Just wondering have any of you guys had eproms that blank check, program, and verify good but do not work? I had a new batch of chips come in and none of them work. I have tried 9 in a row. Mind you in between I have programmed other chips and they work fine. I have also used the same exact chip I had from my last order and that programmed.
It makes zero sense. I have erased these and reprogrammed and nothing. Could my programmer be flaky. I even purchased an external power supply and that didn't help. I even read the chip after programming it and the code "from what I compared on the screen" looked identical to the roms I loaded.
I thought maybe the 3 roms I was using got corrupted so I chose a different rom and programmed on old stock and it worked. I then programmed it to a new chip and it doesn't work. Maybe someone with more expertise in how eproms work could chime in.
the programmer is a GQ-4X True USB the chips in question are m27c160
Need more info, post the full part number of the chips that worked and the ones that didn't.
They are identical. Exact same chips. Could something have caused this shipment to go bad? They are from same seller as well. I just don't get how they program and verify then have scrambled graphics or no picture at all.
m27c160
-100F1
58854
ST 1104L
Singapore
Hmmm it looks like 100F1 is 0-70 Celsius. 0 is freezing so if the temps got down under 32 F could that have caused damage to them during shipping. It did just get cold finally.
They were shipped on Dec 29th and delivered Jan 7th.
the F1 rating just means it's only guaranteed to work correctly in a range of 0-70C, not that permanent damage will happen.
The datasheet says "Absolute maximum ratings: Tstg = Storage Temperature - -65C to +150C" so.
EPROMS get slower every additional time they're programmed/erased. It's more likely that the ones you have there aren't fast enough anymore.
Well they are suppose to be new old stock. I don't buy used for that exact reason. I read every time they are erased they lose shelf life. I'm not sure how they act after sitting for years.
I don't think a never-programmed EPROM should age? But...
Not getting the update wouldn't cause problems would it? They released a new update but I never got it as it was working fine and honestly I couldn't figure out how to install it LOL. I hate to admit that as I am quite well with computers but this software they give is not user friendly at all. It has a way of blocking out options for you so you can't do what needs to be done.
Ok I am 100% confident these chips are bad. I did continuity checks on every pin on adapter board and everything. Furthermore I had another m27c160 from my last order that I needed to erase. I tried programming 2 more new chips and bam both bad. I then programmed the chip from last batch after I erased it and it worked perfectly.
I am not losing my mind here LOL these chips are bad somehow. I even uninstalled the software and reinstalled w/the new upgrade. The driver was what doesn't want to update and wat I gave up on last time. Now I see they say it doesn't work on certain models. Maybe that is why I couldn't get it before.
Man soon as my new tip for my Hakko FR-300 comes in I have a LOT of rework to do. Yeah I kind of tried drilling the nozzle cleaner in as it wouldn't go and I broke the end off inside the nozzle..... yeah I knew it wasn't a good idea but I was a bit annoyed at the time w/all the chips failing lol.
Lesson learned I now have on order some zif sockets so I can make a tester board, enough of this crap lol.
A couple of suggestions.
When reading back on your programmer make sure vcc and vpp are at 5v not the higer programming voltage of 6v etc... vpp should be at 5v or gnd.
Write the rom twice with no erase between the two writes.
An eprom is more or less a lot of tiny solar pannels tied to capacitors to hold a charge. They are charged when exposed to UV (set to logical 1). Writing a byte replaces the negative charge in a memory cell placed there by the solar panel with the VPP voltage. (This is not 100% accurate but good enough explanation to the non technical tinkerer)
If the memory cell was not erased enough its charge may not be sufficient to trigger a 1 or 0 condition when read at a high speed. This is likely what is happening. A long erase followed by two writes without an erase will make for a better read. If you can increase the write pulse time in your programmer this well help too.
BennVenn wrote:
A couple of suggestions.
When reading back on your programmer make sure vcc and vpp are at 5v not the higer programming voltage of 6v etc... vpp should be at 5v or gnd.
I do not know how to change that. I just click read. Unless this is somehow done in the software, I am still novice to the ins and outs. The programmer itself is fully encased the only thing I can do on actual unit is plug usb in and power.
BennVenn wrote:
Write the rom twice with no erase between the two writes.
How is this possible? If the chip is not blank it will fail the write. Therefore soon as I go to write again it will just error on me.
BennVenn wrote:
An eprom is more or less a lot of tiny solar pannels tied to capacitors to hold a charge. They are charged when exposed to UV (set to logical 1). Writing a byte replaces the negative charge in a memory cell placed there by the solar panel with the VPP voltage. (This is not 100% accurate but good enough explanation to the non technical tinkerer)
If the memory cell was not erased enough its charge may not be sufficient to trigger a 1 or 0 condition when read at a high speed. This is likely what is happening. A long erase followed by two writes without an erase will make for a better read. If you can increase the write pulse time in your programmer this well help too.
These chips are suppose to be new, they are not suppose to have ever been programmed and they will read blank. They look new as well. Unless they know a way of removing chips w/out bending the pins at all and they know a way of getting every spec of solder off because these chips have none. I do not see how this can be the problem.
You're not erasing them with a uv lamp?
You should be able to write the same rom to a chip unless it does a blank check first. What kind of programmer do you have?
And yes, many of the cheap eeproms from china are pulled from scrapped boards. EPROMS typically are socketed so it would only be a matter of them removing the sticker and glue remnants and a quick erase. It would be hard to tell.
I only erase if it is needed. These chips are listed as "new". They are suppose to be new old stock that have just been sitting for awhile but never used.
http://www.ebay.com/itm/351493027694?_t ... EBIDX%3AITI have ordered a lot from this guy different types of chips and rarely do I need to erase them, only when they fail, which up until now was maybe 1 or 2 out of every 50. I completely stopped blank checking as it didn't matter. If it was not blank it would just fail the write immediately anyhow.
I have a GQ-4X true usb universal programmer.
http://www.ebay.com/itm/TrueUSB-GQ-4X-V ... XQTT9RqYLF
If you have an oscilloscope you could check the eprom speed. Having a whole batch fail in the same way is odd. Bad batches do happen though. I've got 10 mega128 IC's each with blown port drivers. All new. I've been sent fake rebadged mosfets, fake YM synth IC's etc.
If you have extracted the bin from the eprom via your burner and it is a bit for bit match I'd suggest try cooling (butane can upside down) or heating the ic then try it in your console. If it is speed related (due to aging or abuse) temperature changes could show it.
Oh and use caution with butane as a freezer spray!
So I was able to take 2 32 zif sockets I had and cut them down and place them side by side. Luckily for me the pins in exact spot after the 1st zif socket ended 17 and other side would be 26 were unused so I had a tiny gap to work with to make a test board. It looks like crap and will probably pop a trace eventually it is so tight but it does the job for now until I get the bigger zifs in.
Anyhow I programmed 3 more chips and same crap. I then had another chip from an old batch still and just like every other time I programmed an old chip in between these new chips it worked fine. I did see a "double write" on the programmer and I did that and still got same results.
Chip wrote twice and verified and had garbled graphics. I have already started a return and purchased another batch from another seller. He sold me good ones a few months back. so hopefully they will work fine.
I have programmed like 13 out of these 50. I am tempted on programing them all and see if they all suck. I now have very little confidence in the seller I was buying from.
Ok what is going on here people. Look at the date I made this thread. It was exactly 1 year ago. I purchased a batch of chips in December and January they gave me problems.
Now the same exact thing happened. I purchased a batch of chips last month "December" and they program bad again just like the batch did a year ago.
However here is the kicker, Only on snes. I was trying to diagnose the problem and since I make genesis games as well now, I tried a genesis game and it worked. So Then I went back and tried snes again, multiple games and same thing, bad glitches on all.
So I grabbed the batch of chips I got over a year ago that would not program correctly and tried a genesis rom on them and it worked!!! Unreal. What is going on here. I'm glad I can use the chips now but what is going on here.
They are all 100F1
Sounds like your SNES board is defective or your soldering is wrong. If ICs usually erase, blank check or verify they're 99% fine. I never had problems with EPROMs unless they didn't blank check or erase, those have been thrown in the trash.
How did you solder the 27c160 to a SNES board? Which SNES board did you use as a donor?
SHVC-1AXM
SHVC-1AXB
SHVC-1JXM
SHVC-1JXB
etc..
Genesis is simple so there's barely any error potential at all.
Maybe they are too slow for SNES. (Disclaimer: I have no idea how fast the Mega Drive ROMs have to be.)
The 65816 requires memory to return a value within a half cycle, which is 200 ns (slow ROM) or 120 ns (fast ROM). But the 68000 accesses memory only once every four CPU cycles. Because of this, a Genesis cartridge can use much slower memory if it isn't going to be DMAing from ROM.
Ok, what makes zero sense is I can order some m27c160 chips today and get them in and they can be identical and work fine.
Like I have programmed snes games all last year using these chips and the same adapter boards and same games. Here is a picture, the 2 chips on the left programmed fine, the two chips on the right will not.
If you look at the two chips in the middle they are 100% identical as far as the "I assume batch numbers" one programs fine, they other will not. It is only for snes though. All 4 of those chips will program perfectly genesis games but only the 2 on the left program correctly for snes.
So going over this speed thing, could it possible that some of these chips for some reason have some how "slowed down" and as such they are not fast enough to program snes properly now?
It is not a soldering problem or a wiring problem. I don't even solder until they go through my tester board which I have used to test 100's of snes games with. This problem is chip specific that is 100% guaranteed. As far as the boards I use, I use them all it doesn't matter as long as the donor has 36 holes instead of 32. so boards like 1aon-20 or 1a3m-11 anything if they are lo rom. Any hi rom works as well as long as they have 36 holes which I don't think I have come across a 32 dip hi rom donor yet.
So I am leaning to 1 of 3 problems.
1. The speed of the chips like thefox and tepples mentioned has slowed down.
2. The programmer/software is buggy
3. The adapter board is buggy
The only one that makes sense is 1. As I can and do program many of these chips perfectly fine.
Is there a way I can test the speed of these chips to see if some of them are not as fast as they should be?
Attachment:
100_8069.JPG [ 2.96 MiB | Viewed 1629 times ]
...