If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#11
|
|||
|
|||
Problems rebuilding system
[snippage]
The GA-EP45-DR3LR has dual BIOS, two tiny surface mount chips near the battery and CL_BIOS pins. It is designed for safe upgrade, with networked OS software. Gigabyte gives recommendations on what software to use. BIOS is V1.0 It is an unused MOBO, never sold. It came in a brown cardboard box with the IO shield attached with a plastic strap. It is a virgin. It had been in storage so long the battery was dead. It has nice features, like dual channel memory. At one time, high end RAM was advertised as 'dual channel;. It is not a property of RAM but of the MOBO. With dual channel memory and a 3 GHz 4-core Q9650 with 12MB cache, why worry? The advantages of 8-core may be exaggerated. A GTX 970 GPU gives it more horsepower. Its contribution showed up in "Task Manager". What I am saying is that Big Bucks Boyz may get better economy with an old system, by trading labor for capital. |
#12
|
|||
|
|||
Problems rebuilding system
From manual, here are the beep codes: Q: What do the beeps emitted during the POST mean? A: The following Award BIOS beep code descriptions may help you identify possible computer problems. (For reference only.) 1 short: System boots successfully 2 short: CMOS setting error 1 long, 1 short: Memory or motherboard error 1 long, 2 short: Monitor or graphics card error 1 long, 3 short: Keyboard error 1 long, 9 short: BIOS ROM error Continuous long beeps: Graphics card not inserted properly Continuous short beeps: Power error Since I have not heard any beeps when the table says I should, that would imply a problem with the front panel connecter. I'll be back after Xmas duties. It requires a working motherboard+CPU combo. Then you'll get beep codes on SPKR. SPKR is not polarized. Just make sure the span is correct for the connector (some OEM cases might have a 1x2 connector, while some retail motherboards use a 1x4 pattern). You can use connector strips and lift the tab on each, to shift wires within shells and make the correct connector for the job. The SPKR body should be electrically isolated from the two wire leads. The wire leads go to the speaker coil, and the coil moves in and out with the cone. Even so, a few computer cases here use plastic mounts for SPKR. The front panel header is reasonably bulletproof. RESET and POWER are SPST momentary contact switches, normally open. You push the button, and the switch closes for a moment. The LEDs are polarized as you say. The LED is rated for 5V PIV (Peak Inverse Volts) so cannot be harmed by reversing the leads. Each driving circuit has a series resistor to limit current flow. If they don't light, you reverse the 1x2 connector and try again. Normally the spacing and span of various signal assignments on the Front Panel connector, avoids the possibility of shorting two power pins placed next to one another. If there is a SPKR section of a Front Panel connector, the "hot" end should not be near a GND pin. The ones I've looked at, generally have fairly safe signal assignments. While there are connectors on a computer that get crushed and you cannot visually check them, the Front Panel connector isn't one of the ones where people habitually do bad things. I don't think I've had any reports of anyone managing to start a fire using nothing but a Front Panel header problem :-) (If you pinch the "hot" wire of SPKR between the computer case door and the computer case, which is GND, then the wire will get smoked. And that has happened. Normally those wires aren't sitting near the door.) You can get a "no beep" condition, by using a reset button crushed in the ON position. Usually OEM computer cases are the ones with sufficiently cheesy buttons on the front of the computer, to make incidents like this possible. I have one computer here, where the buttons are such, I know that some day that's how those buttons will fail. The buttons speak of cheapness. Paul Thanks Paul, I'll look into it. |
#13
|
|||
|
|||
Problems rebuilding system
[snippage]
You can get a "no beep" condition, by using a reset button crushed in the ON position. Usually OEM computer cases are the ones with sufficiently cheesy buttons on the front of the computer, to make incidents like this possible. I have one computer here, where the buttons are such, I know that some day that's how those buttons will fail. The buttons speak of cheapness. Paul Paul, I may have found the problem, no beep. On the one hand the GA-EP45-DS3L provides a 10 pin keyed audio header while the MX330-X provides a compatible plug. The mike and line-out go to connectors on the IO shield. I have headphone-mike connected but that is not where the beep goes. GA-EP45-DS3L provides a 20 pin front panel header. I have insured that polarity of POWER, RESET and POWER LED are correct. I do not yet know about HDD Activity LED. But when I look closely at GA-EP45-DS3L manual, polarized Speaker Leads are connected to pins 20 and 14. I suppose that is a good idea. However, the MX330-X front panel header ribbon cable does not provide these, none that I see. I guess I need to send an email to Cougar customer support. Odd, cannot see this problem flagged in any review. Maybe I need to dig into the MX330-X with a flashlight and magnifying glass? Remember the pin connectors are real tiny. The MX330-X manual is one page, Mickey Mouse. I need better documentation. The Power Switch is confirmed. The Power LED is confirmed. The Reset Switch is confirmed. Maybe the HDD Activity LED is in fact polarized Speaker Leads. These pins are so tiny it is hard to tell. But it is only these that need closer inspection. |
#14
|
|||
|
|||
Problems rebuilding system
[snippage]
The Power Switch is confirmed. The Power LED is confirmed. The Reset Switch is confirmed. Maybe the HDD Activity LED is in fact polarized Speaker Leads. These pins are so tiny it is hard to tell. But it is only these that need closer inspection. This is a URL to an image of the MX330-X front panel connector: "How-to-connect-case-cables-system-panel." https://assets.rockpapershotgun.com/...stem-panel.jpg Note there is no HDD Activity LED between pins 1, 2. They belong elsewhere. On the GA-EP45-DS3L they belong on pins 20, 14. In this image they appear on pins 18, 14. Now to sleep. |
#15
|
|||
|
|||
Problems rebuilding system
The Power Switch is confirmed.
The Power LED is confirmed. The Reset Switch is confirmed. Maybe the HDD Activity LED is in fact polarized Speaker Leads. These pins are so tiny it is hard to tell. But it is only these that need closer inspection. This is a URL to an image of the MX330-X front panel connector: "How-to-connect-case-cables-system-panel." https://assets.rockpapershotgun.com/...stem-panel.jpg Note there is no HDD Activity LED between pins 1, 2. They belong elsewhere. On the GA-EP45-DS3L they belong on pins 20, 14. In this image they appear on pins 18, 14. Now to sleep. [correction, even and odd # pins are on opposite sides of header] Houston, we have a problem. HDD Activity LED pins 1, 3 and MX330-X connection is fused and clearly marked HDD Activity LED & polarity. While Power LED MX330-X connectors are not fused, polarity marked and might be a candidate to connect to pins 20, 14, they are confirmed as Power LED. When ON the Power LED glows steady and does not pulse as expected for front-panel speaker beep. Maybe an email to Cougar is needed. Strange, this problem was not flagged in reviews. |
#16
|
|||
|
|||
Problems rebuilding system
On 12/13/2019 6:11 PM, Paul wrote:
You can get a "no beep" condition, by using a reset button crushed in the ON position. Usually OEM computer cases are the ones with sufficiently cheesy buttons on the front of the computer, to make incidents like this possible. I have one computer here, where the buttons are such, I know that some day that's how those buttons will fail. The buttons speak of cheapness. Anything mechanical is always suspect. I've been using a couple of identical Antec cases, not normally noted for being cheap, for computers now residing in the basement "computer room". Not too long after purchase (but outside any hope of warranty) the power switch failed on one of them and drove me crazy for a while. Rather than give up I swapped the connections between "power" and "reset" on the MB and it has worked fine for years now. Only real problem is that I can almost never remember which of the cases has the swapped switches. If I had any sense I'd apply some masking tape over the dead switch to remind me... |
#17
|
|||
|
|||
Problems rebuilding system
Norm Why wrote:
[snippage] You can get a "no beep" condition, by using a reset button crushed in the ON position. Usually OEM computer cases are the ones with sufficiently cheesy buttons on the front of the computer, to make incidents like this possible. I have one computer here, where the buttons are such, I know that some day that's how those buttons will fail. The buttons speak of cheapness. Paul Paul, I may have found the problem, no beep. On the one hand the GA-EP45-DS3L provides a 10 pin keyed audio header while the MX330-X provides a compatible plug. The mike and line-out go to connectors on the IO shield. I have headphone-mike connected but that is not where the beep goes. GA-EP45-DS3L provides a 20 pin front panel header. I have insured that polarity of POWER, RESET and POWER LED are correct. I do not yet know about HDD Activity LED. But when I look closely at GA-EP45-DS3L manual, polarized Speaker Leads are connected to pins 20 and 14. I suppose that is a good idea. However, the MX330-X front panel header ribbon cable does not provide these, none that I see. I guess I need to send an email to Cougar customer support. Odd, cannot see this problem flagged in any review. Maybe I need to dig into the MX330-X with a flashlight and magnifying glass? Remember the pin connectors are real tiny. The MX330-X manual is one page, Mickey Mouse. I need better documentation. The Power Switch is confirmed. The Power LED is confirmed. The Reset Switch is confirmed. Maybe the HDD Activity LED is in fact polarized Speaker Leads. These pins are so tiny it is hard to tell. But it is only these that need closer inspection. ga-ep45-ds3l(r)_e.pdf Page 27 FPANEL PWR PWR LED Switch SPKR + - + - + - X X X X X - - X X X X X X + - - + HDD RST LED Switch Polarity of RST Switch, PWR Switch, SPKR, does not matter. All assemblies are electrically floating. PWR LED and HDD LED can be reversed if they refuse to light. The nine pin section does not need to be completely wired up while testing. The two LED indicators are superfluous, until final build time. Just the PWR switch can be connected for the style of testing to come. What is not optional at this point, is the SPKR connection. You need the SPKR connection to verify that the "missing RAM" and "missing video" functions work properly, which tells you that the motherboard/CPU is working, and that when you plug the RAM in, the RAM is detected. Once you have completed that phase of testing (successfully), then you can go back to solving the video card issue, whatever it is. Since the chipset is P45, I can't see a reason for the motherboard to be screwing this up. You should have good luck with PCI Express compatibility with such a motherboard. If the card has too much onboard GPU RAM, then it's possible the OS won't start (address space issue). When doing the "RAM detected, video not detected" test, you'll be using a single stick of RAM. Use one of your *lower capacity* sticks. The smallest I have is probably 1GB for this particular test you're doing. In years past (year 2000), you could get 64MB as a stick, but the "smallest" has gone up a lot since then. Your GPU has 3.5GB of onboard RAM (GTX970). But the BIOS screen should always come up, unless the "max address space" is exceeded. Let's say that is 8GB, just pulling the wrong number out of the air. I could use a 2GB single stick of RAM plus that video card, for 5.5GB total. I could be reasonably assured that the BIOS screen would then come up. Once you can see the screen, *future* tests can use your bigger DIMMs. The objective is to get to see the BIOS screen first. Once you prove there isn't a motherboard hardware failure, then lots of other test cases will make sense to do later. Windows 10 can start with as little as 256MB of system memory (tested in a VM environment, to see if it was possible). While the stated minimum is 1GB, that's the stated "comfortable" minimum, not the "absolute" minimum. If you run Windows 10 with 256MB, a task called the "Memory Compressor" runs almost continuously, as a measure of how "annoyed" the OS is about this development :-) When you use 1GB of RAM, the memory compressor almost comes to a stop. Paul |
#18
|
|||
|
|||
Problems rebuilding system
"John McGaw" wrote
You can get a "no beep" condition, by using a reset button crushed in the ON position. Usually OEM computer cases are the ones with sufficiently cheesy buttons on the front of the computer, to make incidents like this possible. I have one computer here, where the buttons are such, I know that some day that's how those buttons will fail. The buttons speak of cheapness. Anything mechanical is always suspect. I've been using a couple of identical Antec cases, not normally noted for being cheap, for computers now residing in the basement "computer room". Not too long after purchase (but outside any hope of warranty) the power switch failed on one of them and drove me crazy for a while. Rather than give up I swapped the connections between "power" and "reset" on the MB and it has worked fine for years now. Only real problem is that I can almost never remember which of the cases has the swapped switches. If I had any sense I'd apply some masking tape over the dead switch to remind me... Thanks John. Maybe reassignment of connections is the answer. Because of propagation delays, the group seems not to have seen my latest post or "How-to-connect-case-cables-system-panel." https://assets.rockpapershotgun.com/...stem-panel.jpg |
#19
|
|||
|
|||
Problems rebuilding system
Norm Why wrote:
The Power Switch is confirmed. The Power LED is confirmed. The Reset Switch is confirmed. Maybe the HDD Activity LED is in fact polarized Speaker Leads. These pins are so tiny it is hard to tell. But it is only these that need closer inspection. This is a URL to an image of the MX330-X front panel connector: "How-to-connect-case-cables-system-panel." https://assets.rockpapershotgun.com/...stem-panel.jpg Note there is no HDD Activity LED between pins 1, 2. They belong elsewhere. On the GA-EP45-DS3L they belong on pins 20, 14. In this image they appear on pins 18, 14. Now to sleep. [correction, even and odd # pins are on opposite sides of header] Houston, we have a problem. HDD Activity LED pins 1, 3 and MX330-X connection is fused and clearly marked HDD Activity LED & polarity. While Power LED MX330-X connectors are not fused, polarity marked and might be a candidate to connect to pins 20, 14, they are confirmed as Power LED. When ON the Power LED glows steady and does not pulse as expected for front-panel speaker beep. Maybe an email to Cougar is needed. Strange, this problem was not flagged in reviews. No, Mr.Engineer, we don't have a problem. https://i.postimg.cc/JnXsBwGc/LED-driver-circuit.gif LED driver circuits tend to have current limiting resistors in series with the power source. This limits current flow to a safe value. In the above picture, the rail was 5 volts, the resistor 330 ohms, and the series current flow cannot be greater than 5V/330ohms = 15mA. Commonly available indicator LEDs (T13/4 size) are rated for 20mA or more. (I have one here rated for 100mA but that's more unusual.) Fuses on motherboards, they cost money, and they're only fitted in certain places. For example, industry practice is to *not* put a fuse in the fan power path. On occasion, the fan power path is burned out because of this. Places that fuses are used: 1) Power pin on the parallel port connector. 2) Power pin on the PS/2 keyboard connectors. 3) Pairs of USB headers have a fuse (each 2x5 gets a fuse). The very cheapest of motherboards, use a single fuse running to 1,2,3. You might find mainstream motherboards with five fuses or so. It's also possible to use silicon fuses. These are an eight pin DIP, sufficient to police current flow in two circuits. When you're counting fuses, those would be easy to miss in the analysis. A silicon fuse uses a MOSFET to interrupt measured current flow. Silicon fuses are a preferred solution inside a laptop (where there is poor cooling available). ******* There are various ways to do LED drivers. "Series" or "Shunt". This is series drive. The bipolar transistor can be a 2N2222 or the circuit can be driven by an open collector 74F06 or a 74F07 (as examples). + - +5V ---Rlimit--- Anode Cathode -- C E ---- GND LED B | transistor drive In shunt drive, the circuit looks like this. + - +5V ---Rlimit--+-- Anode Cathode ---- GND | LED +-----C E ------------ GND B | transistor drive In the shunt case, the Vce of the transistor, when it is on, must be lower than Vfb of the LED. The LED might be a 1.8V to 2.2V indicator. The transistor Vce might be 0.4V. When the transistor in the shunt case turns on, it shunts all the current (the 15mA) to Ground. And since the transistor conducts until Vce is 0.4V, the 0.4V across the LED is not sufficient to cause it to light up. When the shunting path is ON, the LED is OFF. In the series case, the transistor Vce of 0.4V causes minimal loss to the 5V supply. Now, the circuit sees 4.6V. It's delivered through the 330 ohm resistor, and well less than 15mA flows through the LED. (That's since the LED drops some voltage too, causing even less current to flow through the 330ohm.) The circuit above is generous enough to work with red,yellow,green,blue LEDs. If the supply voltage drops to 3.3V for some reason, then you might have a problem getting a blue (or white) LED to light. The ones I own at the moment run from 2.5V. 3.3V - 2.5 - 0.4 leaves only 0.4V across the 330 ohm resistor, or 1.2mA, which doesn't always yield a good lighting experience. The engineer designing the board, would not normally assume blue LEDs are being used, as back when this nonsense started, the LEDs were always red in color. The design is resistant to most "expected" failure modes. If the LED fails short, there's no fire. It is possible, with enough shenanigans by the computer builder, to connect the wrong things together and damage something. The motherboard designer is under no requirement to protect against "stupid stuff". All parts cost money, and the motherboard makers are so profit-incented, they will remove useful gold pins costing only a penny or two apiece, to make more profit. And that's part of the engineers job, is cost-reducing the product, balancing "reasonable" feature sets against BOM (Bill Of Materials) costs. When they put "toys" on a motherboard, the marketing department made them do that. Things like Port 80 displays, Bitchin Betty (Winbond acoustic POST output chip), those are gimmicks to sell more motherboards. And not essential to running an OS with the door closed on the computer case. Paul |
#20
|
|||
|
|||
Problems rebuilding system
Norm Why wrote:
"John McGaw" wrote You can get a "no beep" condition, by using a reset button crushed in the ON position. Usually OEM computer cases are the ones with sufficiently cheesy buttons on the front of the computer, to make incidents like this possible. I have one computer here, where the buttons are such, I know that some day that's how those buttons will fail. The buttons speak of cheapness. Anything mechanical is always suspect. I've been using a couple of identical Antec cases, not normally noted for being cheap, for computers now residing in the basement "computer room". Not too long after purchase (but outside any hope of warranty) the power switch failed on one of them and drove me crazy for a while. Rather than give up I swapped the connections between "power" and "reset" on the MB and it has worked fine for years now. Only real problem is that I can almost never remember which of the cases has the swapped switches. If I had any sense I'd apply some masking tape over the dead switch to remind me... Thanks John. Maybe reassignment of connections is the answer. Because of propagation delays, the group seems not to have seen my latest post or "How-to-connect-case-cables-system-panel." https://assets.rockpapershotgun.com/...stem-panel.jpg What is that a picture of ? It doesn't seem to be your GA-EP45-DS3L. And the black wiring depicted, is a mistake when doing builder work. You want classically colored twisted pairs for each 2-pin thing needing connections. It makes it easier to get the polarity right on the first try (for LEDs). No harm will come from hooking up a two pin pair the wrong way around. I prefer the color code, because it tells me "what they had in mind" when they wired it. Antec in particular, has made mistakes on front panel wiring more than once. The part I consider weird about this, is when they do make a mistake, it's *never* with the VCC pin, never a risk of smoke or fire. How do they make these mistakes, without ever involving a "hot" pin ? I could never figure that part out. If you bring me an Antec case today, I'll get out my ohmmeter and buzz out the wiring harness, before using it. Paul |
Thread Tools | |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Problems rebuilding system | John McGaw | General | 5 | November 27th 19 02:42 AM |
rebuilding computer, need help please | tomgasgas | Homebuilt PC's | 12 | April 5th 07 12:04 PM |
Rebuilding IP5000 | Bullitt | Printers | 27 | November 14th 06 11:58 AM |
Rebuilding from CDs and missing applications | ~~Alan~~ | Dell Computers | 10 | July 5th 06 12:04 PM |
Rebuilding System A8N-SLI | Dave McCaleb | Asus Motherboards | 9 | March 25th 05 11:11 PM |