Jump to content
RESET Forums (homeservershow.com)
urbangenie

Microserver Gen8 fan PWM signal inverted?

Recommended Posts

urbangenie

Hi everyone

I've been going through the process of upgrading my G1610T Microserver. I started off by replacing the fan and PSU (I threw a 160W PicoPSU in there - works brilliantly). I'm next planning on replacing the CPU and adding active cooling - I've got an i3-2100 at home and I've got a Noctua NH-L9i coming in the post.

During my research into modifying the Noctua fan to fit the motherboard holes, I came across this post by Spivak. His last paragraph caught my eye:

Quote

Now I have mounted it and supplied with the molex 12v and using the PWM from the mobo connector. But the fan behaves just the oposite as the case fan. Means if the case fan speeds up the cpu fan slows down, and if the case fan speeds up the cpu fan slows down. Any suggestions?

When I replaced the main fan, I noticed that the fan started off slow at POST and then sped up once the OS kicked in. This is exactly the opposite to the stock Delta fan, which started off fast and then slowed down. I had put this down to having shorted the RPM lead to ground, and just put a 100 ohm resistor on the 12v line to slow the fan down. But after reading the above it made me wonder if the Gen8's motherboard is actually sending out an inverted PWM signal rather than a regular one.

I don't own an oscilloscope so I have no way of definitively confirming this. I haven't seen any mention of this anywhere else (I spent a few hours researching this last night) so I'm quite ready to be told that I'm wrong! But given everything I've seen so far it makes sense that this is what's happening. And I don't think this would be the only time that HP have used inverted PWM fans.

So, my next step is to work out how to build some sort of inverter circuit. I've read bits about Inverted Schmitt Triggers using NE555 chips, and basic transistor inverters (NOT gates) using a BC548. But my electronics knowledge is basic at best - give me a circuit diagram and I'll solder something together, but I would struggle to design the circuit in the first place.

Does anyone here have any experience with this? And could anyone test my hypothesis?

Thanks!

Share this post


Link to post
Share on other sites
urbangenie

Hi JBark

Thanks for the link, but I've already done this. Connecting the fan wasn't too much of an issue. What I'm trying to do is stop fans from doing the exact opposite of what the Microserver's motherboard is telling them to do.

At the moment the Gen8 motherboard sending an inverse signal, because that's what the stock Delta fan wants. So, whereas normal PWM fans look for a 100% signal for full speed, the Delta fan looks for a 0% signal. As there's no way of changing this in the Microserver's BIOS, I need to build something which inverts the signal so that normal fans work they way they're meant to.

I did some more research last night, and it looks like I can make a PWM inverter circuit using an NE555 chip. There's a circuit diagram in this thread over at the Overclockers forum which amplifies a PWM circuit. This uses two Schmitt Triggers in series (an NE556 is basically two NE555 chips in the same package) to invert a PWM signal, and then invert it back again, which ends up making the PWM signal stronger. If I were to use half of this circuit (i.e. only use the first half of the NE556) I should be able to get the signal the right way around for normal fans.

As luck would have it, I actually have an NE555 in a box of old electronics given to me by a friend, so if I have time this weekend I'm going to give this a go. Hopefully I don't end up blowing up the motherboard!

Share this post


Link to post
Share on other sites
JBark

Ahh, I didn't read close enough to see you wanted PWM control of your fan as well, most people just either use a quiet fan at full speed, or use some other controller for the fan speed.

But yeah, 555 should do the trick.  Other thing to watch out for is that the PWM signal on the G8 is only 3.3V, and the PWM spec is 5V.  However, I'm pretty certain most decent fans these days have a 3.3->5v pullup on the PWM line for just this reason.  So I'd hope something like a Noctua would just work once you've inverted the signal.

  • Like 1

Share this post


Link to post
Share on other sites
ellnic

@urbangenie did you ever solve this? I'm having the same issue. 

From what I can gather, the PWM signal isn't inverted, but becomes inverted for the second fan. I don't pretend to know too much about this, but I was reading on a different forum that a PWM signal may become inverted or is too weak as more fans are added to the same socket - ie. we took voltage from the SATA loom, but the control signal is also too weak for more than the delta. 

 

To check this, I connected just the CPU fan to the board - and sure enough, it behaves correctly (spins up at boot, then increases in speed in the same way the delta does) but then the system flags improper fan. I am sure it's behaving correctly though. If I hook both fans up but swap their positions on the loom, the CPU fan behaves normally and not the delta. So the signal really does seem to be different for each fan. 

 

At the moment I have my CPU fan (K25) 12v'd just to give sufficient cooling under load, but I would like to solve this if possible. 

One thing I don't undetstand is how just a few of us are having this problem. Having the board unintentionally slow down the CPU fan as the heat rises in the cores is surely something that would have been noticed. I can only think that:

 

1. Something is fundamentally different in our wiring that is causing this but I have tried several things including @overcoat's exact wiring

 

2. Others in the thread have observed the CPU fan spinning, reassembled and not actually manually checked the revs. After all, with no sense wire - there's no iLO data on it. 

 

If it's as simple as a Schmitt buffer, i'll grab a 555. But if I can't solve, I'll probably add a potentiometer and set to about 9-10v. 

 

Edit: here is the thread I was referring to: http://www.overclockers.com/forums/showthread.php/642297-Multiple-PWM-case-fans-this-easy?s=03bb45ed5e3768d88fdd3688da70e896

 

Edited by ellnic

Share this post


Link to post
Share on other sites
rustywood
On 6/8/2017 at 2:43 PM, urbangenie said:

Hi JBark

Thanks for the link, but I've already done this. Connecting the fan wasn't too much of an issue. What I'm trying to do is stop fans from doing the exact opposite of what the Microserver's motherboard is telling them to do.

At the moment the Gen8 motherboard sending an inverse signal, because that's what the stock Delta fan wants. So, whereas normal PWM fans look for a 100% signal for full speed, the Delta fan looks for a 0% signal. As there's no way of changing this in the Microserver's BIOS, I need to build something which inverts the signal so that normal fans work they way they're meant to.

I did some more research last night, and it looks like I can make a PWM inverter circuit using an NE555 chip. There's a circuit diagram in this thread over at the Overclockers forum which amplifies a PWM circuit. This uses two Schmitt Triggers in series (an NE556 is basically two NE555 chips in the same package) to invert a PWM signal, and then invert it back again, which ends up making the PWM signal stronger. If I were to use half of this circuit (i.e. only use the first half of the NE556) I should be able to get the signal the right way around for normal fans.

As luck would have it, I actually have an NE555 in a box of old electronics given to me by a friend, so if I have time this weekend I'm going to give this a go. Hopefully I don't end up blowing up the motherboard!

 

Hi @urbangenie can you update on this at all?

Share this post


Link to post
Share on other sites
Cap_peeps

Yes, I would be interested about the results too. It looks that I have a same issue as you guys. it means: new CPU fan is slowing down once case fun is getting faster and exactly opposite, when case fan getting slower, the CPU fan starts to run faster. 

 

I would be more than thankfull if there is any kind of circuit board scheme or any guide how to make a PWM signal invertor. I dont understand why the HP had to complicate everything so much. Why to make life easier if you can struggle ? 

 

Thanks all.

Share this post


Link to post
Share on other sites
Cap_peeps

Hi all, 

I was searching and googling for another day for the solution and just found this solution.

https://electronics.stackexchange.com/questions/322390/12v-pwm-fan-seems-to-be-running-inversely-of-what-it-should-be

 

But again no information if the mentioned setup fixed the issue or not. I also dont have the mentioned parts and would like to discuss it with my colleague first. So if anyone of you has a parts and time you can try and reply back here. It really helps to people who are trying to cheat the HP :D. 

 

Many thanks.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...