What is the spec of the psu in the car?
I came here in hope that someone could help me out with an issue that I have been having with my new carputer project. First off, I just want to get some of the specs out of the way. The motherboard of for the computer is an Intel DN2800MT Cedar Trail board with an Intel Atom N2800 Dual-core 1.86Ghz processor on it. I have 2 GB of Crucial DDR3 RAM installed, a 32GB OCZ SSD, and an internal 802.11n Intel Centrino Wifi mini-PCIe card. The power supply powering the board is the Mini-Box DC-DC Buck-Boost USB Programmable PSU. The PSU is currently set in automotive mode and programmed to output 12V at all times.
The problem I am having with my carputer is that I cannot always get it to POST when I have the mobo in the car. Sometimes it will post very quickly, while other times it will take about 15 seconds to POST, and then there are times it just won't POST at all. When the computer does POST however, it operates completely fine and doesn't have any problems at all. But here's the thing, when I take the computer inside and hook it up to a 12V DC supply from the wall outlet, it POSTs fine all the time, and the POST is significantly quicker as well.
Thinking it was an issue with the motherboard, I got that replaced and realized the error kept happening with the new board. I then thought it was a display issue so I got a new logic board for my LCD display and that is also not the cause of the problem. The only difference between taking the computer inside and having it in my car is the power source. But why would that cause the computer to get suck when POSTing? I have eliminated all other sources of problem problem in USB connections as well. I am stumped. Does anyone have any ideas? Could it really be a problem from a ground loop?
What is the spec of the psu in the car?
Like I said the PSU is manufactured by Mini-Box and is a DC-DC Buck-Boost power supply. Here is a link to the product website: http://www.mini-box.com/DCDC-USB
These are the main specs they have listed:
- Wide range input, 6-34V
- Programmable output 5-24V
- Default output 12V
- High current, up to 10A
- Applications: power for motherboards with single rail power from any voltage to any voltage, laptops, custom electronics with buck-boost DC-DC.
- Programmable via USB interface
- Sends ON/OFF pulse to motherboard
- 4 modes of operation: DUMB, Automotive, Script or UPS
- 100% Solid Polymer Capacitors, Japan
This isn't like the typical picoPSU or Mx PSU since the output voltage is a level that is programmed into the unit, and I can write custom scripts for it. I cannot use a different type of PSU because the DN2800MT motherboard only has a 1x2 power adapter on it, and a cylindrical DC power adapter on it.
I wish I would have found those power supplies sooner or I would have gotten one! But regardless I can't understand how the computer being underpowered would cause the system to 'hang' at post. When I initially started working on the project, I didn't have the AC-DC wall adapter, so what I would do is use a car battery (an extra one I have) to power the DC-DC regulator. Everything would work fine when I had that setup, so I'm not so sure it would be logical that the carputer would be underpowered. The one thing I am seeing with the supplies that you mentioned and the DC-DC power supply, is that those supplies do seem to have more circuit protection than the one I have. I'm thinking that alternator noise could be the problem, as I have studied ground loops in the past and know they are capable of affecting sensitive circuitry. I'm just not understanding how the computer is hanging on POSTing, but then works completely fine once it has posted.
You mentioned USB devices. I assume that:
1. When you test on the bench you have no USB devices plugged in?
2. When you test in the Vehicle you have no USB devices plugged in?
3. No line out, In or any other connections to equipment in the vehicle.
If so, have you tried powering the unit on the bench with the input voltage at say 10 volts and if possible try quickly swinging the power supply input voltage down and up a few volts on the bench supply.
I have seen various intelligent power supplies take a dive if the input voltage dips and rises just as they load up ( power up the PC).
Try duplication the power source on the bench. Power it from the vehicle battery with the same length and gauge of cable and include any fuses that are inline in the vehicle.
In most cases like this we really have to logically duplicate the vehicle system on the bench if we're going to find the culprit and rectify the problem. As SNO said it may be border line on the PSU. Although the PSU can power the PC all day long once it's running, often the highest load on the supply is right at start-up and BUCK SMPS draw more current as the input voltage drops. If USB devices have been eliminated then something on the power side of the installation is most likely just tipping this over the edge.
GA-Z77N-WIFI, i5-2400S, 8GB, Intel 520 128GB SSD, M4-ATX Modified, 2 Rev Cams, 2 Web Cams
8" 16:9 TRANSFLECTIVE, Win8-64, Dual GPS RX and Garmin PC + Odyssey Nav, FM-DAB+, BB-Rec
T-Screen HVAC control, custom microcontrollers, code and FE. CarPC Project
1 & 2. Correct. I have tested the computer with no USB devices plugged in both inside and outside the car. The only device that is plugged in when inside the car that is different from outside it, is the VGA LCD screen, however, even that I have taken outside the car to test, and operates normally when I test it inside.
3. You are correct.
What you're saying makes sense. I think what I will do is take my volt meter and do some current measuring with the circuit and see if i can identify the problem. It may be due to a bad ground since for some reason I get a lot of engine noise through my speakers before the sound drivers for the computer kick in, so that may be the source of the problem. If it's not the issue, then I will look into upgrading the power supply. I will also give your idea of dropping the power and then raising it a couple volts during the POST a try. If anything I would think when I have it in my car and the alternator is running, the power supply would have more available current to draw from. The intelligent PSU does have a couple second delay before switching the computer on, so the alternator of the car would already be going when it tells the PC to fire up. I will still give your idea a shot though regardless.
I appreciate the help! Thank you!
Hey guys I don't really mean to bring back a semi-old thread here but I bit the bullet on the CarNetix CNX-P2140 power supply. I hooked the DC-DC PSU directly up to the battery in my car with a direct ground and power connection through 6 gauge wire with no substantial results. It did help a little bit but POSTing still takes forever, and sometimes it will still get stuck during the post, it just doesn't seem to get stuck as much. It's still a PITA though and the POST times are way too slow for what I want so I'm at the point where I just said screw it and bought the CarNetix. (Literally half the time the computer takes to boot is spent on the BIOS screen) There is no other way I could have replicated my lab power source any better in my car. When I took my computer back out for the first time in awhile and hooked it up inside using my AC-DC power adapter, the BIOS posted in about half a second, so I can only pin it on the power supply that's the problem. Hopefully this will be the last bit of money put into my project. Thanks again for the help guys!
Have you checked that your cars battery is healthy?
My car battery is healthy. I have a 1100rms total sound system hooked up in my car through 0/1 awg wiring, an Optima Yellow-Top battery, and a big-3 also done with gold-plated battery terminals. This issue happens regardless my car is turned on or not, so I doubt it's the battery or any issue on that end.
I have come down to a conclusion on the problem though, and it deals with the Intel DN2800MT motherboard (and apparently other Intel motherboards as well). The issue IS BIOS related, and it seems to have problems that jump out with the connection of different USB devices. Since my project has a very custom display unit and is using the gutted-shell of an old aftermarket head unit for an "OEM"-looking Mazda3 head unit, some of the extra storage options previously available by the old head unit needed to be refabricated and remapped to the new computer. (2x SD expansion slots and 1x front USB input) The easiest way I could make this happen was through USB, so I bought 2x USB card readers, and custom made a female USB connector in the front of the unit, and then plugged them into a 4-port USB hub. I also had a joycon EXR controller plugged in the 4-port splitter as well because of some buttons that also needed to be mapped to the computer from the aftermarket shell. Then I can just fabricate another female USB port on the back of the display for the 4-port splitter input, and use a standard male-to-male USB cable to connect the display/4-port splitter to the computer.
Okay so here's the problem. Intel, the very people responsible for the birth of universal serial bus communication, cannot get the software responsible for enumerating the ports properly on their own motherboard during boot sequence to work properly. Having done some in depth research, the dreaded "EB" post code seems to be a common problem amongst more than just Intel's Mini-ITX motherboards, but rather effects a lot of other boards as well, and the code seems to get stuck and be most finicky with the removal and insertion of certain USB peripherals as others seem to figure out.
It's interesting to note that this year Intel has ALREADY released yet another BIOS update for the DN2800MT motherboard. In fact, as long as I have been working with computers, I have never seen so many BIOS updates for one brand of motherboard well after a year since it came out. The Intel website currently lists 13 since the first on November 30th, 2011.
Bottom line, don't buy an Intel motherboard unless you plan on using minimal USB devices with it, or it will annoying the heck out of you. It's a great board, but wow, not what I expected from Intel.