Another Failure to hibernate
For anyone interested, I spent some time chasing another failure to hibernate and/or resume under certain conditions.
Once every few days I would still get a boot failure and/or hibernate failure which also makes the ATX “emergency power off timer” drop the rails if the PC hangs during shutdown.
This type of problem is a real pain to find and initially I started looking at software and drivers.
I found that the “Ultimate DAB-FM” radio module was “now” causing a failure to suspend; more to the point, if the com Port was open when the PC went into hibernate.
I uninstalled the USB driver and reinstalled, same condition. I had recently installed numerous updates to the Win7 64 OS so I put it down to that and coded the FE to close the port prior to hibernate and on resume coded the FE to reload the Radio (if it had been playing at hibernate). This solved that problem.
Now to the “every few days” random boot failure (resume from hibernate) and/or failure to shutdown. The fault only happened in the vehicle, not on the bench, so I went looking for a power glitch (again).
USB devices can cause a boot or shutdown failure if they are plugged in or unplugged during the start-up or shutdown process, any disturbance or glitch on one of the dozen USB devices can look like a sudden disconnect to the OS.
My system has 4 separate custom microprocessor devices that communicate with various vehicle systems and each is interfaced via an RS232 to USB device. These have device drivers and I know from using them in various projects they will hang a system if not powered correctly- they can phantom power a connected device through the RS-232 data lines and will often hang if forced to do so. To overcome this problem during times when the supply to the micro regulators may be off, I secondary-power the Micro via an isolating diode from the +5 USB supply available on the USB-232 convertors.
I decided to rewire the system (again – talk about a learning curve) and in the process I wired the system to power the 4 Microprocessor devices from the PC-PSU +12v ATX line as well as +12v via the ACC/IGN switch (car battery).
It means that during cranking, the USB +5v line is not being called upon to suddenly supply extra current to the Microprocessors when the ACC voltage disconnects. These devices run perfectly from the USB supply alone and although drawing low current I wondered if the sudden (slight) transition from regulated +5v (via +12 Acc) to +4.6v USB power was causing a problem. I also got around to running ATX “+5v and common” to the front of the car via two 25A cables. This is to power 3 front USB hubs. I use 25A cable for no voltage drop and good common supply grounding.
I have not had a single boot/hibernate failure over the past week, even during all manner of abusive cranking and battery conditions since rewiring the system and reprogramming .The FE stops 2 WEB cameras (if they are recording) and now closes the DAB-FM port on hibernate or shutdown, it works perfectly.
http://www.brisdance.com/Honda/files...onnections.png
New Layout - part of the Car-PC wiring.
PDF format
As a side note, I have 8 software graphic meters programmed into the FE and watching these (for the first time believe it or not) revealed something I was not aware of or had even thought about.
This vehicle adjusts the Alternator voltage to meet the demands of current consumption in the system by monitoring current flow across a sensor “and” by interfacing with the inbuilt CPU. It knows when various lights are turning on/off and when the AC compressor kicks in/out etc.
The alternator output is 14.8 volts after cranking, after 5 to 10 minutes this drops and soon settles on around 13.4 volts. Turning the Climate control or lights on will instantly raise the output. Now this can be a problem for an AUX battery in the trunk. The system does not appear see 4 to 6 amperes drawn by the AUX battery and PC, and will drop the alternator output to 13.4v as soon as the Main battery reduces charging current sufficiently and the vehicle lights and AC are off.
With around 200mv drop across the cable to the Trunk (Car-PC “and” aux Battery both drawing current) you won’t charge the Aux battery on short trips. At the moment I make sure I have the climate control on to get over 14v to the aux battery for a period of time.
More thought needed! - If I knew what I know now this would have taken a lot less time, headaches and problem solving. Isn’t that always the case!
Smart charging systems are designed for a car in factory state, no allowance for small Car-PC’s or AUX batteries. This is a high current charging system but it won’t charge what it can’t see.
Rule1. If you are going to mount “even” a small AUX battery in the trunk and run a PC that only draws 3 to 4 amperes, run 55A cable.
Rule2. Do as OldSpark say’s on this forum! For the lowest voltage drop to any AUX battery use a relay isolator, or if an isolating Diode is needed, use relay switching across the diode.
The size of the cable is not only about current capacity, sometimes it’s all about a "very small" voltage drop, especially over a 15 foot run and even to a small boot mounted AUX battery.
The biggest problem with a complex system with a lot of vehicle interface and hardware devices spread around the vehicle is thoroughly testing it. I can’t sit there for hours trying every possible cranking scenario just to find the one that causes a problem, (batteries, starter motors and vehicles are expensive) so it’s down to the “solve it as it crops up” method. However over the past week the PC has resumed around 200 times during testing, dozens of firmware upgrades and FE code changes and run through around 30 or so engine cranks without a failure to resume.
Another bit of tech info - voltage drop.
I decided to test individual components in the supply path between the Vehicle battery and my new 26AH PC battery that’s mounted in the trunk.
I measured the voltage drop across components and leads, firstly on the bench and then in the vehicle once any needed mods had been made.
The bench setup is simple and consists of a 100 ampere adjustable linear power supply, digital multimeter and a 50 watt 10 ampere @ 13v resistive load.
The voltage drop across 15 foot of 56A lead and a 60A rated 5AG fuse holder with a 30A fuse running from the vehicle battery to the power input terminal (100A rated) on the PC case was around 4mv @ 10A, which is almost nothing.
The voltage drop from the input terminal to the electronic low voltage main battery protection circuit via a 20A fuse was a staggering 280mV.
Testing across the older design 25A rated fuse holder showed it to be the culprit. I tried a new 3AG 20A inline unit from Jaycar and the voltage drop went down to 18mV using the same 20A fuse. I may look at fitting a 5AG inline unit down the track but they are just a bit large for the available position.
Next was to measure the drop across my FET low voltage isolator and this turned out to be around 40mv. Now the hard part with these 80A FETS is getting a big lead to a tiny connection on the FET. I have done this and I consider this drop to be acceptable, especially as the “FET switch” draws virtually NO current and this is important as it’s always powered.
That 40mv has to taken in light of the voltage drop across the Directional charge isolator. This was initially a FET across a 35A Diode (the diode is used for a specific reason in this installation) but I replaced the FET with a 30A relay. This relay shorts the Diode whenever the alternator is running (voltage above 13V.) The relay is also powered via the 12v ACC line and this voltage drops to zero when the car is cranking, hence “instant disconnect” and no drain from the PC- battery, but main battery voltage is still fed to the PC battery via the diode in the event of the PC-battery being partly discharged, this stops the PC from crashing if it’s started booting.
The voltage drop across this 30A relay is 50mV at 10A. I tried a few and they all read about the same, these are new units from Jaycar. So for smaller relays in lower current installation below 30A the FET I used is not bad.
The lead from the isolator is 56A and runs to the battery via another 5AG 20A fuse and 60A holder and there is virtually no voltage drop here.
So from an initial 374mV (4mv + 280mv + 40mv +50mv) I’m down to 104mV with the new fuse holder. That extra 270mV may not sound like much but makes a big difference to a charging battery.
With 14.3 volts across the vehicle battery I’m now reading around 14.1v across the PC-Aux battery @ 10A. The alternator output can go as high as 14.8 in this vehicle
It pays to take the time and double check everything when you’re trying to make an installation as robust as can be.
Bit more info on some relay tests.
BTW Thanks koltron5543 for the nice feedback.
Tested:
Jaycar - 30A SPST with inbuilt 30A spade fuse. Voltage drop = 148mV @ 8A. Drop mainly due to inbuilt 30A fuse.
Jaycar - Mini 30A SPDT no fuse – from 40mV to 75mV @ 8A.
Jaycar - 30A SPDT (PCB mount version of the fused 30A relay above but no fuse.) VD = 8mV @ 8A.
Guess I’ll be changing a relay.
The battery I’m using is a 26AH deep cycle sealed unit that needs a quoted 14.4 volts min for charging. Initial current draw is quoted at 7A and so far it has not gone above that value. The PC draws around 3.5A to 4A so worst case total so far from the charging system has been 11A. PC standby current in sleep is around 380mA and battery 20hour rate is quoted at 1.3A.
You normally need at least 3 fuses and one relay between the Charging system and the turnk mounted AUX batt and PC. Those small voltage drops across these and the cable add up to effect the aux battery’s ability to charge correctly and hold standby.
Basically, if you want a low loss feed for a smallish aux battery and PC system (5A to 15A draw), then you really need to use 56A cable runs, 5AG fuse holders and “larger sized” 30A relays “without” inbuilt fuses.
Last of the voltage drop mods.
I decided to mod the last few items causing a voltage drop from the Main battery to the Aux battery.
Here is the revised circuit.
http://www.brisdance.com/Honda/files...nnections2.png
New Layout - part of the Car-PC wiring.
PDF format
I was trying to find a way to use a relay across the Low-voltage protection circuit FET. The problem is I want this circuit to be on at all times unless the main battery is getting low.
The FET draws NO current when everything is off, a good 30A relay draws anywhere from 70ma to 130ma. This is not going to cut it here.
There are ways to make a relay use less current and stay energised by taking advantage of the fact that the hold-on voltage is far less than the operate voltage. In the relays I tried I could get it down to 3.5 volts at 20mA. However making that work was not going to be reliable in this situation unless I went to some trouble.
I had a sudden realisation that if I left the FET in circuit for when the car was off (no current wasted) and placed a relay (RLY2) across the FET, driven from the charge transistor, I would have the best of both worlds. No standby current when the vehicle is off and no voltage drop when charging.
The FET and D1 allow the Main battery to help support the Aux battery in sleep and in starting the PC if the aux charge is low. Although diode D1 allows Main battery voltage to flow to the Aux Battery, it will not allow the Aux Battery voltage to flow back to the Main battery during cranking (or any other time really.) It is only shorted by RLY1 when the charging system is producing over 13v. Under this condition both RLY1 and RLY2 are on and there is now a direct low loss path between the Aux battery and the charging system.
I also replaced the input fuse going to the Low voltage protect circuit with a 5AG unit.
The voltage drop between the Main battery and the Aux in the trunk when heavy charging with all mods finished is now around 100mV. Most of the time it's around 40mV.
I also added a Power Off/Reset relay and switch to the +12 input of the M4 to allow changed settings to be updated. It was getting to be a pain disconnecting two fuses every time I made a change.
