Carputer/ CarPC - Has anyone here built one?

[Scheer]

11 hours ago, Huntsman said:

There are reasons people don't go full on buttons and dials for everything and a million sensors feeding data into the carputer thing.

 

1. Power. Most car can only cough up like 20W of extra power after allocating them to the electronics already in the car. Lightings, power windows, wipers, stereos etc already takes alot of power from the car's alternator. The problem isn't about what voltage and how stable a voltage can you get (this is relatively easy to solve if you know electronics) but about how much power is still available for you from a stock car. The power has to come from somewhere. Modding the car to get more electrical power meant you sacrifice somewhere else, most notably horsepower at the wheels.

I'll agree for the most part with everything you said other than number 1.

 

Cars have a lot more than 20W of extra power to spare, in most vehicles adding in a 300W amplifier for speakers is no problem at all, above 500W and you may run into problems, but not many computers are going to be pulling that unless you are gaming/rendering and driving for some reason. Its also going to depend on the car, cheap cars typically skimp on the alternator and may be an issue, a big diesel truck and we've ran 1500W inverters for power tools without a problem.

[straight_stewie]

9 minutes ago, Scheer said:

a big diesel truck and we've ran 1500W inverters for power tools without a problem.

I agree. You just have to do research into your own car. My Ram 1500 will run two 1500 Watt (RMS@1Ω) with the only hiccup being dimming brights during heavy bass.

[Huntsman]

4 hours ago, straight_stewie said:

I guess I got us off track, but I'll agree with that: The original problem was just to control the audio. The hardest part of this is going to be getting the monitor to look right in the dash. The second hardest part will be designing an easily useful UI for it, so that you can use it while driving. I assume that OP has basic skills with an Arduino or Pi, so he would be able to map the steering wheel buttons to some type of interface and then pass it to the computer for processing easily enough. 

Keeping in mind that the point of this is to control the audio system, you can really think of this as nothing more than a very complex car audio project. That said: There are other, easier (although more expensive) options that will drop right in.

I'm pretty sure he's talking about more than just audio lol... 

 

" I'm aware you can implement tablets into vehicles, but I'd like to do something that can be better integrated into the vehicle; integrating wheel controls, great sound quality DSP's for the accompanying SQ system build, most likely with a 10" Liliput touch screen and perhaps secondary screen modded into the dash. "

 

Unless wheel controls meant buttons on the steering wheel... then disregard the wall of text I posted just now.

[Huntsman]

3 hours ago, Scheer said:

I'll agree for the most part with everything you said other than number 1.

 

Cars have a lot more than 20W of extra power to spare, in most vehicles adding in a 300W amplifier for speakers is no problem at all, above 500W and you may run into problems, but not many computers are going to be pulling that unless you are gaming/rendering and driving for some reason. Its also going to depend on the car, cheap cars typically skimp on the alternator and may be an issue, a big diesel truck and we've ran 1500W inverters for power tools without a problem.

For more than 20W you will need to run your own cable from the battery terminals itself. If you do that, yes, draw as much as your alternator can provide. I'm providing the worst case scenario there.

[straight_stewie]

6 minutes ago, Huntsman said:

I'm pretty sure he's talking about more than just audio lol... 

I assumed that because he said "wheel controls" and then listed a bunch of audio only stuff he was referring to the audio wheel controls, like volume and tune...

OP has been mighty quiet. Where'd you go @RogueCow

[Viperxavier]

I went with an atlas cpc1000 a few years ago for my eclipse show car.

But now you could just do a micro atx system in the glove box fairly cheap Or mini or micro system and get a ton of features and get a monitor off ebay fairly cheap with a touch screen.

[RogueCow]

On 7/20/2016 at 10:18 AM, Huntsman said:

There are reasons people don't go full on buttons and dials for everything and a million sensors feeding data into the carputer thing.

 

1. Power. Most car can only cough up like 20W of extra power after allocating them to the electronics already in the car. Lightings, power windows, wipers, stereos etc already takes alot of power from the car's alternator. The problem isn't about what voltage and how stable a voltage can you get (this is relatively easy to solve if you know electronics) but about how much power is still available for you from a stock car. The power has to come from somewhere. Modding the car to get more electrical power meant you sacrifice somewhere else, most notably horsepower at the wheels.

 

2. Performance. Sure, you could build a 22 core Xeon with 10x 980TI and get some rockin performance but you'll hit the power constrains before you hit the accelerator. Not to mention are you going to wait 1 or 2 minutes for everything to boot into Windows, launch whatever app you would like and what not? No, in the realm of car infotainment things have to be ready as soon as the engine starts. Imagine having to wait 60 seconds before you could attempt to load up your songs or change the air conditioning settings. Going low power? Off the shelf SBCs aren't going to boot any faster than a Windows 98 machine unless heavily modified and slimmed down.

 

3. Integration. Unless you're a very very dedicated engineer, chances are that you wouldn't be able to do what Tesla or Ford or [insert manufacturer name here] can do. Yes, they have a "carputer" that can handle everything from air conditioning to engine tuning. But that takes a big team of hundreds of staff and not to mention specialized/custom parts that are specifically picked out to work with each other. You car might have most of the sensors already but tapping data from them and getting them to work together while being controlled by that carputer thing you've made there is not going to be easy.

 

4. User experience. Unless somebody with the required engineering skills port something like Android Auto to something like a Raspberry Pi or Odroid, chances are that ready made software for these platforms don't really work well in an automotive environment. A automotive software which distracts the driver from driving are just not intuitive to use in that kind of environment. Kodi isn't going to cut it if the driver have to fumble with those small buttons while driving. And we haven't gone into things like connectivity with phones yet...

 

The point is, if the car doesn't already come with a good enough infotainment system, it's really really hard to DIY yourself a solution to make it worthwhile. It's cool as a swag thing to show off to friends but it is by no means functional nor practical.

Hahaha, I like this post.

 

1. Sounds system will be about ~1500+ watts, the car will most likely have a beefed up alternator or an extra one, so power draw while the car is running is of very little concern (though still the goal is to be drawing 20-50w when all is said and done). And this will be going into a Lexus; not very concerned about robbing the horsepower.

 

2. The vehicle will have keyless entry, so I would look for power off of a line that detects when the key is in range (the body control module would most likely receive some form of power when this happens), and begin booting the system then (is the plan at least). Then modifiying the bios so some sort of dummy screen is immediately displayed upon vehicle start if the comp has not yet fully booted.

 

3. This is kind of funny, I actually am an engineer that works in powertrain feature calibration for Ford lol; so I know for a fact that you are right about that.
This is going in a luxury car, so I don't care about engine performance parameters at all (which can be siphoned off of the OBDII port easily enough anyways),

trying to write your own software to be controlled by a custom made PC would be very very dangerous and take forever to do... Just audio, nav, and steering wheel integration for me:)

 

4. I am leaning towards starting with something like Centrafuse Auto, and then perhaps doing a custom skin. I've run a trial version on my PC and it seems to

do everything I'd need it to do, and operates like my current double din head unit in one of my other cars; it could just look a little better,
so I think a skin for it would be a nice touch.

 

In the end you are right, it is a challenge, and could be done easier/better (buy a better car, buy a double din head unit, stick it in the dash and be done; though it'll look like aftermarket garbage), but where's the fun in that?

 

---

 

As for what I want out of the project:

I want to have navigation, audio, steering wheel controls; maybe even some RGB ambient lighting control (cuz everything is better with RGB).

Still undecided on touchscreen/ having a custom controller put into the center console

 

I'm an all or nothing kind of guy, as most of the times it's easier to do a complete system standalone, than integrate into what's already there.

So i would most likely do a custom dash panel for the screen, fully custom audio (bypass amps and existing system entirely), and a custom carputer to run it all.

 

The only "integration" from the car would be the power, and I'd like to have the steering wheel buttons input as well (this shouldn't be too difficult as I have

seen custom Arduino-based chips that decode the control buttons to act like keyboard strokes; just setting those up as macros to control the carPC front end

would be cool)

The only thing I don't really know too much about is how to build a pc that isn't a power hungry gaming rig; I watch most of LTT's videos, but very

seldom do they ever talk about power-sipping parts, or components that aren't gaming oriented, so I'm looking for background from someone who

has built a carpc with some success, or a super low power pc.

Pro-Tip: Here is a great resource for custom CarPC parts (power supplies and such) and info
https://store.mp3car.com/
 

[RogueCow]

 

On 7/20/2016 at 9:45 PM, Scheer said:

I'll agree for the most part with everything you said other than number 1.

 

Cars have a lot more than 20W of extra power to spare, in most vehicles adding in a 300W amplifier for speakers is no problem at all, above 500W and you may run into problems, but not many computers are going to be pulling that unless you are gaming/rendering and driving for some reason. Its also going to depend on the car, cheap cars typically skimp on the alternator and may be an issue, a big diesel truck and we've ran 1500W inverters for power tools without a problem.

Agreed on this; I am running 2000W in an otherwise stock car, and just get a bit of headlight dimming when a blast my system, so far no alternator or battery issues after 4 years.
 

On 7/21/2016 at 1:43 AM, Huntsman said:

I'm pretty sure he's talking about more than just audio lol... 

 

" I'm aware you can implement tablets into vehicles, but I'd like to do something that can be better integrated into the vehicle; integrating wheel controls, great sound quality DSP's for the accompanying SQ system build, most likely with a 10" Liliput touch screen and perhaps secondary screen modded into the dash. "

 

Unless wheel controls meant buttons on the steering wheel... then disregard the wall of text I posted just now.

Sorry, wheel controls did just mean the buttons on the steering wheel
No autopilot for me ahahah, didn't work out too well for Tesla, don't think I can do any better... Just volume up/down and track changing is all i need

 

On 7/21/2016 at 1:49 AM, Huntsman said:

For more than 20W you will need to run your own cable from the battery terminals itself. If you do that, yes, draw as much as your alternator can provide. I'm providing the worst case scenario there.

Yupp, most wiring in vehicle is pre-fused and seperated so each line is very low power and unusable for amps/ other equipment (carPC)
 

 

Sorry for being quiet, forgot that I turned off email notifications and just thought no one was interested enough to reply lol
 

[Huntsman]

Damm.. At least you had some background on the vehicle's powertrain. I've built a OBD module trying to get telemetry and do some controlling but can't seem to get anywhere except display RPM, speed and temperatures. Attributed it to my limited ECU.

 

I did tried to built a carputer/infotainment system last year but immediately got discouraged when I found out that Android Auto got big and the fact that it's just not going to perform well enough to be practical. Also the existing diy solutions didn't really fit my needs.

 

Regarding the low power PC thing, try some Intel Atom boards or the J1900 line of products if you had to go x86. PSU is easily solved by buying those micro psu from ebay, those that stick on the 24pin power and gets 12V in.

[Huntsman]

5 hours ago, RogueCow said:

2. The vehicle will have keyless entry, so I would look for power off of a line that detects when the key is in range (the body control module would most likely receive some form of power when this happens), and begin booting the system then (is the plan at least). Then modifiying the bios so some sort of dummy screen is immediately displayed upon vehicle start if the comp has not yet fully booted.

Smarter and less wiring solution to this is to detect battery voltage. Have a small low power micro sense the battery voltage. When the car is off, the voltage is going to be lower since that's the steady state for not charging, but when the car starts up the voltage is higher and that's when you know the car is started.

 

I did this for my car's automation and it works reliably well.

[johnny5c]

On 7/18/2016 at 2:18 PM, manikyath said:

the biggest issue is finding a power supply solution for in the car that makes sense.

 

On 7/18/2016 at 2:23 PM, Julian5 said:

I would say a 12v to 240v adapter and a pico power supply 

Then an itx board with an i3 and 4gb of ram

 

On 7/18/2016 at 2:32 PM, manikyath said:

that's not a sensible solution. you're converting 10-14 volts to 230 volts AC, then to 12, 5, and 3.3 volts.

 

the losses there are so hilareous that you just shouldnt..

i'm sorta maybe looking into solutions that would allow me to molest a UPS to run my server straight off the battery voltage, and i'm stuck in the same void.

 

On 7/18/2016 at 2:46 PM, Julian5 said:

yeah i know, you could regulate the voltage to a steady 12v and then step  down the voltage to 5v and 3.3v, I have done similar things before.

 

On 7/18/2016 at 2:49 PM, manikyath said:

well, if you can make/get your hands on a decent step up/down converter for 10-14 volts to a steady 12 volts, add a beefy cap, and some chunky off the shelf 12->5 and 12->3.3 regulators and you're good to go.

(and the vareous voltages i missed, an enable circuit, and a car USB charger bodged as a +5Vsb)

 

I built a 12 volt HTPC for my RV using this PSU,

https://store.mp3car.com/products/m4-atx-hv-250w-intelligent-dc-dc-psu-power-supply-unit-6-34v-input

 

Puts out a constant voltage on the 3.3/5/12v rails no matter what the input voltage is.  Has no problem running a i5 2400 on a matx board with a SSD and two HDDs.

[idiesel2]

I obtained an old Microsoft Surface Go 2 Tablet. I set up a "cradle" in the dash with a USB interface. The other end of the interface connects to a USB hub with pass-through charging, running from a high-wattage cigarette lighter charger, because it also has a port for charging my phone.

 

I set the tablet to hibernate or sleep when the car is turned off. I settled on a always on cigarette port and setting the tablet to sleep after 2 hours of idle. The screen is set to blank in 1 minute of disuse. 

 

I like the Surface Go  because the form factor is the same from ver. 1 to 3. I also like it because it is removable from the dash to do updates and has a decent screen brightness.  I could put a SIM in it, but it's easy enough to just tether.

 

My favorite feature is the traffic map integrated with camera links in Edge. Not a big windows fan, but its easiest to customize. Still working on the interface customizing. Any ideas in that direction would be nice!

 

PS. I use a USB DSP to output great-sounding audio!