Ongoing saga of PC Boot Problem 4

[Crustyoldbloke]

OK, new thread without the 90 odd previous postings. Now where were we?

Jay, just tried disconnect on HDD with boot up disk. POST beep was good - same result though. I've rechecked all of the jumper settings.

Die dulci fruere!
Phil
24 Beers in a case, 24 hours in a day.......coincidence??

 

[Crustyoldbloke]

Jay

I think you may be right about trying a different CPU and possibly PSU. I am assured that these MOBOs are very reliable and the liklihood of having a bad one delivered is very remote.

Let me ponder a while to see if I can source either without taking my good PC apart.

Die dulci fruere!
Phil
24 Beers in a case, 24 hours in a day.......coincidence??

 

[GrandpaCarl]

HI PHIL,

My oh my! Thought you guys would have this figured out by now. I am somewhat apalled at the lack of good information on this boadr. The above reference almost tells you what you need to know:

The P5SS-Me+ clock frequency can be selected from a choice of frequencies ranging from 66Mhz to 133.3Mhz including: 66Mhz, 75Mhz, 83.3Mhz 95Mhz, 100Mhz, 112Mhz, 124Mhz and 133.3Mhz. The clock multipliers can be selected from a choice ranging from 5.5X to 5.5X in steps of .5X. The processor voltage core can be adjusted from 2.1 to 3.5 vdc.

The thing that hit me the hardest is that 66MHz is the lowest frequency for which the clock can be set - and that is the frequency shown at your post test. So we are back to jumpers. I know you have indicated you checked and rechecked the jumpers, but I am going to ask you to check them again. This time I have an additional strange request: For the jumpers that are on, look inside and make positive the little metal shorting bar is present. Quite some years back in encountered a system with jumpers that would not do what they were supposed to. Found that one "jumper cap was missing part of its metal shorting connector. One pin had a socket, but the other had nothing! I'm looking at straws here.

Another thing to look for is that motherboards frequently have a silk screened chart for frequency settings. May or may not be jear JP7 - anywhere there is space.

Another thing to try. is the following I saw on one system long ago. While you power on bootup, holt the Page Up key pressed the entire time; that was supposed to load the CMOS bios with default settings.

Keep trying! I will keep racking my old brain cells. Wish Google worked on it.
GrandpaCarl

 

[DBSSP]

Well, there's something we hadn't thought of! Boy we are looking through proverbial straws! Jay
"If the words up and down were reversed, would you trip and fall or trip and fly?"

 

[Crustyoldbloke]

Hi Carl & Jay

Straws are indeed what we are clutching at here. OK, the metal shorting bars are present. I tried booting with page up and afterwards with page down selected the entire time. No change. The keyboard is not powered up at all.

Looking at JP7, four jumpers and eight pins, current config is 1&2 not connected, 3&4, 5&6, 7&8 all selected(shorted).

I've taken the graphics card out and selected on-board graphics. Result still the same.

Why doesn't the MOBO see or power the keyboard? The old board did.


Die dulci fruere!
Phil
24 Beers in a case, 24 hours in a day.......coincidence??

 

[CapsMan]

If pageup doesn't reset your cmos memory then disconnect power from the mainboard and move the jumper cap into the CLEAR setting for a few seconds, (I guess on JP4, probably short pins 1&2, then replace to pins 2&3, use a magnifying glass, not a pint glass!) Please put all the wires back the same way they were, and keep a record, before you fiddle and break it, which it sounds like youv'e already done!

Finally if all else fails, I highly recommend you visit this fun page: The illustrated guide to breaking your computer: by Tom Murphy
Click the computer big icon pic.

http://www.webtree.ca/newlife/

CapsMan.
The Computer Unprofessional.

 

[Crustyoldbloke]

CapsMan

I think you're missing the point. The keyboard does not power up. If it did, I really think that I would be in with a slight chance. I believe that there is a problem elsewhere which we haven't found as yet.

I've booted skinny, to no avail. I've checked out the RAM - OK. The CPU reads albeit wrongly, and the BIOS is working. I've tried various configurations of different drives. That leaves the PSU or MOBO.

I don't think much of coincidences, so that leaves the PSU. That needs to be checked by someone more technically minded than myself; hence the wait whilst a friend, who is on holiday in Scotland, checks this piece of the puzzle.

In the meantime, I will try and calm myself and perhaps get a few rounds of golf in. Die dulci fruere!
Phil
24 Beers in a case, 24 hours in a day.......coincidence??

 

[edfair]

Does the kbd set of lights flicker on and off when power is first applied. Mormally they do while the keyboard controller initializes. Then later in the POST the keyboard is sent a ID question and the lights flicker again.
If the first lights don't flicker you have power problem coming off the M/B.
Do you have a video splash screen showing at the start of POST? Shorted address or data lines would put garbage out there. And if you have garbage there you'll never get thru POST to do anything else.
With all that has gone before I have a stromg vibe about the CPU. Ed Fair
unixstuff@juno.com
Any advice I give is my best judgement based on my interpretation of the facts you supply. Help increase my knowledge by providing some feedback, good or bad, on any advice I have given.

 

[Crustyoldbloke]

Hi Ed

No lights on KBD at all. The MOBO has a jumper setting for KBD power. The old MOBO was configed as Disabled, but it worked, and KBD lights lit up. This new MOBO is also set to Disable and does not power up, even when I change the jumpers to Enable nothing happens.

I have an initial splash screen, looks normal enough, and then a second screen which I copied on the previous thread around post #96 or so. The previous thread was entitled "pc boot problem"

I was contemplating changing the MOBO back to the original, just to see if the KBD lights flash and then stay on. Now that might tell me that this new MOBO is no good.
Die dulci fruere!
Phil
24 Beers in a case, 24 hours in a day.......coincidence??

 

[GrandpaCarl]

HI PHIL,

No lights on the keyboard at all. Now that is a different clue. There are more than one pin supplying +5V from the power supply to the motherboard. Since a broken wire from the PSU to the motherboard connector should be rather noticeable, I will assume that is not the case.

If you remember my analysis of the problems that could have occurred due to the bent pin on the mouse connector, I indicated how it would be possible to force parasitic diodes to become forward biased in the chip on the old motherboard that communicates with the mouse. This sometimes could cause an extremely high current flow on the motherboard. It is possible that part of the power supply was damaged, but you should have noticed no lights on the old motherboard as well. Let me know the results of reverting to the old motherboard.

Wish you were next door. I would just lend you a power supply, as I keep a spare.

Going back to the classroom again today, so I will be out of the circuit for a while.

GrandpaCarl

 

[Crustyoldbloke]

I wish I was next door aswell, California sounds good to me, better than "rip-off Britain" where we pay more for most things than do our European neighbours. And they get more sunshine.

Die dulci fruere!
Phil
24 Beers in a case, 24 hours in a day.......coincidence??

 

[Crustyoldbloke]

Ok, so I've reloaded the original MOBO and it behaves in exactly the same way as it did when I first entered this forum i.e. it POSTs, goes through BIOS OK, everything seems OK, correct drives and CPU speed, but that MOBO is probably no good.

So do I have two knackered MOBOs or is there something really stupid staring me in the face that I can't see?

The two boards are identical even down to batch numbers.

Well, it's a bank holiday this weekend, so everyone has Monday off work. I'll work on it and see what I come up with, but I'll stay in touch with this thread.
Die dulci fruere!
Phil
24 Beers in a case, 24 hours in a day.......coincidence??

 

[edfair]

But you didn't answer whether the kbd lites flash at power up.
There may be legitimate reasons why they do or not, but this seems more a power problem at present.

All +5 lines should be common, so one being broken wouldn't generally stop the kbd from initializing. Same goes for common.

Wish you had a logic probe or digital VOM. Sure would like to know what voltages you are dealing with. Ed Fair
unixstuff@juno.com
Any advice I give is my best judgement based on my interpretation of the facts you supply. Help increase my knowledge by providing some feedback, good or bad, on any advice I have given.

 

[Crustyoldbloke]

Ed

The keyboard doesn't flash or light up in any way when using the new MOBO, but works fine with the old MOBO. Therefore, I would conclude that PSU is OK. Is my logic incorrect?

Die dulci fruere!
Phil
24 Beers in a case, 24 hours in a day.......coincidence??

 

[GrandpaCarl]

HI PHIL,

Well I am back. Got your report on resilts of the old MOBO back in place. That seems to rule out the PSU, I think. By the way, I have some schematics of a couple of PSUs where there are several 5V outputs that are not common. An ATX PSU has to have a minimum of 2 5V outputs, one of which stays on even after shutdown.

Look real carefully at the newest motherboard, through a magnifier light, if possible and look for possible tool or handling scratches. Some of the printed circuit runs are very fine and quite vulnerable. And though you thought straightening of the mouse pins was microsurgery! Wait until you have to VERY carefully scrape through the solder mask of a 5 mil line and then bridge a solder connection using a fine strand of tinned wire. Careful on soldering, for the solder can dissolve the copper. I used to have a roll of solderwhich was saturated in copper for soldering fine wires. It was British and was called Ersin. It is a challenge for sure.

There is definitely a difference between the two motherboards, you have proved that for certain. We know that likely one or more chips got damaged from the shorted mouse pins on the first board, but the second board is acting differently in a different area. The results you are getting definitely point to a portion of the motherboard not getting +5V. A tiny tiny tool mark can cause that, and you sometimes have to have very good eyesight to see the problem. I use a magnifier lamp along with a pair of 3.0 diopter glasses stacked together to really scan every line on the board on both sides.

By the way, did you find any information at all silk screened on the motherboard regarding setting of clock frequency. I looked at several old motherboards and each one had a chart, but it was always some distance from the actual jumpers! On one of the boards I found the chart, but never could find the jumpers!

If I come up with any other nutty ideas, I will let you know. We have exausted the reasonable explanations other than a CPU problem which you have not completely exhonerated.

Good luck,

GrandpaCarl

 

[Crustyoldbloke]

Hi Carl

I've found the jumper charts. The config on both MOBOs is set to 133 CPU and 88.9 SDRAM. The voltage is set to 2.9 and the other jumper set is marked BF1, BF2, BF0 and gives a result as a ratio of 4.0 (I always thought that ratios were expressed as 4:1 etc).

Scrutinising the new MOBO, (presbyopic these days)I can see a light, fine, 4-inch scratch on the underside of the board. It does not appear to have severed any of the circuits, but it may have reduced their capacities.

I think the next thing is to test the CPU and if that is OK, another MOBO might be the answer. Die dulci fruere!
Phil
24 Beers in a case, 24 hours in a day.......coincidence??

 

[GrandpaCarl]

HI PHIL,

Voltage of 2.9V seems high, but don't remember what CPU you have.

Glad you found the clock charts. I just thought they ought to be there somewhere!

One hint to aid in checking that scratch to see if lines were severed is to shine a light through from the other side. I would look VERY closely.

It might not be a bad idea to try another motherboard if you can swing it.

GrandpaCarl

 

[Crustyoldbloke]

Hi Carl

CPU speed is 333. The scratch is located on the reverse of the onboard graphics gizmo and I know that works fine. The board is a 4 layered affair, so light does not shine through unfortunately. My eyesight nearpoint is around 3 inches without my specs, so I can see very clearly each line. My son has 20/20 vision, but he can't see things really close up like I can (I knew getting old had to have at least one advantage and that's probably it).

I don't understand the finer points of CPU speeds. To an idiot like me, if the CPU is rated at 333, then why is the board set at 133? What's happened to the other 200?

This is not my first encounter with a MOBO problem. I had this PC built for my son, some 3 years or so ago. Originally it had the gigabyte board (which is why I've got a booklet for it), which fried big style. The smell was awful and I could see the damage. I swapped that PC for this one.

Since then, I've changed the modem (lightning strike took it out along with 4 TV's, a satellite system, Hifi and 2 video recorders) and the heatsink fan, apart from installing a graphics and sound card.

Kids! Who'd have 'em?




Die dulci fruere!
Phil
24 Beers in a case, 24 hours in a day.......coincidence??

 

[GrandpaCarl]

HI PHIL,

I have skipped learning all the details of the clock speeds on a motherboard. I do know that the bus speed that the RAM works on is usually much less than that of the internal core of the CPU. The 133MHz is referred to as the font side bus speed, and it needs at least PC133 memory speed. PC100 RAM could only be used if you reduced the front side speed to 100 MHz or less.

The voltages required by a CPU chip is different for the core part as opposed to the outside interfacing section. Each CPU type requires a different set of voltages. Ususallt the manufacturer's web site will tell you the recommended voltages.

I forgot about the extra challenges of a multilayer printed circuit board. Still, a gash on the circuit board can do a lot of damage. Most of the runs on the outside layers are signal lines, with the inner layers being dedicated to power (+5V) and ground.

Sounds like you were lucky if the lightning strike only took out the limited things you mentioned. Have designed watering systems for golf courses in the Midwest U.S., and have had lightning burn a 3" diameter hole in the steel case, yet no damage to the circuit. The client was impressed.

Your best bet seems to be to get a third MOBO and preferably a new CPU.

GrandpaCarl

 

[mainegeek]

If you are running a 333mhz cpu, then try setting your speed at 66mhz and 5.0x

 

[DBSSP]

So did we get any farther in the saga of "The Little PC that Wouldn't"? Jay
"If the words up and down were reversed, would you trip and fall or trip and fly?"