Installed the cap and diode... waiting for RMA'd motherboard to test
Installed the cap and diode... waiting for RMA'd motherboard to test
Hi, I have been away from the forum due to health issues and just had a quick read of your post. I originally posted that diode cap solution on the M4 IGN line and although this does solve some problems related to some installations of the M4 it does not solve all problems.
In fairness to the M4, I have recently found that the cycling problem while resuming from Suspend to Ram is actually a Motherboard design problem.
Firstly, the mother board will always work perfectly when resuming from Hibernate. The problem is only apparent when resuming from STR (suspend to ram) due to +5v standby rail.
In my case, the problem is intermittent and the unit can resume from STR in 2 seconds a dozen times throughout the day and work perfectly every time, then for no apparent reason it would cycle on/off trying to boot. As you noted, the only way to get it to boot was to remove power from the PSU for a few seconds.
From tests I recently carried out it appears that the problem is caused in two ways.
The initial failure to resume from STR is caused by the PSU thinking there is a problem with the 12v power input. It appears that most PSU’s will drop the +5 standby rail if they detect a problem, this is chicken and egg situation though as the MB will signal power good failure if the various supply rails don’t come up in a pre programmed time/sequence. This in turn drops the +5v Standby and in either case you have the MB failing resume and trying to reboot.
Changing my current MB for a newer MB from the same maker completely solved the rebooting cycle issue. Changing the PSU has no affect on this constant rebooting loop. It always follows the MB (BTW, latest Bios firmware)
I carried out further tests into the cause of the initial failure to return from STR. In other words, why does the +5v standby drop as soon as the PSU signals the MB to start?
The M4 was actually much better that some of the other 12v PSU units I tried.
In some cases putting a 40,000uf cap across the +12 supply input would stop the JAYCAR PSU/ MB combination from resuming from STR almost every time. It appears that in this case the Generic Switching Convertor controller does not like a very stiff +12 input and drops +5v standby causing the MB to fail on resume from STR. It should then do a cold reboot; however my original MB would lock in a power up cycle.
The newer MB correctly reboots after the first failure.
The M4 fails with the right combination of input voltage AND voltage change just as it powers on, and will drop the +5v standby rail every so often. Once again it may be a timing issue between the MB and the PSU causing the MB to signal Bad Power. The M4 allows you to change PSU rail timing for all supply rails, however I had no interest in going down that road.
I now use Hibernate, its fast enough for me and the CAR PC has been absolutely faultless in operation. The M4 performs perfectly and after rebuilding the complete system into a larger case and combining the PSU, MB, DAB radio and Remote volume control unit into the same large case all audio noise and RF noise has finally been eliminated (exorcized).
Hope this is of some help.
Mickz - bluddy brilliant as usual! (Great description and scope.)
Sorry to hear about your health issues and thanks so much for stopping by and breaking bread. You have definitly shed some light on all of the issues that i have been having. I have been pulling my hair out with these issues because they are intermittant and hard to reproduce at will. Like you STR will work properly 10 times in a row and then fail randomly.
As for the using hibernate. I would much rather use standby since I do have an isolator and Aux battery. I also use a Carpc Joycon EXR and a USB captre device for my rear view camera. For this i need a ~5 second resume time.
One thing i would like to mention that we do have in common (if you specs in your signature are up to date) is that we both use a core i3 LGA 1156 (although mine is the 3.2) and we both have motherboards with the H55 chipset.
Currently i have a Zotac motherboard (H55-ITX-C-E) http://www.newegg.com/Product/Produc...otac%20H55-ITX
I had a couple of follow up questions.
What other Power supplies did you test? Any Opus? In the past i have used an opus 150 and i never had any issues.
Is there any way to safely make the M4 fail (drop standby and or reboot cycle) at will? If i do decide to play with the timing of the rails this will help with testing and verification
If it is a timing issue and i could dial in the correct settings would that potentially resolve both issues (standby power cut and power on cycle)?
Is it safe to say that you never resolved the STR problem with your motherboard and that a workaround that was acceptable for you was to use hibernation?
Thanks again for your help on this
Hi antimatter. The MB I’m using now is a GA-H67N-USB3-B3.
I tested the M4, a unit from Jaycar which has a common controller and two older industrial units with earlier chipsets.
I kept the reply fairly simple but the interaction between the PSU and the MB is complex when the PSU decides, for whatever reason, to delay its start-up. Resume from STR requires the PSU, once started, to be ready almost instantly. Any glitch or hesitation and the MB signals power bad and start’s its boot failure routine.
Obviously this timing and boot routine varies between Motherboards and I think you may chase your tail forever trying to tailor the M4 timing to overcome this problem.
If you get a good 20A bench Power-supply and lower the input voltage somewhere between 11v and 13v you will find a point at which the PSU will not resume correctly. Sometimes rocking the supply rail voltage slightly around this point will cause a 60% fail rate for resume from STR.
Like I said this has no affect on any other mode – Just STR and instant resume by the MB.
Yes, I gave up on STR and now hibernate. Every so often (1 in 20 times) I would also have the Garmin-PC GPS receiver (USB) fail to initialise with STR. Hibernate is faultless.
I have completely revised the Car Supply system but I still use a small “part isolated” 12v battery for the CAR-PC. I have two isolators (my design for adjustable 60A solid state isolators). One is set to isolate the Car battery below a pre-set voltage which stops the Main battery bleeding down, the other bypasses an isolation diode to the PC battery when the input voltage is above 13volts – allowing the PC battery to fully charge.
I don’t trust the M4-PSU Main battery cut-off logic; others have found it to be – flaky.
I still use a PC battery to reduce the chance of a supply glitch/spike affecting USB device initialisation on resume when the car is cranked. I have a number of USB Microprocessor interface devices running throughout the car and with this system everything is once again faultless.
It seems to me that this is related to an incompatibility with the M4-ATX and the motherboard.
I have high hopes that moving to a higher quality power supply will solve these problems. I had good experience with Opus in the past. I actually have had two Opus 150's that worked flawlessly with other systems so maybe the Opus 320 would work out better.
For me the ultimate goal would be getting STR with instant resume to work near flawlessly.
What you suggest me to do? Change boards or change power supplies or both
going off the findings that mickz posted, i would say that your issues are related the psu more then the board. so i think the psu needs to go..
I would tend to agree. I think like mickz has said that there are issues with the board however things are exacerbated by the PSU
Hi to all, quite honestly, I don’t know if changing the PSU to any other supply would solve the occasional failure to resume from STR. The Switching supply and the environment it works under are so complex when it comes to the very fine and critical timing required by the MB to resume instantly.
Chicken and egg syndrome is all I can liken it to. Both the PSU and MB design from each manufacturer most likely meet the “GENERAL” timing and signalling requirements but simply won’t agree 100% of the time in this critical start-up mode.
If you select a MB that resumes from STR without going into a loop like my previous model and your current model then the occasional failure to resume will only cause a delay and a longer cold reboot.
I am going to switch to that for a trial when I feel up to doing some tinkering.
I read somewhere about Win7 dual suspend mode allowing a resume from hibernate if STR fails? Not sure if it works but someone might chime in.
I guess I feel stuck and am trying to decide what the best next step to do is. I don't want to change my entire platform (motherboard, memory, cpu) and drop 3-4 hundred dollars. I also don't want to spend another 200 dollars on the Opus 360.
One thing I did notice while going through the settings for the M4 is there was a delay for the 12v rail and the other rails were set to 0ms. I haven't changed it yet but maybe setting the 12v rail to a 0ms delay would change things... somehow I doubt it.
I have to think that if this board is finicky about having steady rails (which I have read @ the zotac forum) that having a higher quality power supply would be helpful.
I don't know... I am open to suggestions. I also posted a thread in the general hardware section asking if anyone with the H55 chipset and socket LGA1156 have flawless resume from standby and if so what PSU.