USB Time Delay Fan Control

[Karaethon]

Holy crap, I just checked intel, max operating on my specific processors is 67°C.

From the Intel site: Thermal Spec 67°C

Ok, I will definitely make sure it never gets THAT hot again.

 

[hantto]

Holy crap, I just checked intel, max operating on my specific processors is 67°C.

From the Intel site: Thermal Spec 67°C

Ok, I will definitely make sure it never gets THAT hot again.

Just a thing that hapened once to me: I was poaking around my computer (had a Celeron 466 at that time) and when I was done, I forgot to reconnect the CPU fan. Happily not knowing it, I started my computer and started playing Homeworld. After a couple of minutes the computer started making beeping noices from the PC-speaker. I didn't understand what it was all about, so I just ignored it for a while. It wouldn't stop beeping, so I want to bios health status menu, and to my suprise it showed that CPU temp was 101°C!! A quick shutdown, and fan reconnection The computer still works with no complications whatsoever. Robust those celerons

 

[akg]

Just a thing that hapened once to me: I was poaking around my computer (had a Celeron 466 at that time) and when I was done, I forgot to reconnect the CPU fan. Happily not knowing it, I started my computer and started playing Homeworld. After a couple of minutes the computer started making beeping noices from the PC-speaker. I didn't understand what it was all about, so I just ignored it for a while. It wouldn't stop beeping, so I want to bios health status menu, and to my suprise it showed that CPU temp was 101°C!! A quick shutdown, and fan reconnection The computer still works with no complications whatsoever. Robust those celerons

It's amazing that the machine didn't crash!!!. (maybe the bios uses the waterfall method??)

 

[DirtyLude]

No, there's basic thermodynamics here that you are missing, if it's your CPU you are worried about.

As soon as you turn off the computer no more heat is being generated. Overall heat will not rise, only move from high concentration to low concentration. Your CPU is the hottest thing there when you turn off the computer. No matter how little circulation you have, it will never get hotter than it was when you turned off the computer. Like someone mentioned above, the heatskink will heat up as it continues to soak up the CPU heat, but that's not hurting anything.

You are just way over complicating things here. If you really want this for some reason, a simple 1 minute timer would be sufficient, there's no need for variable delay.

edit: Also, the accuracy of motherboard temp monitors it not very good.

 

[Karaethon]

believe me, I understand thermodynamics, and you are paraphrasing the second law, but I seem to not be getting the point across.
When the computer is on, the CPUs generate Heat (heat is just another form of energy, and all energy flows from a point of high concentration to a point of low concentration until entropy is achieved.)
This heat flows from the High energy CPU to the Low Energy Heatsink thus rasing the energy level in the heatsink.
the heat gets distributed throughout the heatsink to the fins
The Fins have a higher Energy level than the moving air( being accelerated over them by the fans) so heat is transfered from the fins to the air
finally the air exits the case taking the heat with it.

BUT once the computer powers down this changes because even though no more heat is being generated, the CPUs still have more Latent heat in them then the heatsinks, so heat continues to move into the heatsink. HERE is the problem Because the heatsink is no longer being cooled as eficiently the temperature of the heatsink rises untill entropy between the CPUs and Heatsinks is reached. of course the heatsink contiues to radiate heat and the temperature inside the case begins to rise BECAUSE there is no air flow moving heat OUT OF THE CASE. if the fans continued to run after shut down then the heatsink never reaches the same energy level as the CPU and the CPU temperature drops FASTER then without the fans on.

The whole reason I want to build this circuit is so I can control how long the fans run, and can set a shutoff point based on the temperature of the air moving out of the case. I KNOW a simple timer would keep the fans on, I WANT more control over them then that.

 

[DirtyLude]

Then I still don't see the problem. What is being damaged here then?

What are you going to get out of more control of the fan delay that will justify the effort?

If you want to build a nice real fan controller, an Analog Devices ADT7460 is a really nice chip. Gives closed loop fan speed control based on temp probe readings and you can run it automaticly or manually based on control over a simple I2C network.

With a microcontroller and a simple FT232 USB->serial chip you would have far more than you are asking for and would justify the effort, IMO.

The ADT only comes in a tiny SMD package, though.
**broken link removed**
**broken link removed**

 

[Someone Electro]

What did they teach you in phisics class.The temp can go up unles you are inputing some power.(But an heater that neads no power would be quite welcome in this winter)

Upsy! 67 °C I had myne at 80 °C for weaks an kept working.It was very hot to toch.So it was runing way out of spec for a long time.

Oh and if you want to compicate stuff you can put in an temp probe so the fan continues working until the temp drops enugh.

 

[Someone Electro]

Damit beaten by an minute!

 

[Karaethon]

Nothing has been damaged yet. and I want to make sure it stays that way.
I know a simple circuit like a timer, or a timer with temp probe would work, but I want control, I want to be able to set a temp that the fans will run the temp down to before shutting off, I want to do his from software, thus I need an interface to the computer, while Parallel or Serial ports would work, they are heading the way of floppy drives and 5.25" disks, I want to make sure this device will be usable for the future, thus the USB interface.

 

[DirtyLude]

I didn't ask what has been damaged, I asked what is being damaged. We've established that the CPU is not getting any hotter than it's running temperature, therefore it's safe. The heatsink will get a little hotter on average, but that's not going to effect it. The air in the case is going to get fairly warm, but everything in the case is easily durable enough to handle a bit of warm air.

What possible advantage will having manual control of the delay time using software give you in venting your case post shutdown; rather than giving a generous amount of time, say 10 minutes of fan time?

Now, if you're doing this, just because you want to, that's fine and dandy. I have no problems with that at all.

I'd suggest the Z8 Encore over the PIC as the uC of choice to start with. I wont repost all the reasons as I probably already sound like a Z8 Encore zealot to people here; just search for previous posts on it for the reasoning behind it.

The FT232 chip is a cheap USB serial converter that Jay mentioned earlier; You can use it to read USB with your uC. Drivers and specs are available on the site: https://www.ftdichip.com/

If you still want to maintain your requirement #4, I'll go back to my previous post. The Analog Devices chip is the best option if you want to monitor fan speed and control it. I've played with directly controlling the fan speed and trying to read the actual fan speed at the same time, and it's very difficult. I never managed to get it working fully by just using the uC and some passives. I think the problem with it is talked about in some General Chat section threads.

Once you have those three chips, the uC of your choice, the ADT fan control chip, and the USB->serial chip, you have a full scale PC controllable fan controller with thermal feedback. It's fairly simple to program the uC to control the ADT chip and relay data and controls back and forth from the PC, especially if the uC has a hardware I2C block.

 

[Karaethon]

Now, if you're doing this, just because you want to, that's fine and dandy. I have no problems with that at all.

Pretty much, I know its a complex solution to a simple thing, you are right its not even a poblem, but with the amount of money I dump into my computers, and the amount of money they make me, I'm a bit overprotective of them. Plus, as my fiancee says, I am a control freak, I have this unstoppable urge to control everything in my life.(she says she likes that because it frees her from having to worry about stuff because she knows I already did.)

Ok, I looked at the ADT7460 datasheet, and your right that is perfect for my needs, i'm looking at the FT232 and Z8 datasheets now, but the only thing I'm not 100% about is how to connect them all together...

 

[DirtyLude]

Everything you need is available from either Digikey or Mouser.com.

The datasheet for the ATD has some pretty nice sample circuits for controlling 3 fans. It includes some suggested logic level MOSFET part numbers. The MOSFETS I found on mouser.com. The picture of the control board I showed is a reproduction of one of the sample circuits.

The Z8 Encore or whatever uC you plan to use connects to the ATD chip over a 2 wire I2C network. Many uC's have an I2C hardware block so it's fairly simple to communicate over the I2C network. You can send commands to the chip e.g. FAN1 30% and it runs FAN1 at 30%, or request data from it, e.g. FAN1 RPM, and it will return the fan RPM.

Now, I have never used the FT232 chips, but I plan to use one in a upcoming project. As far as I can see from the datasheets, it's a pretty simple affair (someone correct me if I'm wrong). You need the USB connector, which the USB TX and RX pins connect to the FT232, then the FT232 connects to your uC over it's standard UART lines. Like I2C, many uC's have UART hardware blocks, which simplifies programming the serial communications.

That's pretty much it. The UC acts as simply a middleman between the USB communications and the fan control chip. You program it to take your commands that will come from the computer and relay them to the fan control, and take the data from the fan control and relay it back to the computer.

Eagle is a free (limited PCB size) schematics and PCB design software that I use. If you want my limited schematics you can have them. Actually I might take this project up myself. I've already made something similar without the USB control.

In order to start with the Z8Encore, I suggest getting the development kit from either Mouser.com or Digikey (part #:Z8F64200100KIT). It's $40.00 USD and includes everything you need to get started with it, the USB programming cable, a development board with a 64k chip on it, and the Windows development environemt. The ZDSII C compiler and assembler is a free download from zilog.com. Otherwise, you can use the PIC line. There's lots of support for the PICs here, and Nigels tutorials are very nice.

Putting this all together is not extremely difficult, but pairing it up with introducing yourself to microcontrollers is a fairly big project.

I'm assuming you live in North America, otherwise, I'm not certain availabilty of some of this stuff will be all that good.

 

[Karaethon]

Yep, North America, only a small distance form one of the most historic electricity/electronics birth places and only a hop skip and a jump away from THE birthplace of electricity.

does anyone know if USB is a daisy chainable system? can I connect this Inline with another USB port so it doesnt take up the port? or do I need to add in a USB hub Chip.

One more Question, in the datasheet for the ADT7460 I noticed that it used PWM to a mosfet to control the fans speed, the mosfet was on the Negitive power side of the fan? is there a reason for this? some reson to prefer regulation on that side rather then on the Positive power side? onlt readon I ask is in all the electronics i've built before I always use the negative power as a common point.

 

[Karaethon]

Ok, Reading the FT232's specs it mentions an eeprom, as I understand it this is used to identify the device to the USB system, now will I need one, or will the Z8 be used for that instead?

 

[DirtyLude]

My electical theory is craptacular. It's not very good at all. Swtiching ground is a very common practice. As for the reason, I have no idea. I'm sure someone else here would be able to explain why this is common. I think it has to do with some components ability to sink more current than they can source reliably.

USB is not daisychainable. You really can't give up a USB port for this? I can tell you, if you plan to try and make this into a USB hub as well, I can't help you there, and just looking at the cost/ complexity of the commercial hubs, I can pretty much say it's out of my league.

Integrated EEPROM - Previous generations of FTDI’s USB UART devices required an external EEPROM if the device were to use USB Vendor ID (VID), Product ID (PID), serial number and product description strings other than the default values in the device itself. This external EEPROM has now been integrated onto the FT232R chip meaning that all designs have the option to change the product description strings. A user area of the internal EEPROM is available for storing additional data. The internal EEPROM is programmable in circuit, over USB without any additional voltage requirement.

The FT232R has the EEPROM built in, but even so, the ony reason for it is if you want a custom ID. Unless you're planning on mass producing this thing and marketing it, I think you'll be fine not worrying about this feature for now.

 

[Karaethon]

Ok perfect. I had to trash my schematic drawing about halfway through tho(I swear one of these days I'll start doing them in pencil, but not today, lol) so I ended up just drawing it the way it is on the ADT datsheet this time. The FT232R huh, ok, i'll look for that, i found the datasheet for the FT232BM and was using that, but if the R has built in even better. and maybe, ifthis thing works and works good, I'll market it too. most of the over Clockers I know would use it, they are as anal about heat control as I am.

 

[DirtyLude]

Well, don't get too carried away, cuz there's already commercial product out there. The mCubed T-Balancer already does this. I know another guy that did the same thing as well, but his is much more complex to do the same thing. Can't find his page now.

Also, the ADT chip only comes in that tiny SMD package. If you're not planning on making up a PCB, I'm not certain how you're going to freform solder that thing.

 

[Karaethon]

wow, that thing looks col, but I after searching the web site, like I was trying to find WMDs, I could not figure out if it did one thing, run the fans while the computer is off. so maybe I have a leg up on them?

 

[DirtyLude]

This is my schematic for the ADT's reference design:
**broken link removed**

That's what I used to build this test board:
**broken link removed**

I have some schematics with it connected to a uC, but they are fairly complecated, since my first project was for a fan controller for a watercooling rig that never came out. The board was going to control the water pump, the fans, and communicate with a remote display box with a 433mhz transmitter module.

 

[Karaethon]

ok, let me make sureI have this right, now I connect
Z8 Encore | FT232R
TXD0 | RXD
RXD0 | TXD
CTS0 | RTS
DE | CTS
CLKIN | CBUS0

Is All this right? I kinda pieced tis together form going back and forth between the FT232R and Z8Encore datasheets, trying to match up pin descriptions and where they go.if all this is right, is that ALL the connection i need between the two (besides power?) Oh and which Z8 do I want? Im looking at the datasheet for the Z8EncoreXP but it only has 4K, i see theres another with 64K