Minimig failed RAM upgrade

[Yod@]

Hiya,
Started my minimig RAM upgrade from 2Mb to 4Mb today, but it's not going as smoothly as I'd hoped.

I've added the first RAM chip as recommended, but when testing it displays garbage text as it loads the kickstart ROM and then hangs on a grey screen as if part-way through a reboot.

Any suggestions with regards to how I go about troubleshooting it?

I've done my best to look for poor solder joints or solder bridges but as far as I can tell it seems ok. Do I have to actually start resoldering all of the joints anyway?

I was also a little dubious about the length of Kynar wire used to connect pin 6 to the IO header, should I try a thicker or shorter wire?

BTW it looks as though pin 6 is connected to the pin below it in my picture, but I can assure you that I've bent it up to leave it disconnected.

Thanks for any help!

EDIT: I'm using this RAM as recommended in the official RAM expansion guide:
http://pl.mouser.com/Search/Refine.a...57EV30LL45ZSXI

[FOL]

You need to resolder it, you have legs that are'nt connected.
Test each leg with a multimeter, You will find some are'nt connected.

[Jope]

Another easy test is to take a drawing pin or a small knife and gently nudge all the pins sideways in turn. You'll notice that the loose ones will move with quite a gentle nudge, whereas the ones that are connected will be rigid.

Gently, mind you! Anything will move given enough force. :-)

[Yod@]

Thanks guys. I've already checked them with a multimeter but will try resoldering again. Will post back with results later!

[FOL]

Originally Posted by Jope

Another easy test is to take a drawing pin or a small knife and gently nudge all the pins sideways in turn. You'll notice that the loose ones will move with quite a gentle nudge, whereas the ones that are connected will be rigid.

Gently, mind you! Anything will move given enough force. :-)

Aye, but didnt want to suggest that, as we know about doing this. Someone who isnt experienced enough may push to hard and rip the legs off.

I try to think of the simplest solution, it will deffo be a leg not connected (if there are no legs short to each other), as I have come across this before, I have moded 30+ Minimigs.

[DDNI]

Hi yod@, great to see more NI folk on here! You know the user Tension? He is Bangor man too.

[Yod@]

Originally Posted by DDNI

Hi yod@, great to see more NI folk on here! You know the user Tension? He is Bangor man too.

I haven't come across Tension, no - but I'm glad to see there are others from NI on here

[Yod@]

Right, I've resoldered all the connections, and I'm still getting *exactly* the same garbaged picture

The F12 menu options still work perfectly, but I'm assuming that's just because the ARM board isn't reliant upon these RAM chips.

Tried to take some slightly clearer pics of my RAM chip, I swear I can't find any bridged pins or solder breaks. It doesn't seem to make any difference if I leave pin 6 of the upper chip disconnected or not, is that normal? Should the Minimig run as normal with pin 6 left floating?

Is it possible that one of my RAM chips is actually dead?

I swear, I'm at the point of removing BOTH chips and starting fresh...

[yaqube]

The garbage in the bootloader screen are most likely caused by some shorted pads or legs (address lines are suspected). Check with a multimeter any neighbouring legs pair.

Not connected legs (except power and chip enable) of the extra memory chip cannot cause corruption of the bootloader or kickstart screen.

You have put too many solder on the legs (IMHO). If it still fails remove all the solder and redo the chips.

[Yod@]

Originally Posted by yaqube

The garbage in the bootloader screen are most likely caused by some shorted pads or legs (address lines are suspected). Check with a multimeter any neighbouring legs pair.

Not connected legs (except power and chip enable) of the extra memory chip cannot cause corruption of the bootloader or kickstart screen.

You have put too many solder on the legs (IMHO). If it still fails remove all the solder and redo the chips.

Thanks yaqube, so that means I *must* have CE connected? After my original failure, I removed it when I was troubleshooting and it didn't make any different. Should I reconnect CE for further testing?

[yaqube]

Originally Posted by Yod@

so that means I *must* have CE connected? After my original failure, I removed it when I was troubleshooting and it didn't make any different. Should I reconnect CE for further testing?

The CE must always be connected. For troubleshooting connect it to 3.3V (pins 1-2 of the I/O connector) to disable the extra memory chip.

[Yod@]

You are all absolute geniuses.

I thought I glimpsed a tiny trace of solder hiding between pins 43 and 44, I used a sewing pin to scratch between them and then tested, it's all working perfectly!

It's taken me two days on and off to do just one RAM chip, I hope the second one doesn't take so long!

My sincere gratitude to you all.

[Yod@]

The second RAM chip went straight on, worked perfectly first time. Case closed.

Thanks again for the help, chaps.

[yaqube]

You can bypass R51 resistor and activate Fast SPI option to improve emulated hard disk performance and ADF file browsing speed.

[FOL]

Originally Posted by yaqube

You can bypass R51 resistor and activate Fast SPI option to improve emulated hard disk performance and ADF file browsing speed.

This only works though, if you have an ARM controller.

Just incase people using the normal pic (ARM Core), think it will speed things up.

[Yod@]

Originally Posted by yaqube

You can bypass R51 resistor and activate Fast SPI option to improve emulated hard disk performance and ADF file browsing speed.

I've just done this mod - ADF directory listing does seem to be slightly improved, I look forward to putting a hardfile image on to see how that works.

[cosmicfrog]

well done Yod