1 Topic by wiley

  • wiley
  • Member
  • Offline
  • Registered: 2016-05-16
  • Posts: 45

Topic: Hotend Fan Switch

I was researching the hotend fan and what it would take to switch the 12v E3D v6 fan like the factory fan was switched today, and I had an interesting discovery. Over on Voltivo, someone traced the wiring on the daughterboard for the two fans and found that they actually share an always-on 5v rail. The "switch" is actually a GPIO-switched NPN transistor that opens and closes the ground connection for each fan.

This means that if the ground is the same across the 12v and 5v rails and the current through the 12v fan isn't too much for the NPN transistor, you can just wire your fan from the 12v supply line everyone is tapping to the same switched ground the original fan used. A quick check with a voltmeter showed that the ground lined up and fan currents are pretty low anyway, so I went ahead and tried it. It worked. I wouldn't recommend this as anything permanent until I get a chance to check the parts on the mainboard to make sure the currents are okay in the long run, but it's promising.

Has anyone else poked at this before? Has anyone here tracked down where this NPN transistor lives on the mainboard (and more importantly, where the second NPN transistor that's only populated on the Duos lives?)

2 Reply by Telemachus (edited by Telemachus 2016-06-24 05:44:27)

  • Telemachus
  • Member
  • Offline
  • From: Riverside, CA
  • Registered: 2016-01-19
  • Posts: 289

Re: Hotend Fan Switch

So what you are saying is supply 12V+ to your e3d directly from your power supply, but connect the ground to the (switched) ground wire on the daughter board that the original fan connects to?

We know that the current limiting resistor going to the base of the NPN transistor from the micro controller is rated with enough ohms to open the "gate" of the transistor enough so sink the 5V that the original fan runs off of,  but potentially not enough to sink 12 V.   

That having been said, ground is ground, and the only thing you would risk is the E3D fan not running as fast as it usually would.


I like your idea, but I think I would rather use it to control a 12V layer fan, and just keep the E3D fan on permanently.

Da Vinci 1.0a.  Repetier 0.92.  E3D V6.  Titan Extruder.  Sllic3r.  Octoprint on Raspberry Pi B+.

3 Reply by wiley

  • wiley
  • Member
  • Offline
  • Registered: 2016-05-16
  • Posts: 45

Re: Hotend Fan Switch

Aye. My evil plan also includes getting the second fan output working so I can control a layer fan too. I need to pull the main board out to get a proper look, but I think XYZ may have switched to SMD components.

4 Reply by Telemachus

  • Telemachus
  • Member
  • Offline
  • From: Riverside, CA
  • Registered: 2016-01-19
  • Posts: 289

Re: Hotend Fan Switch

The spots are pretty clearly labeled on the motherboard for the components to the second extruder.  Using through-hole components shouldn't present that great of a challenge.

Da Vinci 1.0a.  Repetier 0.92.  E3D V6.  Titan Extruder.  Sllic3r.  Octoprint on Raspberry Pi B+.

5 Reply by wiley

  • wiley
  • Member
  • Offline
  • Registered: 2016-05-16
  • Posts: 45

Re: Hotend Fan Switch

I pulled the board out this evening and had a look, and I do see the 1k resistor, capacitor, and transistor. On my board, though, the resistor and transistor are SMD components. Were they always SMD and people have just been soldering through-hole components onto the board because it's easier, or did XYZ change this across board revisions? My board also has the drivers built into the board, which seems to have been a recent change.

I'm also a bit confused about the transistor. If I trace the wire for the working extruder fan, it leads to a SOT23 package labeled "W1P 52" which seems to correspond to a BFT92 PNP transistor, but a PNP transistor doesn't make any sense at that place in the circuit. I keep looking for a way I'm reading the label wrong or somesuch, but I can't seem to figure out where I'm getting it wrong.

6 Reply by Telemachus (edited by Telemachus 2016-06-24 13:52:04)

  • Telemachus
  • Member
  • Offline
  • From: Riverside, CA
  • Registered: 2016-01-19
  • Posts: 289

Re: Hotend Fan Switch

Yes, from what I read they are soldering jumper wires to through hole components.  The stepper drivers are all soldered to the board on late 1.0 and 1.0A.  We don't really have a valid option to replace them.

Here is a thread on github that offers limited info about the mod itself, and good info about changing the FW to support it:

https://github.com/luc-github/Repetier- … /issues/64

As you work this out, please document it well.  I've seen very little useful info on doing this.

Da Vinci 1.0a.  Repetier 0.92.  E3D V6.  Titan Extruder.  Sllic3r.  Octoprint on Raspberry Pi B+.

7 Reply by wiley

  • wiley
  • Member
  • Offline
  • Registered: 2016-05-16
  • Posts: 45

Re: Hotend Fan Switch

So as it turns out, according to NXP (formerly Philips) Semiconductors...
- W1% corresponds to the BFT92 PNP transistor
- *1P corresponds to the PMBT2222A NPN transistor

And it apparently didn't occur to them that this breaks the component search engine on their website.

By the datasheets, this has to be a PMBT2222A. Even if I had completely misunderstood the circuit and a PNP was correct, the BFT92 has a tenth of the necessary current.

The good news is that a PMBT2222A has a max rated current of 600mA, which is just fine for running the E3D v6 fan. It's also enough for the set of four 60mm fans I plan to put on the sides of the case and try to use for layer cooling without mounting a fan on the carriage. The only question at this point is whether I stick to SMD or try to put through-hole components on the SMD pads. Not sure about that yet.

8 Reply by wiley

  • wiley
  • Member
  • Offline
  • Registered: 2016-05-16
  • Posts: 45

Re: Hotend Fan Switch

http://soliforum.com/i/?ggQ5v1r.jpg

Now to see if it works.

9 Reply by wiley (edited by wiley 2016-06-27 18:59:19)

  • wiley
  • Member
  • Offline
  • Registered: 2016-05-16
  • Posts: 45

Re: Hotend Fan Switch

The short answer is yes, it works.

The long answer is at https://github.com/luc-github/Repetier- … issues/165 - there are some firmware modifications involved. It's also worth checking your board to see if you already have the resistor and transistor that I had to add - apparently there was a previous bug where someone already had them from the factory on a 1.0A.

And yes, reflashing the firmware preserved my printing timer. I exported the EEPROM to XML via Repetier-Host, though, and you can also just restore it afterwards.

I'm still waiting on an order of fans, so I'll have to wait to see how the whole setup functions.

10 Reply by Telemachus (edited by Telemachus 2016-06-28 07:28:08)

  • Telemachus
  • Member
  • Offline
  • From: Riverside, CA
  • Registered: 2016-01-19
  • Posts: 289

Re: Hotend Fan Switch

Thanks for the feedback on the eeprom timer.  I think it is time to repurpose my extruder fan and conserve my 53 day counter.

Did you end up using SMD components?

Da Vinci 1.0a.  Repetier 0.92.  E3D V6.  Titan Extruder.  Sllic3r.  Octoprint on Raspberry Pi B+.

11 Reply by wiley

  • wiley
  • Member
  • Offline
  • Registered: 2016-05-16
  • Posts: 45

Re: Hotend Fan Switch

Nah, as you can see in the picture, I went with through-hole. I rediscovered the pitfalls of through-hole components on surface mount pads when I accidentally ripped off the top pad of that current limiting resistor while trying to trace it to a pin on the microcontroller, but I managed to scratch away the board surface above that trace leading to it and solder to that. The result is quite fragile, but the board lives on a nice mount inside a box. Shouldn't really be an issue.

It seems a lot of the confusion around this mod is that the DaVinci main board is actually capable of switching four fans. There's the one used from the factory, the one I filled in, and another two on that Fan2 header to the left that I assume is used on the Duo. It's more obvious when seen from directly above, but a picture from above doesn't really show the components I added.

12 Reply by Telemachus

  • Telemachus
  • Member
  • Offline
  • From: Riverside, CA
  • Registered: 2016-01-19
  • Posts: 289

Re: Hotend Fan Switch

Ahh... I see them now.  I was looking in the spot where the Fan2 would be... didn't think to look over there!

I would be inclined to solder small jumper wires to the board that then lead to the through-hole components.  That helps relieve the strain on the traces.

Da Vinci 1.0a.  Repetier 0.92.  E3D V6.  Titan Extruder.  Sllic3r.  Octoprint on Raspberry Pi B+.

13 Reply by Hecot (edited by Hecot 2016-07-11 09:39:30)

  • Hecot
  • Member
  • Offline
  • Registered: 2016-02-29
  • Posts: 59

Re: Hotend Fan Switch

For second fan support on Davinci 1.0A please refer to my great working mod: https://github.com/luc-github/Repetier- … issues/152

I also designed a fan duct for it: http://www.thingiverse.com/thing:1634657

14 Reply by wiley

  • wiley
  • Member
  • Offline
  • Registered: 2016-05-16
  • Posts: 45

Re: Hotend Fan Switch

I wired my cooling fans the same way you did, but I didn't want to mount them on the carriage. Instead I have four 60mm fans mounted two in each hole in the sides of the printer with a variation on http://www.thingiverse.com/thing:1339148 . It also turned out that 600mA isn't enough for all four fans, so I'm actually using that current to switch two PNP transistors on a pair of small PCBs that live next to the fans. I'll grab some pictures at some point, but the whole setup does work and it's definitely improved my prints, although I'm not sure how well it compares to a carriage-mounted cooling fan.