Fan and overheat controller for amp/radio

[sparky17]

I didn't see alot of mention of people really pimping out their radios, etc.. I was going to post up some random designs and schematics of stuff I came up with..

First thing I designed was a fan controller and overheat alarm.
Basically you set a temperature with a dial for your fan to turn on for your amplifier/radio, and another dial sets the 'overtemp' alarm buzzer at a different temperature.. The overtemp alarm could also be replaced with a power relay, to cutoff power to the amp on overheat instead.. This lets you run your amp at NO noise until it needs the fan, run a faster fan since you don't have to put up with the noise the entire time and also increase your duty cycle from the increased fan speed.

Thermistor wired up as a voltage divider, with a variable pot used as a voltage divider going into an opamp circuit acting as a comparator..

Basically a clone of this circuit, with different driver transistors:
[Please login or register to view this link]

(basically it compares voltage from thermal sensor to a reference voltage. Thermal sensor voltage changes with heat. Once the limit is hit, the opamp powers on the opamp which drives whatever).

R3/R4 are replaced with a variable 28 turn board mount potentiometer. Since opamps normally are sold as 'dual', there's an entire unused side! The thermistor is split between the two sides of the opamp, and an identical circuit with another potentiometer is used on that side as well.

Using the tiny prototype boards and through hole components, I was able to get this down to about half the size of a pack of smokes.

Circuit functions fine as described until you put it inside an amp.. Additional filtering work is needed with some more caps, and perhaps shortening the thermistor wires (Before filtering, it would not turn on at all while keyed up. After a few filters were added, it keys up EARLIER. When I get around to it I'm going to add a filter for the thermistor which hopefully should do it). After I get all the filtering points worked out, I'll post a more accurate schematic.

You could source all the parts and make your own for about $5-10 each..

[sparky17]

Both of these functioned fine in my lab oven tests [read: toaster oven with a thermometer probe]. Design works, it just needs shielding/capacitor filtering.



Version one. I had put antistatic foam underneath, heatshrinked and hotglued it for potting. Had to rebuild it several times as it burned up during testing.. Later realized that antistatic foam is very conductive.. Nice shiney excess hotglue gunk. After the fourth or fifth rebuild, I gave up and went to a larger board that was easier to debug.. Binding post so that I don't have to keep cutting my fan wire everytime I want to reoil it.. Buzzer/LED goes to the other high side of the binding post. Center negative is shared..

[ external image ]

No more duct tape. I found fingerguard for dirt cheap. Threaded straight into aluminum fins.

[ external image ]

Lug mount thermistor just barely peeks out of the case. Bolted onto the faceplate.. Not the best spot for a temperature reading, but I didn't have any thermal epoxy and my last attempts at making some were horrible... I'd rather lug mount than tape mount..

[ external image ]

Hotglue is much better than my usual epoxy for installs that need testing, lol. Buzzer mounted. Wires routed through vents for fan. Power wires soldered. I made a few 'posts' of hotglue at the four corners, then hotglued the board on top of that.. 1/8" gap so it doesn't short out, and I don't need to drill holes and bolt in nylon standoffs... and it's cheaper to boot.

[ external image ]

Probably be a few days before I continue work on this and get a revised schematic...

Thanks for looking.

[sparky17]

Redid circuit dead bug style. This time I didn't do the alarm, but it would be trivial to even add it to this.. All the wiring aligned beautifully for this construction method. Took me an afternoon to piece it together like a jigsaw compact...

Going to test then pot it.

Almost the same size as my index finger. Bulk is filtering capacitor..
[ external image ]

[ external image ]

[sparky17]

I don't know why the image won't rotate.. It's rotated locally, but here's the schematic. It's basically identical to the linked one in the first post above, but the voltage divider is replaced with a pot.. Goes together beautifully. A little dab of superglue until everything is soldered.

Tested it last night. Shorter wire lengths has almost eliminated interference problems. Fan kicks on a degree or so sooner if you're keying down, but it's barely noticable and kinda a good thing actually.. I think I'm using a PNP transistor, so I think I'm going to add a pull down resistor to have a fan 'idle speed' and then a 'high mode' with only one extra part..

Total cost is about $8 local.. Maybe get it down to $3 if the parts were all sourced online..

Everything gets a healthy dose of heatshrink...
[ external image ]

[sparky17]

Here's where things diverge... lol.

My local radio shack stopped selling thermistors, but I did find some at some others..

The circuit as described and shown above should be able to be built with parts at radioshack, except perhaps the thermistor.. Although they do list it on their website still..

I'm going to 'pump it up a notch' and use an ATTiny85 microcontroller which is the same size and pinout as the opamp but will let me get fancy...

ATTiny are $1.50 - $3 microcontroller chips, 8 pin DIP. I will convert the thermistor value to digital, convert it to a temperature, and then have variable fan speed with pulse width modulation.. Similar to how computer fans adjust speed. Will use the hall effect sensor/RPM sensor on the common computer fans to determine if the fan is stalled so I can prevent the mosfet/transistor/fan from burning up in such situations..

The circuit above works as a single speed, a 'dual speed', and I'm going to build a 'variable speed' perhaps with a push button to increment the temperature up/down a few degrees.. Could even have an LED blink out the temperature..

[sparky17]

Made another batch of four today to hand out to friends.

Working on learning Attiny so I can build a PWM one with variable speed and an RGB LED that fades to different colors to indicate temperature for a remote monitor.. I'm pretty sure it will remain the same size with those features.. If only I could put stuff INSIDE the capacitor, lol..

If there's any interest, and I still have parts, I'll write up a howto.

[ external image ]

[petriach]

I'm definitely interested.

How hard would it be to get a temperature sensor with a LCD screen showing the value(C or F doesn't matter)?

Would be nice to be able to show the temp of each transistor,maybe with a control board to disable the amp on an overheat situation.

[sparky17]

I'm learning as I go, but I know quite a bit of the details on how to add i2c LCD displays.. Be able to use a common cheap cord for remote installs, like cat6 or phone cable.

Panels aren't cheap tho. Cheapest panel is seven dollars plus supporting hardware...

I like the idea of temp per each transistor, you'd really need to cycle the display through all of them unless it was a big panel.. You'd need thermal epoxy to do it right.. There's also tape on flatter probes, but I haven't had any luck with the one I bought... Maybe I should try again.. I really liked the lug mounts, but the longest lead I can find online is two inches..

I'm going from cheapest/easiest and incrementally improving the design. remote LCD will be after I get the RGB led.. Could also flash out the temperature by blinking the light..

This analog design is as minimalist as you can get without going to surface mount... It's inexpensive, requires no board, and most if not all of the parts can be obtained at radioshack <thermistor might or might not be sold>.

[Hypo]

Really nice project article....Better than most that have been on some Ham sites lately.
Pretty good Photos, almost as good as VE7BPO does, just a little more light in his shots.
I have some computer fans and heat sinks laying around here.
This thread and article really raises the bar for this site.
i would love to see more like it. Maybe someone with a DDS controlled by a PIC or Atmel chip.
Looking at getting an 80M AM kit for myself and doing some NVIS around the South.

[petriach]

BTW Sparky,if you want thermal epoxy/adhesive,I would suggest the stuff made by Arctic Silver,it's mostly used for computers,attaching small heatsinks and the like to MOSFETs on the boards or controllers especially when overclocking to help with heat dissipation. It's not conductive,does have some capacitance though,they say. Keep it away from traces. Cheap'ish.

[sparky17]

Took about 20 hours, but I made a programmer, got code loaded on a chip, and got it to do something. Woohoo. I had to make 3 different programming cables, 2 different programming cords, and try about 10 different programming programs to get it going.

Now I just have to learn all the code/APIs since it's quite a bit different than normal C programming...

I added a PI network on the power IN line feeding the analog device. A 220uH inductor and two 0.1uF filtering caps, one before and after the inductor.. The 'interference' seemed about the same. You can modulate and have the fan kick on a few degrees earlier as a result. This tells me the noise is coming in from the thermistor line most likely... I mounted one outside a box and had similar results..

When board mounting these however, it would kick on 20-30+ degrees earlier, so interference is still vastly improved by using the dead bug method..

Working on getting analog to digital conversion going for the temperate into the microcontroller. Hopefully I can just take several temperature reading 'samples' and average them out or something if it's still a problem in that design..

I got basic PWM working with an LED for testing. I can dim it in and out (but not control the level of dim), and I wrote a quick program that cycled through mixing colors with the RGB LED.. For ease in install, I'm going to try to get the panel mount LED to also be a 'button' by putting your finger over it..Since the LED is three LEDs in one, I can have one light up, and another one 'watching' for a change in brightness caused by the light reflecting off your finger.. I'd hate to drill TWO holes for panel mount. One for the LED and one for a button.

Still have a ton of reading and learning to do. I ordered a used oscilloscope online for Christmas which will hopefully help me out in development..

Current ideas which I think I can pull off:
-Manually adjustable/programmable fan speed. Can set just 'on' at a fixed speed, set a 'high' and a 'low' speed limit... Some people volt their stuff and their fans run faster than they should...

-LED color indication of temperature (Theoretically I can do 16 million colors, the thermometer resolution will likely limit this to 100 or less)

-Adjustable temperature 'update' time. In case interference is severe, can stretch out how long it takes to update temp for fan and temp indicator LED.

-Programmable 'menus'. Using 'colors' instead of flashes when possible (If you've ever used a programmable LED flashlight, the flashes SUCK for figuring out where you are in a menu. 'Click button until you see blue' is alot easier than 'click button until light flashes 10 times'. If you've ever pulled ODB codes from your car, you know what I mean also. Menus would be where you would configure speeds, cutoff points, etc..

-Simple hand crimp connector. Still need to find a suitable one. Goal is to use commonly available 4 or 5 pin wire, with a connector that can just be 'snapped on' to the wire vs soldered.. This would make it easy to just cut the excess off the LED cable and hand crimp the connector. If you've ever seen CDROM cables inside a computer, that's how those are done.. Should take longer to find your scissors than to hand crimp the connector..

-Fan RPM sensor. Detect stalled fan, keep it from burning up. Also useful when setting a really LOW fan speed to make sure the fan is actually spinning. Sometimes you have to spin them up to speed and then lower them, vs try to start them at a low speed..

I'm cycling through 4 different other projects, and do a couple hours a day on this one, so it's slow and steady progress. Once I get the analog to digital conversion and PWM code figured out a bit better, it should be easier/faster to do stuff.

[sparky17]

As far as the 'probe per pill' w/ LCD display request above, here's my counter-retort

Not cost effective to build. $25 in probes and LCD alone. Add knobs, switching transistors and a case, and you'd be well over the $38 cost of this thing below designed for computers.. Saw this at the local computer store. 4 temperature probes and 4 fan controls. Displays fan speed and probe temperature with a speed adjustment knob for each. Can be set to automatic or manual mode.

[ external image ]

[dime196604]

That is some crafty work there. i like it. you gonna put it into production?

[sparky17]

The analog one no.. I bought all the thermistors in the local area to make some of the analog ones for the locals.. The remaining thermistors I have left (three) are going towards the digital ones.. Depending on how that turns out I might consider going into production...

[sparky17]

Digital one is still work in progress. I have the breadboarded prototype done and soldered most of an 'actual size' prototype. Sidetracked by another project I just finished..

Need to test out the prototype inside/on a box to see the effects of interference from the kicker.. Analog to Digital temperature readings can be heavily affected by interference, so I need to compensate in hardware or software depending.. To keep parts count down, I'm hoping to do it all in software and just do averaged readings..

If that fails, I'll switch to a digital thermal probe that pulses the actual temperature out to the microcontroller.

Played around with the internal temperature sensor on the microcontroller for my last project I just finished.. I'm debating using that to detect the 'ambient' temperature, so that we can detect if the vehicle is hot or just the kicker.. Or perhaps set a 'target' temperature for cutoff, to keep the fan from spinning in vain when it's close enough to 'ambient'...

[sparky17]

Bought a large fan, and wanted a way to vary the speed. It's a 10 watt fan and RadioShack only sells 3 watt Rheostats.. Decided to use a potentiometer as a voltage divider to drive the gate of an N-Channel Mosfet (IRF510). This gives me variable voltage on a knob. Put it in a project box with binding posts so I can easily swap the fan out, and use it as a lab power supply. Also, all my 12v supply taps are tucked away far from my desk, I can use the binding posts as a handy nearby 12v source for my hobby chargers and whatnot..

Super simple. Potentiometer wiring is the far ends go to + and - and the center pin on the pot goes to the gate of the mosfet. Drain goes to + on the pot and source goes to the + binding post/fan. Negative on fan/binding post goes to - on pot.

Right now it's just on/off with the switch. I'm going to go with a 3 position DPDT so I have on/off/temp control. Going to add an RCA jack to the back for a thermistor to plug into and tie the DPDT switch to switch the potentiometer from fan speed adjustment mode to temperature setting mode.

Pot center goes to - on opamp, thermistor and a 10k resistor goes to + on opamp and output of opamp goes to the gate of the mosfet.


[ external image ]

[drdx]

That's some cool stuff. I wonder, after some refining, how much demand there would be? A hot spot, no pun intended, for a nice automatic cooling system would be in some of those really high powered exports.

I used to run a thermometer on my old amps with the remote wire sensor just to watch it. These days, I just run an swr meter on the dash that has the adjustable swr warning light so if I'm keyed and something is wrong the light is there. My home radio and power supply have fans and automatic as I think just about all hf rigs are these days but it would be nice to see more in this level of radio, meaning the cb/export scene.

Then again, if you loaf it and don't be mean to stuff it tends to last on its own. I don't think anyone should ever run an amp without the fans running all the time yet I did it for a while when younger. Why not run it as cool as possible at all times? Is there a benefit to regulating it at "warm" but not hot, as far as stability or something? How much would a typical unit, especially a sideband one, benefit from having some temperature control in a vehicle?

During the summer, I even run the fans for a while and let the cool AC air bring the amp's temp down before talking. Heat is heat and I'm in Texas. It can get to 140F inside a car in no time here. In the winter it is easy. They start cool and if you do a bunch of talking just bump the rear windows down and let the cool air swirl in. My stuff is all in the back on remotes but the same idea applies as long as the amp is not buried under a seat or something.

Don't the Tx star sweet 16's and dx1200's and others have a thermometer controlled fan system, or are they on all the time?

-drdx

[sparky17]

The heathshrinked analog controllers are about the size of my thumb.. The capacitor is the majority of the size. They could be fit into any radio..

I think the analog one could use a silicone diode as the 'thermistor' instead. The reverse current voltage follows a linear path with temperature with pretty good consistency.. You'd have to calibrate each one individually if you ever wanted to correlate that to a temperature however.. I was saving playing with that for the digital one, so I could internally calibrate the diode against my reference '72F' house temperature on the first bootup... Anyways, it would be 10 cents + wire to do that vs 1.50 for a thermistor, and all the parts could be had at radioshack... Last time I checked, all the opamps were mosfet varieties however, which don't go down to 0v (as far as I know). I'm not sure if this would be a problem or not. If you fed a fan very low voltage, you might stall the fan, and depending on the fan, it might burn up. The mosfet opamps on the other hand, might not put out enough voltage to turn on the fan driver mosfet, so it might not be a problem using all radioshack parts...

The problem with the analog design, is that even with keeping lead lengths as short as possible, we still get *some* bleedover into the controller circuitry.. (Hence the heatshrinked packages vs board mounting). I tried various filter caps everywhere, without much benefit. I think a diode on the thermistor line, or even just using the diode for the thermistor, might help this a bit more.. If you board mount this whole circuit, the fan will turn on LATER than expected based on temp.. With the miniaturization, I got it to turn on EARLIER than expected when keying down, so at least it's failsafe, but annoying.. Going from PNP to NPN might help here, or perhaps just going MOSFET and using a pulldown resistor.. I'm not sure.. Anyways, running the temperature on the 'warm' side prevents the bleedover from kicking on the fan everytime with modulation, and now it tends to stay on when its warm, and stay off when it's cool. There's a little bit of hysterisis, but ideally it would be a few degrees wider, so that it stays on/off, but during modulation while you are close to the 'turn on' temperature, the fan can cycle on/off with keydown.. (I looked at doing a schmitt trigger, but I think that would be too big of a hysterisis)

The digital design with microcontroller would add about 3$ to cost, and would give the ability to have adjustable hysterisis.. So we can solve these problems in software (hopefully.. I have my prototype digital board built, I just need to find the time to test it out with RF abusing it).. Side-benefit of digital, is that we can use the unused pins to drive a display (RGB led) and customize its settings/behavior (through a button).. A digital circuit could be built exactly like the analog circuit, I believe with identical pinout, with fixed settings for hysterisis, and not need a button/LED.. (It does need a 5v regulator chip as well however).

So ideally, no.. The fan would turn on the minute the box got above ambient. The higher temperature is to compensate for difficult to solve problems..

Goal here is to make useful accessories, using minimal parts count and cost for friends.. The parts involved are perhaps as low as $3 for the analog fan controller. People don't like to solder, so the binding post gets added instead of just wire leads that people attach to power/fan.. So that's another $1.50.. The more 'commerical' and 'general public' these things are made, the more the parts count goes up as well as the cost and difficulty in building..

At the local surplus place, where I buy all my prototyping parts, the cost is about $10 per unit (for SURPLUS parts). Maybe 15-20$ in parts if you were to just scrounge them up locally at normal local vendors..

What I'd think would be neat is to settle/finalize on designs, and do a 'group buy' for parts, and send out unassembled kits. Detailed picture soldering instructions would be made up for each item.. Other people could place bids on offering soldering/assembly services to that someone could get a little extra money, and people could get a good price on assembled cost.. I spend all my free time prototyping and doing other projects and couldn't offer assembly at a reasonable cost.

For any microcontroller projects, I could upload the software onto the chips for free before kits went out, or people could modify and/or write their own code for them and upload them as well. The programmers can be made for under $10.

Goals are:
1) Low Cost for parts
2) Low parts count - to keep assembly as SIMPLE as possible
3) As many features as possible, while observing goal 1 and 2.

I see all these things that can be made for dirt cheap, but unless they were made in bulk or assembled by the individual, the cost would be high...

Future plans:
~$40 - 1kw (With another chip expandable to 100kw) 0-500mhz digital watt meter/field strength meter with LCD. Could also possibly do modulation and SWR. I think it's possible to have an analog meter with needle as well as the digital meter at the same time, but would 'auto scale' the needle (I think it's easier to see the needle not moving in the corner of your eye and know something is wrong vs looking at a digital readout, but the digital readout offers the choice of looking closer at it for more accuracy)

~6$ - Digital fan controller with intelligence, hysteresis, RGB LED temp display indicator and programmable options.

~10-20$ - computerized VOX. For talking with a headset, through skype (keydown from your cellphone) or noisetoy...

The variable voltage fan controller I made above, with the thermistor temperature options I plan to add would probably cost about $10 bulk for parts..

I think the volume is too low to outsource/save money by having a large factory make these things (maybe I'm wrong), and the profit/effort wouldn't be worth my time to solder things up vs enjoying my time off work (I wouldn't quit my engineer dayjob). Realistically, I don't think you'd have much success selling accessories to CBers unless it is $20 or $30.

Group Buy Kits & Soldering Service Bids.. Keep the parts consistent so people can follow assembly instructions (and share on bulk savings and shipping costs), and offer a way for there to be competitive pricing for the people that don't want to be bothered with soldering..

[sparky17]

Pictures worth a thousand words.. I guess I already spent 5,000 words, so here's some pictures.. Picture of underside was before I added the thermal control. RCA jack and a bunch of heatshrink to make a more durable thermal probe in the dangerous desktop environment.. Attached the thermal probe to the SO259 screw. Manual Mode / Off / Automatic Temperature Mode. Going from a 28 turn pot to a knob is actually pretty good. The sensitivity is more than adequate. As a bonus, the interference issue has gone away with this design. Keying up/unkeying did not change its behavior. Woohoo.

I'm trying to encourage all of you to build some cool stuff, so post a message if you want me to post construction details/schematics, otherwise I'll move onto working on the next version


[ external image ]
[ external image ]
[ external image ]
[ external image ]
[ external image ]

[sparky17]

Added small 28 turn pot to adjust fan speed while in temp mode (Between switch and LED light). Had to match the metal theme... The output of opamp goes into a voltage divider which then drives the mosfet.

I leave the fan at about 30% speed, just barely above the audible threshold while the radio is on. I want to know when it kicks on so I know when to start monitoring temps.. Even at 30% speed I was able to maintain ~80-90F temperature during a 10 minute endurance test. woot.

Got some clear lacquer. I wanted to try doing a few more with a better brushed aluminum job and a clearcoat to preserve it.. I'm doing form over function right now as far as the metal parts.. I just find them so dang sexy. Stainless screws replaced the zinc screws. Wanted to do the Ace hardware chromed screws, but I can't justify $2.00 per screw. They did have some rivet head allen wrench driven machine screws that were also attractive, but I didn't want to risk stripping the threads out while tapping the plastic..

[ external image ]