That's resembling one of my issues where I even desoldered the 8042 and one logic chip in the reset circuitry, just to find out the board is OK and just does not work with a piezo buzzer! It simply resets itself on the first beep. Once a normal speaker is in place, or no speaker at all it works!! This particular board btw also does not work with a CF card adapter but works with any normal HDD or IDE industrial flash module! I have added all those information in to TheRetroWeb so the other guys can save a lot of time and frustration. The board in question is LS-P54CE
F.Y.I. Acer V58 was used on Fujitsu MV205, MVII265, MVII267 Micro Tower sized system. The characteristic is that in order to keep 6 available slots, peripheral connectors are case or bracket mounted.
You are quite welcomed Tony! And nice job with the headers :) Sorry i didn't had the time to design a pcb for it ( the lats two weeks were crazy busy for me ), but anyway if any pcb is drawn then it should be on the updated version i showed you after the fix and it would have delayed you. I think i will do it as soon as i can, and maybe in two versions, SMD and THT and post here in comms a link to all it's files ( tests and gerbers included ). Cheers! Marian.
Thank YOU! I'll leave any revisions of the PCB to the owner of the board, I'm happy to move on to other projects :) When you have the design let me know and I can share them in the description (mentioning you of course). And please don't feel you have to make them!
Wow - this is very weird behavior. I wonder if there is something on this board that can be identified to cause this requirement for delayed switch-on - and if it could be disabled. This seems to be a big inconvenience. One could just waste hours to find nothing wrong with this board and then ... Oh, wait. You did! :) Thanks for sharing this info. Great video!
@@tony359 I'd bet there's something wrong with the reset circuit on this board! It could be next to the ISA-slot where the green bodge-wire is located. Maybe some voltages need time to come up and there are undefined states.
It could be but being an Acer board, it's likely it belonged to a custom system which featured a custom start-up sequence. It was common at some point before ATX came to life. Thanks for your comment!
@@tony359 Yes, this may be the case. I only once encountered a mainboard which i wasn't able to get starting up. It's some custom intel dual xeon p3 board. I'd at least trace the reset pins and startup stuff for fun, maybe theres something :)
If you want to simplify your power on circuit I think I can help. IF you put a large value capacitor in series with the energizing coil of the relay with a high value resistor in parallel with the capacitor. Since DC doesn't pass through the capacitor it will deenergize the coil after a short period of time because the capacitor reaches a fully charged state. This way the coil doesn't have to be in an "always on" state for the computer to run.
Near the end of the "Jumper Manual" for the V58 board, it mentions "Power Supply Selection". It has either Resume, or Traditional settings. Resume = JP1 pins 3+5, and pins 4+6 closed. JP15 pins 2+3 closed. Traditional = JP1 pins 1+3, and pins 2+4 closed. JP15 pins 1+2 closed. I would guess "resume" means it should power up as soon as you apply power / Ground the Green with on your ATX adapter. And "traditional" means it will wait for the power button signal. Could be worth a try?
That was shown on the first video: on "RESUME" it never exits the RESET state. There is a RESUME header on the board: again, designed to work with some special external HW in my view. Thanks for your comment!
@@tony359 Ahh, no problem. I don't think I'd caught that video yet. lol My only other thought was that maybe there was a TPM header or something that has to "request" a full power-up, but I think the board is too early/old for that.
I gotta tell you, I have rescued some pretty high end rack mounted 'dead' PC servers from the EWASTE pile at work and all they needed was a new CR2032 battery. Ditto that for 'bad' UPS units that just needed new batteries. And these two UPS units were 3kw high end double conversion sine wave units.
I wonder if just a capacitor across the power button pins would do it. At power-on it would look like a short as it starts charging up through whatever pullup resistor is in the logic, once it charges up enough to go over some voltage threshold (around 1 or 2 volts) it should look like an open switch.
Someone have mentioned that too, it might have worked. I guess one question would be: how long would it take for the capacitor to discharge so one could power cycle the board? :) Thanks for your comment!
Great work Tony and by the community here as well! Back in the socket 370 days I had a board that had a similarly strange but different power on problem that caused a no boot condition. I had built a plain desktop PC for my now ex-wife's parents who lived on a farm. Everything was just fine for several months but after a bad thunderstorm the modem stopped working. Seems like the PCI modem may experienced some excess voltage via the phone line during a storm. After I removed & replaced the modem the machine would not display any video or POST from being off unless I first pressed the reset button on the case after powering it on. Prior to the modem replacement, even after the modem stopped working, the machine would boot just fine every time! I tried re-seating all the components, flashing the BIOS and even stuck in a different video card and RAM with no change. The really weird thing was I was messing around in the BIOS, trying different system settings to see if any of them would change the no POST w/o a reset on a cold boot behavior and I found one that did! The machine was equipped with a Celeron 600 CPU and the mainboard was a jumperless one. If I changed the FSB to almost anything ABOVE the default 66 MHz the Celeron 600 was designed to run at the machine would POST every time. After a little testing I found that the machine appeared to be 100% stable at a 75 MHz FSB and would happily run at 675 MHz. So I gave the machine back with a new modem and told my ex in-laws that I sped their computer up a little while I was at it. I never did find out what happened to that machine in the long run as I divorced my 1st wife a couple of years after that and didn't see much of those in-laws as they lived a couple of hours drive away.
interesting, I wonder what happened there! One possibility is that this V58 is actually faulty and my circuit is just a workaround. Who knows... Thanks for watching!
Awesome you found out what it was! CF can have problems sometimes, sometimes it’s weird incompatibilities between udma modes and the cables, sometimes it only works with a 40, other times only with 80 cables. You don’t have a cable which should avoid the problem but probably it’s another weird incompatibility of CF.
So, forgive the wall of text (and useless knowledge...) The power supplies in AcerPower computers of a similar vintage had a 3-pin "soft start" ATX style circuit- which ran to a three pin connector in addition to the standard P8 and P9. The connector (if memory serves) has a 5v constant, Power-on, and Ground. The "power bottom"🤣 connector was a momentary switch that, when pressed, would tell the board pull the Power-on to ground, turning the power supply on. Which would explain why the board needs to see that circuit momentarily grounded before it will initialize and POST.
Probably was in some industrial equipment, which has a power switch with a third contact that does this. You can also add on a single molex socket, and feed 12V via the 10R resistor to the HW monitor socket, and have the fan connector work as well, as the original HW monitor board likely did this as well.
My concern is: what if that pin is also going somewhere else on the motherboard? I'll leave this experiment to the owner of the board :) Thanks for watching!
i have an Acer V66LT slot 1 motherboard and it has a similar circuit... i have directly plugged in a floppy power connector to the header pins and it just works like normal. The floppy power connector sends 5v down one of the grounded pins and activates the circuit. Don't use the 12v line... it might do bad things.
Nice to see that finally the mystery solved Tony my congrats! I really do believe it was a difficult riddle cause this behavior was really strange and made the board completely unusable. I have a question thought, does this header also turns off the computer? A soft off before ATX, so you don't have to toggle the main physical switch but use an AT PSU?
One eye is laughing one is crying. Happy to find the solution for this particular problem. Unhappy because the power up sequence still be a misery for us seeing four lines on the post card so many times. It would worth to investigate where the signal pin is connected from that switch (jumper). Possibly to a gate input and the other gate input got an RC timed signal as well. When the two signals are present than the gate output changed. What would be great is to know where the gate output is connected, to which pin of the chipset. I understand you spent countless time on this MB but if there is a canche to get it an another 30 minutes than maybe this direction would be worth to investigate to fully solve this "no start" problem.
I see what you mean and it would be great to reverse engineer the board but I'm happy with this solution which should be reliable and work 100% of the time. In the end, this is not a fault, it's how the board was designed to work - ok, I'm supposing that - so we're getting into a different realm, it's not "broken" anymore :) But I understand your curiosity of course! Thanks for watching!
Maybe the problem with that motherboard is due to a capacitor. Switching on, a current peak occurs, and the capacitors stabilize the power line. But if one fails, the current peak produces an error in the program and crashes. When we introduce a delay, the motherboard starts with the voltage stabilized. It's a possibility because problems due to electrolytic capacitors are quite common. Finally, the solution of adding a transistor was perfect, but there are other ways, such as using an integrated NE555 type. I had fun with the repair, your videos are quite good.
It's possible but it was common back then to have manufacturers come up with their own "power supply standard" .This board has too many "unknown" power connectors which lead me to believe this behaviour was intended. But who knows! And of course, plenty of other solutions for the timer - from a simple capacitor to the 555 as you say! Thanks for watching!
I had a computer doing something like that too. To power it on I had to power it on by either the outlet or the switch on the power supply, wait a few seconds, turn it off, wait a few seconds, turn it on again... and it would power itself on, even when set to not power on after power outage. Power button sometimes worked, most times it didn't do anything.
Hi Tony, I just watched your video and have a suggestion you might want to try out. According to the board, JP1 and JP15 control the power supply configuration between 'Traditional' and 'Resume' mode. After some investigation, it appears the configurations are as follows: Traditional (current setup on your board): JP1: 1 & 3 closed, 2 & 4 closed JP15: 1 & 2 closed Resume: JP1: 3 & 5 closed, 4 & 8 closed JP15: 2 & 3 closed I suggest changing JP1 and JP15 to the 'Resume' configuration and attempting to power on the PSU without your custom circuit. I believe this setting might be this board's version of the 'Turn on after AC power loss' function found in the BIOS of modern systems. By changing these jumper settings, you may be able to power the system on and off without needing your custom circuit, by just using the switch from the PSU. Best regards, Michiel
Hi Michiel Thanks - that was explored in the previous episode. With the jumper set to "resume" the board never comes out of Reset! What I never tried was whether the board would turn on with the power button normally when in "resume mode". But it would still be inconvenient: the AT power supply would keep the board in a "reset" state all the time.
Could be an early case intrusion implementation. Some kind of switch setup that would prevent the system booting when the case was off? I would have to guess it was a workstation board with anti tamper circuitry. I wouldn't be surprised if there is a jumper somewhere on the board to enable/disable it, possibly a bridge if they were particularly stubborn about people bypassing it. An interesting board. I never encountered anything like this on consumer boards in the 90's.
But the system would not power on with that switch in either state, it would need to "go from high to low" within 4 seconds so it wasn't a switch for sure. There must be a way to disable it but I couldn't find anything... I've found online pictures of the V58 installed in Fujutsu machines and those lack the "hardware monitor" connector altogether - it's unpopulated. So yes, chances are there's a link or a jumper or something to change that behaviour!
Wow i forgot about the reset vector. Nice. Interesting board and issue. It being likely a server board, they probably made it to a higher quality standard than the usual you see of them.
Very nice! This reminds me of an issue I had on my first PC build: Shuttle SFF SN45G (NF2 chipset & Athlon XP) with Radeon 9800 Pro All-In-Wonder graphics. The system would not boot normally, and in researching the issue I found that a delayed plugin of the video card extra power connector resolved it. So I built a little relay + RC delay circuit to fix it. :-) That board finally died a few years ago, but I still have the Molex delay circuit harness I built for it.... just in case. ;-) I also built a little circuit for it to manually adjust the memory voltage (for overclocking, due to a BIOS issue)... but that's another story!
@@tony359 Sorry I wasn't clear :-) It was a 9800 Pro AiW AGP card, with a 4-pin molex for extra power. If it wasn't plugged in, the system would halt with a message that it needed plugged in. I used that card in other systems back then, and had no issue; only the SN45G needed the delay circuit mod on the AGP card's power connector. ;-)
So you made a time delay relay to get rid of the Power ON issue? That was something unexpected! Need to remember what you did with the power button just in case I bump into a motherboard that doesn't want to power on. Thanks a lot!
That seems odd. I was thinking maybe the header could be for one of those chassis switches that detect whether the chassis is open, but if there really has to be a delay, I guess it's not.
the acer boards a usually not ordinary ones. and also undocumented because they were sold in systems. great you've found the right starting sequence (shot in the dark, ain't it?). as far as I remember those times the producers gone mad on power saving features. I believe when you hold "power bottom" switch for 4 seconds it comes into a sleep state. but anyway you've sorted it out. one thing I don't like in that control board is that it's getting power from floppy connector. it should be much better to source 12 or 5 volts directly from mainboard.
I didn't consider the sleep state. It would be cool to try to trigger that state WHILE MSDOS is running and see what happens. That said, I don't think I was able to get the board out of that state once it's triggered sooo... Maybe the "resume" connector but that's another rabbit hole! I thought about getting power from the motherboard then I thought a Molex to FDD splitter is like £2... But yes, I could have run a wire from the 5V line from the ISA slots as well.
may be you'll find the supply voltage on pin connector nearby. for example, that one with a blue jumper. some tracing is needed to check if that's a power but not a control line.
So at its most basic level, you could say it wasn't working because you wasn't switching it on, lol. Well done figuring out the timing of that weird switching set up, my overkill unnecessary solution would have involved a 555 timer ;) Great video as always and good to see you won.
"did you try switching it off and on again" always works! Tons of alternative and better solutions for that circuit, I wanted something very simple for this occasion!
Hey thank you so much, Marco. The www was missing, I'm sure I copied that link from the "share" function of the website but thanks a lot, I've updated the link! Edit: no, it was me LOL! I hadn't completed the verification! Do'h! It should work now, with or without "www" :)
Great vid, When you have to make boards like that you know its time to start learning PCB design ;) Trust me when I say there is nothing better than receiving a board you have put together in design software. Top tip, if and when you learn, before sending it off to manufacture, check it and check it again. you'll make many silly mistakes and checking thoroughly is key.
I'll check and check AND I'll receive boards with silly mistakes on them, 100% :) That said, the red boards I showed on video (you might have skipped the advert!) were working 100% from a wiring perspective. So impressed! :)
ciao, complimenti per i tuoi video, un mio amico mi ha presentato il tuo canale. giacchè che ci sono volevo sapere siccome ho la stessa scheda di diagnostica, se mi può mandare per favore un documento per capire come funziona? grazie
Ehi ciao e welcome! I tuoi video dove ripari roba "dalla munnizza" sono gagliardi! Certo, ho un piccolo manuale cartaceo che trovai con la schedina che lista i "post codes" comuni per i vari BIOS. Ma in genere e' roba che puoi trovare su internet - i codici dipendono dal produttore del BIOS (AMI o AWARD). Sul display c'e' posto per due codici e via via che la scheda si avvia i codici scorrono da sinistra a destra. Questo significa che se il processo di avvio si ferma, il codice a sinistra e' l'ultimo e quello a destra e' il penultimo. L'idea e' che puoi consultare la lista di codici e "provare a capire" quale sia il problema. I LED danno indicazione visiva per lo stato della linea RESET e le tensioni su scheda. Se mi mandi una email ti mando la scansione del manuale - l'indirizzo e' nascosto da qualche parte sul mio profilo, YT continua a cambiarlo di posto... Ciao e a presto - e ringrazia il tuo amico!
i wonder if power good signal on atx connector is active high or low because it would flip instantly after power suply stabilizing all voltages so feeding it to that connector with or without some basic logic circuit might also work
I think it's active high but I'm not sure it's going to work - I have a feeling the "power bottom" needs to be pulsed AFTER reset has disengaged. But yes, many other ways to make this work I suppose. Thank you!
Thanks for these videos, Tony. I had a similar experience with an Intel RC440BX in which it did not work because the BIOS Jumper did not work well and went into recovery mode all the time. The problem I have now is that once it works, an unlock keyboard message appears and that's why the keyboard doesn't work, so I can't do anything about it. Can you think of how to unlock the keyboard? I have tried a lot of things without success
Interesting - I've been troubleshooting an older PC that wouldn't start following a power cut earlier in the week. I found exactly what you found: that holding the power button while turning on the PSU made it boot (at least sometimes). It also seems to be sensitive to the total load, and wouldn't start up until I disconnected the hard drives. With everything connected and the power button released, it would power on very briefly and then off (which I assumed was a PSU protection of some sort). I still suspect the PSU and want to do some more testing on that, though it passes the simple "paper clip" test.
When it did start up, it also reported that the CMOS data was corrupt, which is not surprising as the CMOS battery was down to about 2.5 V. I've replaced that now, of course.
just out of curiosity, did you try shorting the pins 19-20 on CN19 ? in the manual the connector you were using CN18 is labeled power button, and CN19 pins 19-20 are labeled power switch.
Welcome to the wonderful world of OEM proprietary mannerisms. I have some Siemens Nixdorf machines that have some extra cables and connector pigtails coming out of the PSU ang going to mystery headers on the board. No idea yet what they do, what the pinouts mean or whatever. Just am afraid that these PSU's might stop working and I have no idea how to replace them. Suppose that motherboard also had one of these proprietary PSUs where you just pressed a single button and the signal went to 3 different places. I feel those were the early days, things like proprietary power connectors and all kinds of other shenanigans seem to have really peaked in the late 2010s.
For sure! Those manufacturers were probably trying to be cool and also make sure that everything was proprietary. I can think of Italian Olivetti, the M series had the 16bit slots with the "16 bit part" further away from the 8 bit one. So the cards HAD to be sourced from Olivetti despite being normal 16bit cards. So sad.
@@tony359 Oh dear. I have several Olivettis waiting for resurrection and I didn't even notice that yet. At least they have the battery upside down, so the acid only ruined the case floor. :D
@@catriona_drummond It might not be the case for ALL Olivetti but some models definitely have those loooong card because the ISA socket is stretched away! :)
As long as unique magnetics of the PSU are fine, which normally you don't expect anything to happen to them, you can just rebuild the PSU with new capacitors and semiconductors, if need be, to like-new condition. Keeping machines mostly-original is a good goal anyway.
Some Mudder Err=motherboards come with an "open case" feature, so I am wondering whether that momentary switch your switching is actually the two pins to facilitate that.. APU 👁👍
Wonder if the bios has an option for power switch behavior i had a few old boards where you can set delays on the power switch and also loss of power conditions like if power interupt to auto turn on or stay off.
Good idea, I didn't consider that - though it would be weird that to power on a board you need to change something in the bios! :) But I wouldn't be too surprised! :)
This would be an interesting reverse engineering project. Doubt the north or southbridge are involved and this might just go through some logic chips or to the bios, shouldn't be to complicated. But would take alot of time thats not worth the effort considering its working.
i knew what the problem was, Tony, but i didn't wanna tell you :)) (i kidd, of course). an interesting repair, nonetheless, love this repair videos of yours
ahah yes, it would have been the perfect advert! They can deliver a PCB in a few days but my designing skills are still basic so... another project maybe :)
look at what I found today. So I have an Acer power supply, with AT connectors... but no power button, not cut, no nothing. I was how the heck to use this? So if I understand correctly, that power supply is for these motherboards!! Very interesting
ahhh amazing! Does it have any other cable sticking out of it? Can you send me some pictures? My email is on my YT main page, hidden somewhere by YT :)
Absolutely, we wanted to keep it simple but there are many other more elegant solutions of course! DJMarik made another version with a trimmer so the delay could be configured.
What do you think that blue jumper is doing in the middle of that four pin header right next to your "delay" board. The pin closest to the edge of the board connects to the fan +ve line. Is it possible that moving the blue jumper to the pins closest to the edge will provide the fan with power, and in some remarkable way do what you've just done with your "delay" board? Does the second pin from the edge have voltage on it when the board starts up? #justasking
If I remove it, an ALARM green LED is lit. I did try moving the jumper around (including removing it) but it would still not power up when power is applied.
10:30 When you have a motherboard with a header labeled POWER BUTTON (POWER BOTTOM) you connect a flipping button to it! RTFSS! SS is silk screen. If had this board, connecting a button to that header is the first thing i would do. You might not even need a relay. try connecting a capacitor across those pins, say a 470uF?
Nice one :) Glad you kept investigating and now we all learned something new. Thumbs up!
Thank you for the FFFFFFF0 :)
That's resembling one of my issues where I even desoldered the 8042 and one logic chip in the reset circuitry, just to find out the board is OK and just does not work with a piezo buzzer! It simply resets itself on the first beep. Once a normal speaker is in place, or no speaker at all it works!! This particular board btw also does not work with a CF card adapter but works with any normal HDD or IDE industrial flash module! I have added all those information in to TheRetroWeb so the other guys can save a lot of time and frustration.
The board in question is LS-P54CE
that's an interesting discovery, thanks for sharing it on TRW!
F.Y.I. Acer V58 was used on Fujitsu MV205, MVII265, MVII267 Micro Tower sized system.
The characteristic is that in order to keep 6 available slots, peripheral connectors are case or bracket mounted.
Very interesting, googling those models I found a picture of the V58 where the "health monitor" is NOT populated!
You are quite welcomed Tony!
And nice job with the headers :)
Sorry i didn't had the time to design a pcb for it ( the lats two weeks were crazy busy for me ), but anyway if any pcb is drawn then it should be on the updated version i showed you after the fix and it would have delayed you.
I think i will do it as soon as i can, and maybe in two versions, SMD and THT and post here in comms a link to all it's files ( tests and gerbers included ).
Cheers!
Marian.
Thank YOU! I'll leave any revisions of the PCB to the owner of the board, I'm happy to move on to other projects :) When you have the design let me know and I can share them in the description (mentioning you of course). And please don't feel you have to make them!
Congratulations!! . The world works when the community helps each other.
I hope this channel grows a lot more.
Thanks for your kind words! :)
Wow - this is very weird behavior. I wonder if there is something on this board that can be identified to cause this requirement for delayed switch-on - and if it could be disabled. This seems to be a big inconvenience. One could just waste hours to find nothing wrong with this board and then ... Oh, wait. You did! :) Thanks for sharing this info. Great video!
Thanks, it would have been sad to use it for parts!
@@tony359 I'd bet there's something wrong with the reset circuit on this board! It could be next to the ISA-slot where the green bodge-wire is located. Maybe some voltages need time to come up and there are undefined states.
It could be but being an Acer board, it's likely it belonged to a custom system which featured a custom start-up sequence. It was common at some point before ATX came to life. Thanks for your comment!
@@tony359 Yes, this may be the case. I only once encountered a mainboard which i wasn't able to get starting up.
It's some custom intel dual xeon p3 board. I'd at least trace the reset pins and startup stuff for fun, maybe theres something :)
I have to say, this is most interesting repair I've seen lately
Thank you, appreciated!
If you want to simplify your power on circuit I think I can help. IF you put a large value capacitor in series with the energizing coil of the relay with a high value resistor in parallel with the capacitor. Since DC doesn't pass through the capacitor it will deenergize the coil after a short period of time because the capacitor reaches a fully charged state. This way the coil doesn't have to be in an "always on" state for the computer to run.
I'll try that thanks!
Near the end of the "Jumper Manual" for the V58 board, it mentions "Power Supply Selection".
It has either Resume, or Traditional settings.
Resume = JP1 pins 3+5, and pins 4+6 closed. JP15 pins 2+3 closed.
Traditional = JP1 pins 1+3, and pins 2+4 closed. JP15 pins 1+2 closed.
I would guess "resume" means it should power up as soon as you apply power / Ground the Green with on your ATX adapter.
And "traditional" means it will wait for the power button signal.
Could be worth a try?
That was shown on the first video: on "RESUME" it never exits the RESET state. There is a RESUME header on the board: again, designed to work with some special external HW in my view. Thanks for your comment!
@@tony359 Ahh, no problem. I don't think I'd caught that video yet. lol
My only other thought was that maybe there was a TPM header or something that has to "request" a full power-up, but I think the board is too early/old for that.
@@electronash No worries, it's an old video. But it's cool, it's when I destroy the Shuttle 591P.....
I gotta tell you, I have rescued some pretty high end rack mounted 'dead' PC servers from the EWASTE pile at work and all they needed was a new CR2032 battery. Ditto that for 'bad' UPS units that just needed new batteries. And these two UPS units were 3kw high end double conversion sine wave units.
Yes, I know. The amount of equipment being thrown away for no reason is staggering.
I’m totally gobsmacked! I never thought you’d get to a solution. Well done and really appreciative of all your efforts
put the cap on the power-on pins.. 100uF the cap will look like a short on power on, but then open up from the 5v board logic.
Other have mentioned that, I'm sure there are a variety of options to make that work, thanks for the suggestion!
I wonder if just a capacitor across the power button pins would do it. At power-on it would look like a short as it starts charging up through whatever pullup resistor is in the logic, once it charges up enough to go over some voltage threshold (around 1 or 2 volts) it should look like an open switch.
Someone have mentioned that too, it might have worked. I guess one question would be: how long would it take for the capacitor to discharge so one could power cycle the board? :) Thanks for your comment!
@@tony359 was about to suggest the same. a high value resistor in parallel can discharge the capacitor again within seconds
Great work Tony and by the community here as well!
Back in the socket 370 days I had a board that had a similarly strange but different power on problem that caused a no boot condition. I had built a plain desktop PC for my now ex-wife's parents who lived on a farm. Everything was just fine for several months but after a bad thunderstorm the modem stopped working. Seems like the PCI modem may experienced some excess voltage via the phone line during a storm.
After I removed & replaced the modem the machine would not display any video or POST from being off unless I first pressed the reset button on the case after powering it on. Prior to the modem replacement, even after the modem stopped working, the machine would boot just fine every time! I tried re-seating all the components, flashing the BIOS and even stuck in a different video card and RAM with no change.
The really weird thing was I was messing around in the BIOS, trying different system settings to see if any of them would change the no POST w/o a reset on a cold boot behavior and I found one that did! The machine was equipped with a Celeron 600 CPU and the mainboard was a jumperless one. If I changed the FSB to almost anything ABOVE the default 66 MHz the Celeron 600 was designed to run at the machine would POST every time. After a little testing I found that the machine appeared to be 100% stable at a 75 MHz FSB and would happily run at 675 MHz. So I gave the machine back with a new modem and told my ex in-laws that I sped their computer up a little while I was at it.
I never did find out what happened to that machine in the long run as I divorced my 1st wife a couple of years after that and didn't see much of those in-laws as they lived a couple of hours drive away.
interesting, I wonder what happened there! One possibility is that this V58 is actually faulty and my circuit is just a workaround. Who knows... Thanks for watching!
Awesome you found out what it was! CF can have problems sometimes, sometimes it’s weird incompatibilities between udma modes and the cables, sometimes it only works with a 40, other times only with 80 cables. You don’t have a cable which should avoid the problem but probably it’s another weird incompatibility of CF.
Yes, I'm aware CFs can be weird and this board is weird so it's weird^2 and I'm happy that an HDD works :)
Is IDE to SD-Card adapter from China less finicky than CF? Or also same issue hit or miss compatibility especially in Pri controller?
@@SianaGearz I do not know, I use spinning HDDs now :)
Congrats on sorting out this one!
Thank you!
So, forgive the wall of text (and useless knowledge...) The power supplies in AcerPower computers of a similar vintage had a 3-pin "soft start" ATX style circuit- which ran to a three pin connector in addition to the standard P8 and P9. The connector (if memory serves) has a 5v constant, Power-on, and Ground. The "power bottom"🤣 connector was a momentary switch that, when pressed, would tell the board pull the Power-on to ground, turning the power supply on. Which would explain why the board needs to see that circuit momentarily grounded before it will initialize and POST.
not useless at all! Yes, I suspected a similar power supply style, it would be nice to have one to see how it behaves! Thanks for the wall of text! :)
15:06 This is the motherboard equivalent to Biff's car from back to the future. Only he knows how to start it up.
ahahahahaha!!!!! Amazing! Love that series of movies!
Probably was in some industrial equipment, which has a power switch with a third contact that does this. You can also add on a single molex socket, and feed 12V via the 10R resistor to the HW monitor socket, and have the fan connector work as well, as the original HW monitor board likely did this as well.
fancy meeting you here-- got [back?] into retro/vintage hardware about a year ago; needed to rekindle more positive childhood memories
My concern is: what if that pin is also going somewhere else on the motherboard? I'll leave this experiment to the owner of the board :)
Thanks for watching!
Great work getting this board back up and running!
Thanks!
i have an Acer V66LT slot 1 motherboard and it has a similar circuit... i have directly plugged in a floppy power connector to the header pins and it just works like normal. The floppy power connector sends 5v down one of the grounded pins and activates the circuit. Don't use the 12v line... it might do bad things.
that's an interesting - but risky! - solution indeed! :) Thanks for mentioning!
Nice to see that finally the mystery solved Tony my congrats! I really do believe it was a difficult riddle cause this behavior was really strange and made the board completely unusable.
I have a question thought, does this header also turns off the computer? A soft off before ATX, so you don't have to toggle the main physical switch but use an AT PSU?
Ah, good point, I haven't tried that! It'd be cool to find out for sure! Thank you!
One eye is laughing one is crying. Happy to find the solution for this particular problem. Unhappy because the power up sequence still be a misery for us seeing four lines on the post card so many times. It would worth to investigate where the signal pin is connected from that switch (jumper). Possibly to a gate input and the other gate input got an RC timed signal as well. When the two signals are present than the gate output changed. What would be great is to know where the gate output is connected, to which pin of the chipset. I understand you spent countless time on this MB but if there is a canche to get it an another 30 minutes than maybe this direction would be worth to investigate to fully solve this "no start" problem.
I see what you mean and it would be great to reverse engineer the board but I'm happy with this solution which should be reliable and work 100% of the time. In the end, this is not a fault, it's how the board was designed to work - ok, I'm supposing that - so we're getting into a different realm, it's not "broken" anymore :)
But I understand your curiosity of course! Thanks for watching!
The amount of commitment that you put for the work and videos is outstanding. Well done. Thanks.😊
Thank you for your kind words!
Maybe the problem with that motherboard is due to a capacitor. Switching on, a current peak occurs, and the capacitors stabilize the power line. But if one fails, the current peak produces an error in the program and crashes. When we introduce a delay, the motherboard starts with the voltage stabilized.
It's a possibility because problems due to electrolytic capacitors are quite common. Finally, the solution of adding a transistor was perfect, but there are other ways, such as using an integrated NE555 type. I had fun with the repair, your videos are quite good.
It's possible but it was common back then to have manufacturers come up with their own "power supply standard" .This board has too many "unknown" power connectors which lead me to believe this behaviour was intended. But who knows!
And of course, plenty of other solutions for the timer - from a simple capacitor to the 555 as you say! Thanks for watching!
@@tony359 It has happened even with the best motherboards, which is why the most modern ones use solid electrolytic capacitors
Amazing Toni! Thank you for sharing your contagious joy with us! I'm as happy as much as you for this success and we all have learned something !
Thank you so much!
I had a computer doing something like that too. To power it on I had to power it on by either the outlet or the switch on the power supply, wait a few seconds, turn it off, wait a few seconds, turn it on again... and it would power itself on, even when set to not power on after power outage. Power button sometimes worked, most times it didn't do anything.
Not a boring system at all :)
Hi Tony,
I just watched your video and have a suggestion you might want to try out.
According to the board, JP1 and JP15 control the power supply configuration between 'Traditional' and 'Resume' mode. After some investigation, it appears the configurations are as follows:
Traditional (current setup on your board):
JP1: 1 & 3 closed, 2 & 4 closed
JP15: 1 & 2 closed
Resume:
JP1: 3 & 5 closed, 4 & 8 closed
JP15: 2 & 3 closed
I suggest changing JP1 and JP15 to the 'Resume' configuration and attempting to power on the PSU without your custom circuit. I believe this setting might be this board's version of the 'Turn on after AC power loss' function found in the BIOS of modern systems. By changing these jumper settings, you may be able to power the system on and off without needing your custom circuit, by just using the switch from the PSU.
Best regards,
Michiel
Hi Michiel
Thanks - that was explored in the previous episode. With the jumper set to "resume" the board never comes out of Reset!
What I never tried was whether the board would turn on with the power button normally when in "resume mode". But it would still be inconvenient: the AT power supply would keep the board in a "reset" state all the time.
You are amazing!
ahah thanks!
Great job Tony! With a litle help from our friends.
Their help was critical!!
Nice to see you *tinker* up a solution for this odd power sequence!
Thanks!
Could be an early case intrusion implementation. Some kind of switch setup that would prevent the system booting when the case was off? I would have to guess it was a workstation board with anti tamper circuitry. I wouldn't be surprised if there is a jumper somewhere on the board to enable/disable it, possibly a bridge if they were particularly stubborn about people bypassing it. An interesting board. I never encountered anything like this on consumer boards in the 90's.
But the system would not power on with that switch in either state, it would need to "go from high to low" within 4 seconds so it wasn't a switch for sure. There must be a way to disable it but I couldn't find anything...
I've found online pictures of the V58 installed in Fujutsu machines and those lack the "hardware monitor" connector altogether - it's unpopulated. So yes, chances are there's a link or a jumper or something to change that behaviour!
Very impressive thinking outside the box! Well done and everybody learned a lot
Thank you!
Wow i forgot about the reset vector. Nice.
Interesting board and issue. It being likely a server board, they probably made it to a higher quality standard than the usual you see of them.
A very good point indeed!
Acer/IBM Aptiva; I'm familiar with the IBM, has AT style power connectors with an extra power connector (3 pin I think) for the 'soft power' button.
Very cool, one day I'll stumble into a working system so I can analyse it! Thanks for watching!
Very nice! This reminds me of an issue I had on my first PC build: Shuttle SFF SN45G (NF2 chipset & Athlon XP) with Radeon 9800 Pro All-In-Wonder graphics. The system would not boot normally, and in researching the issue I found that a delayed plugin of the video card extra power connector resolved it. So I built a little relay + RC delay circuit to fix it. :-) That board finally died a few years ago, but I still have the Molex delay circuit harness I built for it.... just in case. ;-) I also built a little circuit for it to manually adjust the memory voltage (for overclocking, due to a BIOS issue)... but that's another story!
A card with external power on an Athlon XP board? What model was that?
@@tony359 Sorry I wasn't clear :-) It was a 9800 Pro AiW AGP card, with a 4-pin molex for extra power. If it wasn't plugged in, the system would halt with a message that it needed plugged in. I used that card in other systems back then, and had no issue; only the SN45G needed the delay circuit mod on the AGP card's power connector. ;-)
@@jeremywh7 thanks for the details, very interesting indeed!
Thanks!
Thank you so much!
Congrats 🎉
Thanks!
So you made a time delay relay to get rid of the Power ON issue?
That was something unexpected!
Need to remember what you did with the power button just in case I bump into a motherboard that doesn't want to power on.
Thanks a lot!
I suspect there won't be many boards behaving like this one but definitely something to test on next "dead" motherboard! :) Thank you!
World Domination is a good point on ya list ... oh and besides that, a lot of interesting (sometimes simple) points to check.
You’re the first one to notice, well done!! 😂
That was really interesting. Nice finding!
Thank you, glad you enjoyed!
Thank you!
Thank you!
What a weird issue! At least it was figured out in the end, it's really interesting
Yes so nice to have that working, this board has been wandering in my workshop for months!
Awesome work and so satisfying to get an ending to the story.
It really is!
Somewhere out there is an original Acer machine with this board in it...with the original power supply. It would be interesting to see this.
Absolutely! I'd be so curious to see what's the intended use of all those extra connectors!
nice work! .. I was thinking waaaaay to complicated
The manual also says you need to close your eyes and count to 10 to power up the board :)
Great work 👌👏💯
i am impressed
Congratulations
Thanks a lot 😊
Hi Tony!
May be this MB use some proprietary PSU, with separately hardware monitor module?
Glad for resolving the problem with this MB.
For sure - it'd be nice to find out! Thank you!
That seems odd. I was thinking maybe the header could be for one of those chassis switches that detect whether the chassis is open, but if there really has to be a delay, I guess it's not.
the acer boards a usually not ordinary ones. and also undocumented because they were sold in systems.
great you've found the right starting sequence (shot in the dark, ain't it?). as far as I remember those times the producers gone mad on power saving features. I believe when you hold "power bottom" switch for 4 seconds it comes into a sleep state.
but anyway you've sorted it out. one thing I don't like in that control board is that it's getting power from floppy connector. it should be much better to source 12 or 5 volts directly from mainboard.
I didn't consider the sleep state. It would be cool to try to trigger that state WHILE MSDOS is running and see what happens. That said, I don't think I was able to get the board out of that state once it's triggered sooo... Maybe the "resume" connector but that's another rabbit hole!
I thought about getting power from the motherboard then I thought a Molex to FDD splitter is like £2... But yes, I could have run a wire from the 5V line from the ISA slots as well.
may be you'll find the supply voltage on pin connector nearby. for example, that one with a blue jumper. some tracing is needed to check if that's a power but not a control line.
So at its most basic level, you could say it wasn't working because you wasn't switching it on, lol.
Well done figuring out the timing of that weird switching set up, my overkill unnecessary solution would have involved a 555 timer ;)
Great video as always and good to see you won.
"did you try switching it off and on again" always works!
Tons of alternative and better solutions for that circuit, I wanted something very simple for this occasion!
Capacitor! Put a capacitor in place of that weird relay contraption.
I'll leave that to you and DJMarik78! :)
Nice job Tony. I have learned something more now. Interessting. Greetings from Steven from the Netherlands
Thanks Steven!
I enjoyed watching this episode 🙂
Thank you!!
Hey its Marco - your Coffee link leads to 404 error.... On the other side - awesome Video again!
Hey thank you so much, Marco. The www was missing, I'm sure I copied that link from the "share" function of the website but thanks a lot, I've updated the link!
Edit: no, it was me LOL! I hadn't completed the verification! Do'h! It should work now, with or without "www" :)
Great vid, When you have to make boards like that you know its time to start learning PCB design ;) Trust me when I say there is nothing better than receiving a board you have put together in design software. Top tip, if and when you learn, before sending it off to manufacture, check it and check it again. you'll make many silly mistakes and checking thoroughly is key.
I'll check and check AND I'll receive boards with silly mistakes on them, 100% :)
That said, the red boards I showed on video (you might have skipped the advert!) were working 100% from a wiring perspective. So impressed! :)
what about trying a momentary normally closed push button on the power bottom pins?
That won't work, I showed that in the video. The board enters a reset state after 4 seconds. Thanks though!
ciao, complimenti per i tuoi video, un mio amico mi ha presentato il tuo canale. giacchè che ci sono volevo sapere siccome ho la stessa scheda di diagnostica, se mi può mandare per favore un documento per capire come funziona? grazie
Ehi ciao e welcome! I tuoi video dove ripari roba "dalla munnizza" sono gagliardi! Certo, ho un piccolo manuale cartaceo che trovai con la schedina che lista i "post codes" comuni per i vari BIOS. Ma in genere e' roba che puoi trovare su internet - i codici dipendono dal produttore del BIOS (AMI o AWARD). Sul display c'e' posto per due codici e via via che la scheda si avvia i codici scorrono da sinistra a destra. Questo significa che se il processo di avvio si ferma, il codice a sinistra e' l'ultimo e quello a destra e' il penultimo. L'idea e' che puoi consultare la lista di codici e "provare a capire" quale sia il problema.
I LED danno indicazione visiva per lo stato della linea RESET e le tensioni su scheda. Se mi mandi una email ti mando la scansione del manuale - l'indirizzo e' nascosto da qualche parte sul mio profilo, YT continua a cambiarlo di posto...
Ciao e a presto - e ringrazia il tuo amico!
This is wild. Imagine that!
Great video!
Thanks!
A lot of socket 7 boards don't play well with compact flash cards.
Yes, I learnt on my skin that CF's are not always a good thing!
@@tony359Try a Promise Technology GP 0434-03 Sata 300 TX4 Controller Card
I'm good with spinning HDDs. I don't want to introduce further potential issues on old systems - see my "upgrade from hell" video of some time ago!!
i wonder if power good signal on atx connector is active high or low because it would flip instantly after power suply stabilizing all voltages
so feeding it to that connector with or without some basic logic circuit might also work
I think it's active high but I'm not sure it's going to work - I have a feeling the "power bottom" needs to be pulsed AFTER reset has disengaged. But yes, many other ways to make this work I suppose. Thank you!
Thanks for these videos, Tony. I had a similar experience with an Intel RC440BX in which it did not work because the BIOS Jumper did not work well and went into recovery mode all the time. The problem I have now is that once it works, an unlock keyboard message appears and that's why the keyboard doesn't work, so I can't do anything about it. Can you think of how to unlock the keyboard? I have tried a lot of things without success
hard to help from a distance I'm afraid, I'm not familiar with that board!
You could try power on with removed blue jumper from that white connector... of by changing its position.
Tried already, it doesn't work I'm afraid! That would have been too simple :)
@@tony359 did you tried to remove blue jumper and shorten pins which initiates powe on at the same time?
I might have, I've had this board for months, I spent quite some time testing! :)
Interesting - I've been troubleshooting an older PC that wouldn't start following a power cut earlier in the week. I found exactly what you found: that holding the power button while turning on the PSU made it boot (at least sometimes). It also seems to be sensitive to the total load, and wouldn't start up until I disconnected the hard drives. With everything connected and the power button released, it would power on very briefly and then off (which I assumed was a PSU protection of some sort). I still suspect the PSU and want to do some more testing on that, though it passes the simple "paper clip" test.
When it did start up, it also reported that the CMOS data was corrupt, which is not surprising as the CMOS battery was down to about 2.5 V. I've replaced that now, of course.
I'd definitely test another PSU, one possibility is that the PSU is happy with no load but won't work under load.
just out of curiosity, did you try shorting the pins 19-20 on CN19 ? in the manual the connector you were using CN18 is labeled power button, and CN19 pins 19-20 are labeled power switch.
The manual shows polarity on CN19 "power switch", that is the "power on" LED. Thanks for your input!
It is a case check botton? it was in a special case back then??
It could be. Unfortunately the manual doesn't say much...
Welcome to the wonderful world of OEM proprietary mannerisms.
I have some Siemens Nixdorf machines that have some extra cables and connector pigtails coming out of the PSU ang going to mystery headers on the board.
No idea yet what they do, what the pinouts mean or whatever. Just am afraid that these PSU's might stop working and I have no idea how to replace them.
Suppose that motherboard also had one of these proprietary PSUs where you just pressed a single button and the signal went to 3 different places.
I feel those were the early days, things like proprietary power connectors and all kinds of other shenanigans seem to have really peaked in the late 2010s.
For sure! Those manufacturers were probably trying to be cool and also make sure that everything was proprietary. I can think of Italian Olivetti, the M series had the 16bit slots with the "16 bit part" further away from the 8 bit one. So the cards HAD to be sourced from Olivetti despite being normal 16bit cards. So sad.
@@tony359 Oh dear. I have several Olivettis waiting for resurrection and I didn't even notice that yet.
At least they have the battery upside down, so the acid only ruined the case floor. :D
@@catriona_drummond It might not be the case for ALL Olivetti but some models definitely have those loooong card because the ISA socket is stretched away! :)
As long as unique magnetics of the PSU are fine, which normally you don't expect anything to happen to them, you can just rebuild the PSU with new capacitors and semiconductors, if need be, to like-new condition. Keeping machines mostly-original is a good goal anyway.
That was some journey... :) Bravo Tony, always an inspiration... but still cannot find the courage to try making videos myself...
It just takes a mobile and many many many hours :)
Thank you!
Some Mudder Err=motherboards come with an "open case" feature, so I am wondering
whether that momentary switch your switching is actually the two pins to facilitate
that..
APU 👁👍
It might be. If the manual was a bit more clear on the subject...
Wonder if the bios has an option for power switch behavior i had a few old boards where you can set delays on the power switch and also loss of power conditions like if power interupt to auto turn on or stay off.
Good idea, I didn't consider that - though it would be weird that to power on a board you need to change something in the bios! :) But I wouldn't be too surprised! :)
I am guessing that the case that board came in had a power switch that operated in that manner..
For sure, thanks for watching!
Great video.
Thanks!
almost sounds like the power good signal from the psu
indeed!
This would be an interesting reverse engineering project.
Doubt the north or southbridge are involved and this might just go through some logic chips or to the bios, shouldn't be to complicated.
But would take alot of time thats not worth the effort considering its working.
Indeed - and it's really not such a special board so happy for someone else to do that! Thank you for watching!
damn, it was so simple, you even got a response from necroware, you’re becoming famous ! Mi fa piacere ahahah
Grazie! :)
i knew what the problem was, Tony, but i didn't wanna tell you :)) (i kidd, of course). an interesting repair, nonetheless, love this repair videos of yours
ahah I'll check with you next time :) Thank you!
Haha how funny it is?! I would say RTFM but you said there is no such information?
The manual is there but there is no mention of what that "hardware monitor" is and some jumpers are just missing. I don't know! :)
Nice sometimes the solution is simple, need to recheck a 486 board that I do not get working.....
Good luck!
Necro Ware?
Zitten die in 'n dooie componenten niche of zo?
Take a look at his channel, it's super cool! @Necroware
Someone should make a pcb for that with pcbway ;)
ahah yes, it would have been the perfect advert! They can deliver a PCB in a few days but my designing skills are still basic so... another project maybe :)
You must use MOSFET against relay. But also there must be some primitive logic can be disabled
look at what I found today.
So I have an Acer power supply, with AT connectors... but no power button, not cut, no nothing. I was how the heck to use this?
So if I understand correctly, that power supply is for these motherboards!!
Very interesting
ahhh amazing! Does it have any other cable sticking out of it? Can you send me some pictures? My email is on my YT main page, hidden somewhere by YT :)
Hello I speak for the whole "the internet" and our reply is Hello to you
😅😅 Thank you Internet!
555 chip would be perfect here
Absolutely, we wanted to keep it simple but there are many other more elegant solutions of course! DJMarik made another version with a trimmer so the delay could be configured.
Maybe just small capacitor like 20-100uF instead of switch/relay contact would be enough.
Surely plenty of options :)
Well rigged!👍👍
Thanks!
What do you think that blue jumper is doing in the middle of that four pin header right next to your "delay" board. The pin closest to the edge of the board connects to the fan +ve line. Is it possible that moving the blue jumper to the pins closest to the edge will provide the fan with power, and in some remarkable way do what you've just done with your "delay" board? Does the second pin from the edge have voltage on it when the board starts up? #justasking
If I remove it, an ALARM green LED is lit. I did try moving the jumper around (including removing it) but it would still not power up when power is applied.
@@tony359 Oh well just a thought from an electronics ignoramus. I'm a firm believer in doing the simple things first, and it looks like you are too!
@@zs1dfr Of course, all inputs are always welcome!
You got a Power Bottom? 🤣
Apparently! 🙄
Umm yes older motherboards needed a proper switch for power not the same type of switch as modern atx "momentary" ?? Switch
AT motherboard didn't need a switch. The "switch" was on the 220V at the PSU. ATX have the soft-on button. This is neither. Thanks for watching!
Just add a 555! Their only .37
As I said, many options available. Mine is just one of them! Thanks for watching!
power bottom 🤣
🤣
10:30 When you have a motherboard with a header labeled POWER BUTTON (POWER BOTTOM) you connect a flipping button to it! RTFSS! SS is silk screen. If had this board, connecting a button to that header is the first thing i would do.
You might not even need a relay. try connecting a capacitor across those pins, say a 470uF?
What makes you think it was not the first thing I did? I even mentioned that in the video :)
Good Job
Thanks! And cool emoji!
Great video!
Thanks!