between the components that are being used to control these fans.
L298n components really have nothing to do with PWM
Optocouplers with appropriate resisters do control PWM
L298n really only limit output voltages similar to how potentiometers work however this is wasteful as the resistors waste the unused energy.
This post has some really great information as the original poster did the leg work to figure out the proper parameters for controlling PWM on these wildly popular dc/ec fans.
AC Infinity CLOUDLINE S6 - PWM to 0-10v help
Sounds like I've got a super similar problem with just a little different hardware.
I've got a pca9685 http://wiki.sunfounder.cc/index.php?tit ... rvo_Driver (connected to a huzzah32 feather) that I've plugged my lights in to dim/automate and when I tried the same with my AC Infinity fan (Yellow to PWM, Red to V+, black to GND) I was able to get it to turn on but it ran extremely slow, like less than 10% probably. I'm wondering if I need to add something to the setup to get it to work? Maybe it's a frequency issue that's not compatible with the lights?
Right now I have the fan attached to a smart plug with an automation to turn on when I want but it will only turn on 100% this way. It works for now but is really just a band-aid solution.
I've got a pca9685 http://wiki.sunfounder.cc/index.php?tit ... rvo_Driver (connected to a huzzah32 feather) that I've plugged my lights in to dim/automate and when I tried the same with my AC Infinity fan (Yellow to PWM, Red to V+, black to GND) I was able to get it to turn on but it ran extremely slow, like less than 10% probably. I'm wondering if I need to add something to the setup to get it to work? Maybe it's a frequency issue that's not compatible with the lights?
Right now I have the fan attached to a smart plug with an automation to turn on when I want but it will only turn on 100% this way. It works for now but is really just a band-aid solution.
-
PerniciousSnit
- LED-Curious
- Reactions:
- Posts: 1
- Joined: Sun Jun 12, 2022 5:08 am
I wanted to control the speed of a AC infinity A6 lite and ran into this thread! I ended up doing a little digging, and can hopefully answer some of the early questions. Based on measurements I just made, and earlier comments here, I think I can put them all in one place.
The interface is a 10V PWM interface where the duty cycle of the control signal controls the fan speed. This line is pulled up to 10V, so no controller plugged in represents a 100% duty cycle, and the fan spins at maximum rate. Likewise shorting the control input to ground represents a 0% duty cycle, and the fan stops.
Interface specs:
- 4 pin molex male connector. These can be hard to find as naked connectors; it's normally much cheaper and easier to find a cable with the connector inside an old PC. Otherwise any cable like has a male connector already wired up.
- 10 V PWM with 4k7 (estimated) pull up resistor
- threshold voltage seems to be ~5V
- fan speed is PWM, with a frequency of ~5kHz.
- 0% duty cycle represents 0% fan speed
- 100% duty cycle represents 100% fan speed
Pinout:
Pin #1 - N/C
Pin #2 - Control signal
Pin #3 - 0V
Pin #4 - 10V
Note: newer fans use USB-C connector, but AC infinity supplies a USB-C to MOLEX adapter. These pin numbers come from that adapter, which has female pins!
Electrical:
10V levels are NOT compatible with most microprocessor electronics, so some additional electronics will be needed. The role of this electronics is to connect the control signal and 0V signal together, sinking (est) 5 mA. The electronics required depend on how the controlling device is to be powered, and whether electrical isolation is needed.
For applications with modest power requirements it may make sense to power the controller from the fan itself, giving a common 0V rail. In this case a common 0V rail is present and only a simple small signal NPN transistor is required, with a modest base resistor (4k7 would be OK, generally) to limit the current from the control input.
It MAY be tempting to just connect the 0V of the controller, to 0V of the fan and use a simple NPN transistor. I'd seriously suggest AGAINST doing that - it could be dangerous to both sides of the circuit, and to you personally! Unless you're absolutely sure you can safely connect the controller 0V and the fan 0V safely, use optocoupler approach - it only adds a dollar or two and gives a lot of safety.
For applications where it is desirable to electrically isolate the fan from the controller, a good solution, as OP, did, is to use an optocoupler - basically where a LED that turns the transistor on, rather than an electrical signal. Note these are MUCH less sensitive than an electrically controlled transistor, and need more drive current!
The important parameter for an optoisolator used in this application is the CTR - the ratio of LED current to the current flowing through the transistor. As mentioned above, I expect we need to sink ~5mA (to be safe - strictly >1mA looks like it would just work). This should be used to calculate the series resistance driving the optocoupler LED, based on the MINIMUM CTR of the optocoupler, the control signal voltage, maximum optocoupler led current, and maximum control output current - typically the GPIO drive current of an attached processor. Ideally choose the resistor to get 2x the required LED current if you can manage it.
In my case I think I can get away with a '555 running directly from 10V, fan supply but I'll probably need an optocoupler anyway to isolate the signal and ground of the K40 laser cutter that says when the laser is on, from the fan supply power supply - which I want to use to make the fan go fast!
Edited to add: I searched for this using the words ac infinity uic interface or molex pinout , and found nothing! Putting them here so Google could have found this thread!
The interface is a 10V PWM interface where the duty cycle of the control signal controls the fan speed. This line is pulled up to 10V, so no controller plugged in represents a 100% duty cycle, and the fan spins at maximum rate. Likewise shorting the control input to ground represents a 0% duty cycle, and the fan stops.
Interface specs:
- 4 pin molex male connector. These can be hard to find as naked connectors; it's normally much cheaper and easier to find a cable with the connector inside an old PC. Otherwise any cable like has a male connector already wired up.
- 10 V PWM with 4k7 (estimated) pull up resistor
- threshold voltage seems to be ~5V
- fan speed is PWM, with a frequency of ~5kHz.
- 0% duty cycle represents 0% fan speed
- 100% duty cycle represents 100% fan speed
Pinout:
Pin #1 - N/C
Pin #2 - Control signal
Pin #3 - 0V
Pin #4 - 10V
Note: newer fans use USB-C connector, but AC infinity supplies a USB-C to MOLEX adapter. These pin numbers come from that adapter, which has female pins!
Electrical:
10V levels are NOT compatible with most microprocessor electronics, so some additional electronics will be needed. The role of this electronics is to connect the control signal and 0V signal together, sinking (est) 5 mA. The electronics required depend on how the controlling device is to be powered, and whether electrical isolation is needed.
For applications with modest power requirements it may make sense to power the controller from the fan itself, giving a common 0V rail. In this case a common 0V rail is present and only a simple small signal NPN transistor is required, with a modest base resistor (4k7 would be OK, generally) to limit the current from the control input.
It MAY be tempting to just connect the 0V of the controller, to 0V of the fan and use a simple NPN transistor. I'd seriously suggest AGAINST doing that - it could be dangerous to both sides of the circuit, and to you personally! Unless you're absolutely sure you can safely connect the controller 0V and the fan 0V safely, use optocoupler approach - it only adds a dollar or two and gives a lot of safety.
For applications where it is desirable to electrically isolate the fan from the controller, a good solution, as OP, did, is to use an optocoupler - basically where a LED that turns the transistor on, rather than an electrical signal. Note these are MUCH less sensitive than an electrically controlled transistor, and need more drive current!
The important parameter for an optoisolator used in this application is the CTR - the ratio of LED current to the current flowing through the transistor. As mentioned above, I expect we need to sink ~5mA (to be safe - strictly >1mA looks like it would just work). This should be used to calculate the series resistance driving the optocoupler LED, based on the MINIMUM CTR of the optocoupler, the control signal voltage, maximum optocoupler led current, and maximum control output current - typically the GPIO drive current of an attached processor. Ideally choose the resistor to get 2x the required LED current if you can manage it.
In my case I think I can get away with a '555 running directly from 10V, fan supply but I'll probably need an optocoupler anyway to isolate the signal and ground of the K40 laser cutter that says when the laser is on, from the fan supply power supply - which I want to use to make the fan go fast!
Edited to add: I searched for this using the words ac infinity uic interface or molex pinout , and found nothing! Putting them here so Google could have found this thread!
-
xthursdayx
- LED-Curious
- Reactions:
- Posts: 4
- Joined: Fri Jun 24, 2022 7:17 am
Just wondering if you have any more updates on this system after a few more months? I’m trying to decide on the best method for pwm control for my Cloudline S4. I’d prefer not to mess with wiring in the control box if it’s not necessary and it seems like there are some questions about the validity of using mosfets the way some of the other methods do.thevalkyriesride33 wrote: ↑Sat Dec 18, 2021 8:36 pmThe solution I came to doesn't require any other components other the the a PWM signal, an opto and some resistors. Been running very stable with multiple speed changes for over a month now without issue.