lot cheaper to take out the BIOS chip and have it reprogrammed. I had to do this recently and it only cost £8.00 (about $12.00) But if your friend does it for nothing I suppose theres nothing cheaper than nothing.
After painstakely carefully assembling interface for the chip I found out that the programmer did not support EEPROM's big enough! What a letdown. Apparently I was mislead by Microchip's stupid way to mark their chips, as their 25*40 is not the same as Winbond's (read: practically everyone else) and the difference in chip sizes is huge (4kb vs. 4Mb). The upper limit is not imposed by the hardware, but Microchip's software. I guess they don't like using their programmer with other mfg's chips. Perhaps in future I consider PonyProg or JDM as viable option... if I have spare time to make one and I have all the parts. As far as the cost of this attempt goes... well, as I borrowed the programmer, it was free. Free is always better choice for me than the ordering a new pre-programmed chip. For that I would have to obtain floppy drive (for saving BIOS file to a floppy), mail the floppy and hope it would arrive intact, order new chip(s)... and pay. Chips are dead-cheap, but last time I checked, the service did cost about 8€ plus 5€ per chip programmed. No results guaranteed. I usually prefer local services if available, but otherwise domestic ones are enough. Too much hassle with foreign ones. Only bright side with this experience is that it ensured that I will never buy ASRock products in future. I also check more details of the products in my shopping list even more carefully before sales.
SPI Support Oh, yes, I forgot to mention that the spipgmw tool supports absolutely ANY SPI. Even if it doesn't list your chip. Simply enter the source file & the chip size (the size of the file you want to program) & VIOLA. Commands look like this: 1.) spipgmw /e --> Erases the chip 2.) spipgmw /u --> Unlocks the chip (shouldn't take more than 1sec) 3.) spipgmw /p romfile.bin --> Programs the SPI with your rom file 4.) spipgmw /d dump.bin --> Extracts the file from the SPI To program a 1024K chip with a 128K file (like I did for my GFX card), spimply use command (erase the chip first): spipgmw /p romfile.bin --> If it prompts you to enter the chip sixe, enter 128 (If you only want to program 128K). So, you can use a 2048K chip, instead of a 1024K chip also.
Thanks!! I have to save that page for better time slot and try to decypher that translation. I have no idea what "All this will scratch the universal plošňáku looks like this:" really means, but circuit is dead simple! I may arm myself with a few extra chips just in case...
@Carel. Very interesting stuff. I wish i had seen this a couple of weeks ago. Flashed my old HP mobo with a mod and bricked it. The guy at the local computer shop told me spi re-programming was unlikely to sort things out so i bought a new mobo. Gladly not from him.
Lets see if I have some spare time to find the components at the weekend. One small thing is still a bit unclear. The capasitor is polarized (no other info about it's type?), according to schematic diagram. I still can't make my mind if it really is 1mF (1000µF) or is the marking somehow mixed up. I guess it's the latter (and meant to be 1µF), but... To make matters harder to guess, there are different types of capasitors used in different pictures. I have to dive in to my component shelves and try some educated experimenting when I have time. Hopefully I don't fry the chip, as I was astonished to find out how much more they cost now! I'd say there has been about tenfold increase, compared to the last time I checked! Still, thanks for this great information!
Capacitor Yes, there are no specific capacitor values to use. It depends on what power supply you'll be using. Don't worry about the capacitor, the value of the capacitor will depend on the noise / ripple in the VCC+. I didn't use a capacitor in my circuit, because I used a 3v battery (battries don't hav any ripple or noise). If you'll be using a power supply from the AC Mains, like the 3.3V from a computer psu, then you'll maybe need a capacitor, but still, the capacitor is not necessary in all cases. Capacitors store power, to cacel the noise & fall in the amplitude of the waves formed by the psu's vcc+. In the case of a battery, it'll be a sraight line, indicating there is no ripple or noise. If I were you, I'll use a battery, 3V, then you don't need any capacitor. The capacitor is not a required component, and if you really, really want to, then you can use a 1000uF, but it won't have any effect.
Finally got a bit of free time, but no suitable parts found. So, this project will be delayed until I have time to get them from local electronics shop. That being said, I'm no hurry with this project, as the motherboard has been waiting long time already and I just found something more important (at the moment) which I need to solve first... I will report after I have time to try this...
Hi all, I just signed up so I could post to this forum. I've been trying to get this rig to work but despite it's apparent simplicity, spipgm continues to fail to recognize the chip, and often reports different results when I try to ID the chip [spipgm /i]. I am using a 1.7ah battery that delivers 3.5v so I am not using a capacitor or any resistors. There isn't much information about this rig other than a few forums like this one, but another good reference is here: http:||forum-en.msi.com/index.php?topic=129989.0. [Sorry my first post on this forum and I'm not allowed to post a link or a photo...sigh] Someone posted a good few photos of his rig, but his differs from the original in that the ROM's pin 5 (DIO) has a capacitor instead of a resistor. My parallel port connector is actually a resoldered RS232 port (same thing effectively) designed to connect to a desktop motherboard. Thus, the wires which I reused are ribbon cable type wires and are not shielded. I've checked the wiring so many times I see it when I close my eyes and recently had a friend double check me and he thinks it's correct too. I'm going to try shortening the wires to see if that helps any. Oh, and the chip I'm trying to program is an Atmel AT26DF321 SOP8. I originally bought a Willem programmer which was supposed to support the chip but although the Willem can ID the chip, it always fails programming. My guess is the software wasn't tested for the Atmel chip. As I understand it, the resistors are only being used because the original spec used a 5v power supply and the chip needs around 3.5v max (Atmel range spec is 2.7v - 3.6v). Because I have a 3.5v battery I don't need the resistors. And, as I understand it, the capacitor is used to ensure stable power supply. Again, because I am using a battery that produces little noise I don't need the capacitor either, correct? I am wiring the ground to the battery's negative pole. I tried attaching a simplified wire mapping drawing which shows how I have wired my rig, but since this my first post I'm not allowed Hmmm...preview still shows the image so maybe it will get attached...sigh. Any ideas why it doesn't work?
Cool! Looks like MDL allowed me to attach my picture though it said it wouldn't work. Anything wrong with my understanding of the circuit?
Ok, people seems to get confused about the 5v signal & the resistors. Firstly, yes - the resistors really, really, really is necessary... And no, the resistors are not to bring down any voltages... Resistors brings down current & signal drive strengths. Although all parallel ports have a signal voltage of 5v, this doesn't have an influence on the VCC. The signal used for transferring data has got nothing to do with the VCC {power from battery to chip}. The signal, which is 5v, will NOT blow the chip, but the signal drive from the parallel port are too high, which causes the chip to send weaker signals to the parallel port. Please use 270ohm resistors, on the CS, SO, SCK & SI pins. In other words, pins 7, 8, 9, 10 that goes to the parallel port. Your cables should NOT at all be twisted, this will create a magnetic feld & cause signal distortion. The cables should be very, very short. Not longer than 20cm. Keep in mind that the battery should have a very, very short cable. I'm talking about 5cm from the SPI / EEPROM! You would ask why, the reason is simple: You're not using any terminating resistors, so the voltage supply will ripple if the SPI are pulling power from the battery. The capacitor is needed, use a 1000uF {it works the best} capacitor. This should not be more than 1~2cm away from the SPI / EEPROM. Now, keep in mind that this type of programming is not very efficient. This circuit is not able to adjust it's own singal strengths, pulse widths & driving strengths. That's why it is very important to build the circuit as I describe. Another thing, higher VCC from the battery means the pulse signal from the SPI to the parallel port will increase, forcing you to use a higher resistance. {keep in mind the resistors are needed, not for the VCC, but for the signal pulses}. Build the circuit as follows: 1.) no twisted wires, not longer than 20cm {they should be kept straight}. 2.) use 270 ohm resistors. 3.) use a 1000uF battery, not farther than 1~2cm from the SPI / EEPROM. 4.) Then, spipgmw {keep in mind you should be using spipgmw, not spipgm}, should recognise the chip as 371111h or xx1111h. If it doesn't detect the chip, adjust the VCC supplied to the SPI, this is VERY critical, the MOST critical. I programmed 3 SPIs using this circuit & I had to test different voltages for different SPIs, each of them are different. For ST M25P10 -> 2.45v, for AMIC A25L080, use 2.8~3.0v, for Winbond W25X80, use 2.6v. Even if the datasheet states the chip operates at 2.7~3.6v. Adjust the voltages in steps of 0.10, you'll eventually crack it, it's very easy. Adjusting the voltage is the most tricky part {don't use resisotrs on the VCC, it'll cause whining & ripple, use a voltage regulator instead, like LM317}. Some chips are full of crap, others aren't... I'll post a picture in a few minutes...
U D'man! Thanks SO MUCH Carel for your reply! I based my understanding on what others were saying and didn't even think about the juice coming through the parallel port itself. I know the parallel port has power of its own...duh!!! I wouldn't feel SO stupid if I hadn't looked at this for days now! Since posting my question I have built a second rig with VERY short cables, thinking I might be dealing with noise. The parallel interface dongles to the back of the notebook's parallel port, so no cable inbetween, and the SOP8 rig is attached to it. Total length from parallel pin to SOP8 socket is 7-8 cm. I'll have to splice it now to insert the resistors. Well, I'm sure this post will be useful to others since there was clearly confusion about the needed hardware. Thanks again for the quick answer. Since this thread isn't heavily active I didn't expect to get an answer so soon. /\/\ike
One other question, since I'm getting clear answers, if you don't mind... Since I'm using an El Cheapo SOP8 socket and since the lines are so close to each other due to the narrow pitch of the chip, I've wrapped the solder work in JB Quick epoxy to prevent accidental shorts. It also protects the mechanical stuff which is pretty fragile. According to my multi-meter the epoxy does not conduct electricity but since it does contain metal particles in it, I wonder if it is OK to use for this purpose. Opinions? It looks like you may have done something similar with yours? Thanks again!
Allthought the thread is old, I had to come back after I finally managed to dig up my login info from the dead HD. I made the circuit with breadboard, because it was quicker than making a more permanent type adapter. Lo and behold, the thing worked at first try! Huge thanks for helping out! Perhaps I'll make the the adapter type some day, as this may save time and money in future.
I would have to agree take out the bios chip and have it reprogrammed or buy a new mother board. the first option being cheaper