So I have been playing around with Arduino for some time now and its great. But now is the time to move on into packaging a real automotive quality product suitable for the automotive under hood environment. So that means I need to put an Arduino on a circuit board along with power conditioning and power and signal conditioning.
The device will be a PWM radiator fan controller. The controller will take thermister inputs from both the water temperature and AC condenser fin temperature. It will output a 10Hz. PWM signal from 10% to 90% duty cycle. In addition I will use any leftover PWM outputs to connect LEDs to for trouble shooting.
So the things that need to be mounted to the board are an Arduino Nano 3.3 BLE, a 5V power regulator and opticoupler and a few caps and or transistors. Since this is my first attempt at making a PCB, I would prefer to mount the 28 pin nano and the 4 pin opticouler in sockets soldered to the board. Does anyone know where I can find these sockets or even what thet are called.
Then I need to put this all into a box with Automotive quality connectors. That means no USB. -40 to 120°C operation temps and 0 to 100% humidity along with high thermal stress cycles. The box will need 3 connectors. a 2 pin 5V power, a pin connector for temp in signals as well as a PWM pin for motor control, and seperate connector for LED outputs.
So for the moment I am interested in the construction of this box and things like good sources of parts, like boxes, connectors, and materials and proceedures to make PCB boards. I may start a new thread at a later date on the circuits themselves but I would like to keep this thread focused on construction.
Please send ideas or links to anything you feel would be approriate for the construction of this kind of hardware. I also found this video and I would like comments on it.
You can mount the Nano with 0.1" (aka 2.54mm) strip headers.
You can mount the with the same strip header or there is likely a IC socket that can be used for the Optos.
The strip headers come in long lengths and can be snapped off at any length you need. Here is an Amazon link strip headers
Automotive cannot have socketed devices. Vibration will quickly remove them!
Also you will not find automotive connectors to match your hobby demands. Find automotive connectors that will give you the connectivity you need. They will likely have more pins than you need. Real automotive connectors have a minimum order quantity.
Paul
More than likely I wont be the guy soldering the final installation. But I would like to try one myself with sockets. Then get someone who knows how to solder to do the real thing.
Honestly I've not looked. But with the leaning toward SMD parts I doubt the market is there. The strips are more than enough to hold the board once they are soldered in.
So I have been playing around with Arduino for some time now and its great. But now is the time to move on into packaging a real automotive quality product suitable for the automotive under hood environment.
I read your statement about under hood capability and assumed you are still in the test phase as I can't imagine using a production Nano under the hood. First the under hood environment is specified by the OE's as 125°C max. You won't get a Nano to work reliably in the under hood environment even if it doesn't actually hit the 125°C.
The box will need 3 connectors.
Unless there is a overriding need for more than one connector, the automotive industry (even aftermarket) will want only one connector.
high thermal stress cycles
Thermal stress cycles will effect more than the packaging. First to go will likely be the internal connectors, then the IC's if not rated for the stress temperature.
It will output a 10Hz. PWM signal.
Have you verified the fan will not create noise with a 10 Hz PWM? Was this fan previously controlled in an ON/OFF fashion. Typically the fan motors have internal filtering that is more that can be PWM's. When they make a fan for PWM they drop the internal capacitor to about 0.01µF
I don't want to appear negative but you sound like a neophyte in this area. Be cautious you don't promise someone more than you can deliver.
Power and IO could be 1 connector. A second connector could be used for trouble shooting. I really don't have any choice on the trouble shooting connector. Mercedes Benz uses multiple connectors on many of there modules. They even use connectors inside of connectors. Granted they are typically more complex modules. As far as aftermarket, my current module made by Flex-A-Light has four 6" 10 ga pig tails sticking out with instructions to crimp them to the fan wires and the battery and then another set of male spades for signal wires. And I don't trust the Flex-A-Light controller to work at 125C either.
There is no reason to place this module under the hood. I wouldn't do it and I wouldn't recommend it to anyone. But You know that that is where people are going to put it anyway. Just because there wiring is already there for it. That's just a fact of life.
I hope the fan wont create noise at 10 Hz. Its been running that way on every Mercedes Benz vehicle since about 2000. I don't know how they did it. That's between Mercedes and Bosch.
I may be a neophyte is electronics but I have been testing mechanical and electrical components for over 35 years for both helicopters and Jet aircraft. I have been dropping them, freezing them, shaking them, baking them, subjecting them to acidic acid salt fog, blowing sand on them and growing fungus on them. I probably will not be able to make one that will pass all that but I will make one better than Flex-A-Light.
I've not been involved in cooling fan controls, however I did design a HVAC fan controller for GM. That controller ran at ~ 18kHz.
Our electric fuel pumps would start to complain below 4kHz.
Knowing Mercedes I would be very surprised if they ran their fan a anything below 18KHz. But I don't actually know.
Here is a video of a Mercedes fan, using a motor manufactured by Temic. Hes running it at 11 Hz. I have run mine between 10 to 16 hz. Anything outside that and it will not run. Both Bosch and Temic run at 10 Hz on Mercedes
Here is another video of a Bosch fan motor for a Volvo. It looks exactly like the Bosch motor on the Mercedes except hes running it at 100Hz and claims it operates between 100 to 300 Hz.
I understand now. Your controller is commanding an existing Fan motor power controller.
Sorry to have not caught that from your "10% to 90%" statement.
For vehicle testing I'm sure the strip headers will work. For extra security you could have holes in your board and run a ty-wrap around the Nano.
Then I need to put this all into a box with Automotive quality connectors. That means no USB. -40 to 120°C operation temps and 0 to 100% humidity along with high thermal stress cycles.
All our automotive boxes were Nylon. We never looked into ABS or PVC but I doubt they would meet OE specification. I can't say for aftermarket.
This may not be helpful but Hammond Mfg makes a lot of small boxes, even some metal boxes. The state that they will modify them (for a cost). I have no idea if they make nylon boxes.
This was a prototype on vero, but the final product was Nano on strip headers on PCB.
Tom...
PS. Remember you may have to repair it sometime or troubleshoot the first product, make sure you have various test points and access to programming socket. Power supply test points and gnd are a good start.
(In other words make it SERVICEABLE)
ABS is definitely on the edge of melt down at those temperatures. What about the nano itself, where do you think the upper limit of temperature is in that. If you look at the reviews of aftermarket fan controllers you do see a lot of people reporting failures. I believe that's mostly due to heat. For testing the strip headers will be fine. I can have it soldered for the real thing.
Looks good. Is that an LM78xx power regulator. I'm considering that but worried about heat build up because I need to make the box as small as possible. Are you running a 5V Nano or 3.3 in an enclosed box. No issues with heat?
I really don't know about the µP temperature range. From looking at the µP partnumber on the Arduino board on the Arduino website it shows the processor as the industrial range -40 to 85°C
There is an automotive version with an expanded temperature range. Not knowing your timing you might want to check what devices you can actually purchase as a "raw" processor.
I was never able to find that kind of data, but its about what I expected. Where can I find these automotive versions. I can run around for a year with a stock nano. Mine wont be under the hood.
I just googled atmel328P then picked on the automotive version in the search results. Else go to the MicroChip website. You will find the automotive data sheet(s) where the part numbers are. I believe there are only two automotive versions, differing by the case configuration.
I actually started by searching in the partnumber of the processor shown on the nano board.
At this point I'm not going to design for under hood temperatures. The Arduino boards just cant take it. Possibly in the future I may create a design without the Arduino. Maybe I will just use the main chip and fully integrate everything onto a PCB board.
For some back ground and some videos on the application you can visit this thread.
For a discussion on mounting the Nano33 you can visit this thread.
At this point I have paired it down to three possible circuits that will do the job.
Circuite #1 Drive an PC817 opticoupler directly from the arduino. The opticoupler is a LED that shines on to a photo transistor which completes an independent circuit grounding the fan wire. The beauty of this is that the fan circuit is not in contact with the Arduino circuit. Electricity cannot get from the fan circuitry to the Arduino circuitry. There is a physical air gap between the two. The down side of this circuit, is the power required to drive the opti is about 11ma. This is within the 15ma current limits of the Arduino Nano33BLE which is the version I will probably use. But I don't know how that 15ma was derived as no environmental specifications or Mean Time Between Failure (MTBF) data is presented in the data sheet. 15 ma might be good for someone with a drone to peek in the neighbors window and bring the drone home before the battery runs out, but what about taking a 15 hour road trip. Better find something to reduce that power. Other not so good thing is the opti has a fairly low MTBF of 1000hr. That's rated at 50ma 100% duty cycle. Which in the real world driving at 11ma and 50% duty cycle and 2 hours a day 5 days a week equates to somewhere over 5 years. The way I dive this car 30 years.
Circuit 2 Drive fan using two transistors. The beauty of this is, the ultra low power requirement of the PWM IO pin ~1/2ma and the failure rate is so low on these I cant even find the MTBF. The bad is the fan and Arduino circuit can see each other.
Circuite 3 Drive the fan using 1 transistor and an opticoupler. Good news is the PWM IO pin of the Arduino is down to an ultra low 1/2 ma. Bad news is the MTBF is still relatively low. The one thing about the opticoupler is it is a 4 pin component that can be fit into an IC socket and be made replaceable. It can be carried around as a spare along with fuses.
I still have a long way to go on this but it is coming together. First thing I need to do is verify all this stuff works at 3.3V. All testing has been done on 5V to date but I believe its just a matter of changing resistors.
I also still need to provide mounting holes in the board to mount it to a case.
I getting more up to speed on using Eagle and have a board nearly finished, but I have a few questions. Below is a picture of my PCB board as it stands now.
what is the fat blue stripes in the center
As far as resistor selection. Currently I have a stock of 1/2W resistors that I'm using to test these circuits. The max current involved on this board is ~3.3V @ ~ .015A = ~ .05W. The other resistor is running about 0.02W. So is there any advantage of using 1/8W resistors.
