Temp controlled fan installation finished

I just finished the install of a temperature controlled fan. The result is quite good. More info about it here: http://users.pandora.be/miata/english/empeg_fan_installation.htm

Frank

That is killer! Great looking job!

A very decent page, a very nice project. I'm going to link that one from the FAQ. Will the URL remain constant for a reasonable amount of time?

One thing: On your page, you state, "There's an unused cutout for a switch on the back corner (marked security) so I didn't have to cut any holes." Actually, this hole is not for a switch. Click here for an explanation of the purpose of the security slot.

May I also suggest detailing in the text where the fan gets its power from? And whether or not you've noticed if there is any extra electronic noise on the outputs induced by the fan?

Ooo, it would also be cool if you could list the sources for the components, too. For instance, Ratshack/Jameco part numbers or whatever.

That's very nice indeed.
It would be cool (pun intended) if someone offered this in kit form, pre-soldered and heatshrinked etc. ready-to-go

The icing on the cake would be some way for Hijack to take control of the threshold setting (via an unused internal DAC channel within the Empeg?) so it could be controlled via the front panel.

Yes, that page is supposed to remain constant.
Thanks about the security slot, didn't know about it ...
I don't know any part numbers as I bought these in a local shop. They're very common parts though (although the thermistor usually doesn't look like that).
I added a few more details to the page.

Frank

Cool.

I think I may do the same sort of install except onto the outside of my sled. It would be easy to pull in cold air from the ac to pump into the unit due to the position of the ducts in my '98 M3. The sled-fan would line up with the empeg holes in the case, and I'd mount the thermister in the dash below/behind the empeg (away from cold air vent). I'll build the switch into my dash somewhere in case I ever use the heater.

John

Too CL!

Hah, somehow I knew this question would pop-up ...
I *could* make some, but the switch would be a problem. The switch required a bit of modification and was not very straightforward to fit.

Frank

Hi.

Well, how about just supplying the switch unmodified? Either that or leave it out completely, so the user would have to get one that fits. Either way, I think finding a good (fitting) fan and doing the temperature control right is the most difficult thing to most people (besides finding a source for the parts).

Anyway: What exactly was the problem with the switch in its unmodified state and what did you do to fix it?

cu,
sven

Hi.

Another question: What modification to the temp/voltage control circuit is needed to make the output voltage lower (i.e. let the fan turn slower)?
The empeg is has a specified operating temp range that goes up to 55°C, so I would like to have a steeper curve on the voltage that starts later, which would probably mean I need a different thermistor. But even with the thermistor you used, it should be possible to have the fan stay off/inaudible up to at least 30°C or something.
Like in my stereo(home)setup, I know for sure the ambient temperature of the room will never exceed 26°C, but the empeg might get some heat from below, so that it goes up to 50°C while operating (well, after operating for more than an hour), so I would like the fan to still be silent at that temperature, because even a slow fan should be able to cool the empeg down quite a bit under these conditions. Others might have pretty high ambient temperatures, and thus need more airflow. a small poti to adjust the fan speed would be nice.

cu,
sven

Attached is a graph of Fan Voltage vs temp for various values of R2. My suggestion would be to replcae R2 with a 1k pot.

The graphs are theoretical, based on the formula for the regulator and the thermistor resistance/temp values that Franks had. Some errors are evident, but the graphs should be good enough to get you started. Some of them show voltages over 14 volts. Obviously this isnt really going to happen - they will level out at Vbattery minus some L200 specific value (2V ?? - I'm guessing here.)

The problem with the switch is that it was a bit too long to fit in the corner. I had to bend one of the mounting lips to a 90° angle. The mounting holes of the switches are very small and since I didn't have small enough bolts, I had to drill the holes out. Of course, the head of the bolt I used was too big, so I had to grind it down a bit etc etc ... Finally I only used one bolt and some silicone to keep it in place.

Finding the right fan is indeed not a simple thing to do. The specs of the fan I used are 0.6W, 5500RPM, 4.9CFM, 0.12in H2O, 26dBA.
Sunon also makes 6mm fans (but they're 5V only so you'd need to modify the circuit) and 8mm fans (but they're more noisy at 32.5dBA).
I wouldn't use anything thicker than 10mm as it is a thight fit as it is. The capacitors on the back of the display couldn't be in a more worse place . I don't think that even a 8mm fan would be a direct fit. You'd still have to tilt the display a bit.

To lower the voltage at the fan, lower the value of R2. My initial plan was to use a 1K trimmer pot, but used a fixed resistor to keep it simple. The problem with a lower value for R2 is that it lowers the entire curve and doesn't make it steeper (on the contrary). If you want to make it steeper, you'd have to get a thermistor with a higher 'B' value (a measure of the thermistors sensitivity to changes in temperature, higher B values giving greater % changes in resistance). Downside is that the temperature will probably stabilize at a higher value too.
I have no idea of the B value of the thermistor I used. The place I got it from didn't even know about B values. They only labeled their thermistors by resistance.
I actually bought 3 thermistors, 220, 470 and 1K2 and they all gave very similar results (steepness of the curve). In the end I used the 470 because it was easier to install.
So you see, finding the right thermistor isn't easy either.

The fan turns off when the voltage goes under 5V. If you want to keep it inaudible, you'd need to keep the voltage at about 5-6V because the holes in the top lid create a lot of turbulence. The fan made a lot less noise when not attached to the lid.
If you would do that with the thermistor I used, you'd need to use a 400-500 ohm resistance for R2 (thanks for the graph Genixia), but than you'd never get 12V at the fan when you need it.

I have the player sitting at my desk while I'm typing this. With the fan turned off, it stabilized at 38°C. With the fan turned on, at 28°C. This is a very good result, but it isn't really necessary. The empeg is perfectly happy at 38°C so there's no need to have the fan go on. I might look for a thermistor with a higher B value, add a trimmer and do some more testing. In the mean time, I disable the fan when I take it in the house. Iit usually stays in the car so it's not really that big a deal for me.

I'll add this info to my installation page.

Frank

I'll add this info to my installation page.

Feel free to steal that graph

Hmm. 0.6W/12V = 50mA. There's more than a few Op Amps capable of drving that current directly, and that would open up a lot of possibilities wrt to Wheatstone'ing the thermister, and biasing the opposite leg of the bridge so that the fans doesn't start until after the bias has been balanced. You could get pretty aggressive with the gain then...

Thanks, I'll do just that

Good idea about the opamp. I might try that instead.

Frank

As it happens I've just finished my fan controller. I mentioned I was going to have a go at designing something after initially fitting my fan some time ago. I designed it and bought all the components several months ago but I've only just got around to actually soldering the components into a bit of Veroboard. My knowledge of electronics is very limited but it seems to work. It's sitting here on my desk at 32 deg C where it would normally be at about 43 deg C and I can't hear the fan unless I get my ear really close to it. I'll have to see how it goes in the car (where it would always top 55 deg C after about 1 1/2 hours driving.

The fan does increase in speed as the Thermistor gets hotter (tested by bringing soldering iron near to Thermistor) but I've yet to correlate temp against output voltage. I've attached a scan of my original circuit diagram and an Excel spreadsheet of all my calculations in case anyone's interested or wants to take my design and improve it. As you can see from the circuit diagram part of it didn't work so I removed it. That bit was supposed to keep the fan off until it reached about 6V.

I'll do some more experimentation & take a few measurements when I get time. Meanwhile someone with more electronics knowledge might like to improve it.

You may notice I've used a Wheatstone bridge and an Op-Amp (LM324 - quad) as mentioned in a couple of recent posts in this thread.

At the Amersfoort meet Patrick showed an empeg with the remote mount face, as well as video out - too cool! In order to keep things cool it has to have a fan. The one used was really slim and mounted to the lid with plastic rivets. IIRC he said it was from Farnell, you might want to ask him for an exact item number.

/Michael

At the Amersfoort meet Patrick showed an empeg with the remote mount face, as well as video out

Wow! I sure hope somebody snapped some shots of that!

They did... I expect pics to appear soon (didn't bring a camera myself this year :-( ), if they haven't already. Quite possible that piccie links are posted elsewhere on the BBS - amazing what a backlog you get from not reading the BBS in a couple of days...

Edit: It was also the third most expensive player existing (it has two brothers), and you shouldn't be holding your breath waiting for either feature to be available...

/Michael

It was also the third most expensive player existing (it has two brothers), and you shouldn't be holding your breath waiting for either feature to be available...



Heh! Yeah, I thought as much already.

Still...it's nice to drool ! (for the right reasons )

Hmm, I can't understand what you were trying to do with the 7805 in that circuit...theoretically your fan could only get 5V max..

LM324 has a 20mA Iout capability. Assuming you're using a similar spec fan to Frank, you could do a dirty hack to remove the need for a transistor/regulator. Use 1 of the OpAmps to do the detection as your circuit currently shows, and hook up the other 3 in parallel as a current driver to give you 60mA max driving current. (It's not the ideal way to do things but it should work.)

Just a quick update.
I replaced the resistor with a trimmer. At about 600 ohm it seems to quiet the fan done quiet a bit. Where it used to run 31°C on my desk, it now runs 33°C, so I might even lower it more.
I haven't yet had the chance to test it in the car (this morning it only got 41°C with the fan off). I'm hoping it'll be hotter on the way back home.


Frank

In reply to:

---

Hmm, I can't understand what you were trying to do with the 7805 in that circuit...theoretically your fan could only get 5V max..

---

You can use a fixed voltage regulator as an adjustable one by varying the voltage at the reference input just like you do with an adjustable one. It probably doesn't function as well as one which is has been designed to be adjustable but it works well enough for this application.

The Transistor was supposed to remain off until it's Gate voltage had reached a certain value (650mV I think) and then turn the fan on. This would prevent the situation where the fan isn't quite getting enough voltage to start it properly and it sits there oscillating back and forth.

Using the remaining Op-Amps in parallel as a current driver sounds like a good idea and is something I hadn't thought of. I'll look into that. Thanks for the tip .

Small update:
Here's a graph of Fan RPM v Temp. I intend to do a bit more measuring and experimentation but all seems hunky dory at the moment.

[being_stupid]Okay. I might be really stupid and not quite seeing it but why does the temperature increase when the fan RPM increases?[/being_stupid]

[edit]Okay. I've realised now that it must be showing the action of the thermistor as well

- Trevor

I left the player in my car yesterday (thinking it was only 80 F outside and would be OK) but when I got to the car and started driving, it hit the 55C treshold almost instantly and then peaked at 60C before I pulled the unit out of the dash to cool it off in an open air environment... Well, I live in the midwest US and it's ALWAYS hot in the summer and my car interior is black - not a good combination. So I set about to just wire the fan directly last night... And it does work perfectly. The only thing is that the way I have it wired, it's always on - even when the car is off. I KNOW what I did "wrong" (wired it direct to power rather than to the keyed power.

My question is: in the circuit that is described in this thread and web page, you used that pin on the circuit board... Does that have power all the time or is it "keyed"? Meaning, in the car, does it have power only when the ignition key is on or does it always have power?

BTW... on my drive into work today with the fan going full speed (my fan is actually one from a Cyrix 150 CPU and is about 14 mm tall and don't seem to have any issues), I couldn't hear it under normal driving conditions with only a slight whine at slower speeds/RPM's... And I maxed out at 28C with an ambient temp of about 72F (Sorry for mixing my C's and F's... I can't do the conversion in my head)

Jim

OK, what it's showing is that as temperature increases so does the fan RPM. This means that the hotter it is the more the fan will cool the player. When I ran the test I covered up the vent holes so that the temp would climb nicely and whenever the temp indicated by Hijack (thanks Mark ) increased by another degree C I would take a reading of fan RPM. This at least proved to me that the circuit worked over the kind of temperature range that I need. I'd like to do some tests over a larger temp range and see if I can increase the rate of change of fan speed against temperature. It seems to work very well but it's early days and I still need to take it on a longish run to see how it performs in the car.

Hope this clarifies things a bit .

If you use one of the pins marked "Fan" on the main board for the +ve you'll find that the fan is only operational when the player is powered up. If it goes into sleep mode the fan will stop. I believe this is what most people that have fitted a fan have used (including myself).

There's quite an extensive thread about me fitting a fan. Here's a link to that thread (you'll find a photo of the Fan pins attached too).

Thank you. Will do some rewiring tonight...

Yeah. I worked it out a few minutes after I posted that Thanks for taking the time to explain it to me though. I think I've been having a bad brain day

- Trevor

Ok, this may be the real way to do this:
http://www.telcom-semi.com/download/lit/pline/analog/thermal/dcfan/21450a.pdf

8 pin chip + one transistor to drive the current.
The chip has internal temperature sensing with ~ 1-3 degrees accuracy (good enough), and PWM output to drive the fan. You order the chip with pre-determined tempL and tempH parameters, which is a bit restrictive, but it will give us the ability to have a fan that is off until eg 35 degrees, and then ramps up to max at eg 50 degrees.

Or this one: http://www.telcom-semi.com/download/lit/pline/analog/thermal/dcfan/21448a.pdf

This has user-settable temp thresholds, utilising an external thermistor.

Looks good. I see it uses Pulse Width Modulation to control the fan speed. This was one of the alternative methods I thought about when I was first contemplating making my controller. I decided against it because I was worried about the possibility of introducing electrical noise into the system. I might still have a go at it though, just for the hell of it. Does anyone know whether this would be a problem?

Something else I noticed was that it's called a
Tiny
Integrated
Temperature
Sensor

edit I notice both of these use PWM.
edit (again) Aha, I've also just noticed that they recommend putting a capacitor in there to reduce electrical noise. Doh!!

Something else I noticed was that it's called a
Tiny
Integrated
Temperature
Sensor


Yeah, but does it come bundled with an Asymetric Sensor System? How about a Compact Uniform Networked Thermostat?

Sorry, I can't resist.

What about a Wide Area Network Kludgometer.

(to be honest I'm not sure whether the non-brits will get that one.)

or even a Binary Large Object With Java Overlaid Bistable

Yeah, we know what a "wanker" is over here. I think it's what the W stands for in George W. Bush...

Hehe.. Kludgometer. I like that.

In reply to:

---

... they recommend putting a capacitor in there to reduce electrical noise.

---

Did they forget to put it in the diagram (figure 1 on page 5)??? (Or am I even more blind/stupid than I thought?)

I bet he doesn't know what means though...or how to pronounce it

No, I couldn't see it either.

Did they forget to put it in the diagram (figure 1 on page 5)??? (Or am I even more blind/stupid than I thought?)

That'd make 2 of us....

Here's their complete list of fan controllers. I think I'd prefer to use one of the TC64X range so that I could use an external Thermistor. The 64 series also use an external resistor network for setting the temp thresholds which I would prefer. They even have a link to an Excel spreadsheet for calculating the correct component values.