Magnum TTRAM 256MB boards for the TT 030

[majere]

The simms are:

128mb_simm.jpg

adquired in: http://cgi.ebay.co.uk/ws/eBayISAPI.dll?ViewItem&item=360311442827&ssPageName=STRK:MEWNX:IT

and the tos version is 3.0 deusch (3.06 deusch in labels¿?)

sys_info.jpg

The system boot is:

boot.jpg

(all the jumpers on)

[wongck]

What are the specs of memory needed for the adapter ?

Your SIMM is UG232C3674HKT-6 AHT 128MB 72p 60ns 24c 16x4 4K Parity FPM SIMM

[wongck]

Try another sysinfo program
http://www.atari-forum.com/viewtopic.php?f=74&t=10664

[dross]

Just got my board!

Unfortunately I've got the exact same issue as majere. With the PD1x and PD0x jumpers un-set, I get 64MB total. With them set I get 128MB total (reported with the strange number on boot, verified to be 128MB total in the OS).

I went with these:
http://www.memoryx.net/32mbx326e.html

Simeon, if you have a pair of 128MB SIMMs you're pretty sure will work let me know and I'll buy them off you. Otherwise, let me know if you have any ideas.

I've tried swapping the SIMMs into each other slots and also setting/unsetting the wait state jumper.

If I'm stuck with 128MB I guess it's not the end of the world, but it would be a bummer. =(

Also, how do I tell if I have the older TT and I need to set the wait state? I do see a daughterboard under the HD but I don't see the 68030 on it, or at least not on top.

Dave

[dross]

Ahh, found a super informative thread:
http://compgroups.net/comp.sys.atari.st ... ls-Trettin

I posted something in German a while ago here
http://atari-home.de/forum/index.php?bo ... 06;start=0
I talked to Uwe Schneider (the developer) about the problem of not
recognizing 128MB modules. Besides the symetrical issue there is a
problem with the coding of the SIMM sizes.
There are 2 different ways 128MB was coded at the time.
MagnumTT uses:
128MB PD0 and PD1 open.
Many SIMMS use PD0=0 PD1=1.
Just remove the PD1 0 Ohm resistor and it should work. It did for me.


About Niels Trettin, Niels Reimann, AtariGroup, trettintrade or
whatever his name might be right now:
Some people have good others have bad experiences with him.
He is a professional ebay-trader. So do not be surprised if you sell
something to AtariGroup for $5 and find it at ebay for $50 later
offered by trettin-trade.

Joe

I'd prefer not to clip a resistor off the board. Majere -- want to try it and let me know how it goes?
Otherwise I think we'll need to find a good source of "PD0 and PD1 open" SIMMs that meet all the other necessary criteria.

[dross]

This is probably the correct link to the German thread:
http://forum.atari-home.de/index.php?topic=906.0

[wongck]

MagnumTT uses:
128MB PD0 and PD1 open.
Many SIMMS use PD0=0 PD1=1.
Just remove the PD1 0 Ohm resistor and it should work. It did for me.

So it's either 128MB or nothing, once you removed it.

[dross]

That's no problem -- why would anyone want less RAM?

Which resistor specifically is it? I bet I could just clip one lead which would make it somewhat possible for me to re-attach if I screw it up. I'm still more inclined to try some different RAM first though.

[dross]

Ahh, circuit diagrams and jedec files are here:
http://www.uweschneider.de/Download.php?lang=en

I can't read that though... can anyone translate that to identify which resistor would be the one to snip off?

[dross]

I did it, I'm up and running with 256MB. Without harming my Magnum. =)

You need to just cover pin #68 (Presence Detect 2 shown in the diagram linked below) on the SIMMs. To do so, what I did was cut two little slivers of plastic from a silver anti-static bag. I folded the pieces in half and stuck them into the SIMM slot. You can cover a few extra pins on the right (69, 70, 71) without a problem, so my sliver of plastic is a bit wider than one pin just to make it easier to position within the slot. Just make sure not to cover pins to the _left_ though (pins 67 and lower).

Reference diagram:
http://www.interfacebus.com/Memory_Modu ... inOut.html

After inserting the first SIMM you'll have the opportunity to check out if the plastic stayed in the right place. If your technique is good, go for the second SIMM -- on that one you won't be able to check your work after the SIMM is inserted because a socket on the board blocks the view. If you do a good job, the end result will be solid as the SIMMs will hold the plastic firmly in place.

Upon boot you'll see the TT-RAM test takes quite some time due to the large memory size. The memory size won't be reported properly on the boot screen -- this is normal. Once you get into the OS though you can check the RAM and observe 256MB installed.

You'll want to leave all four left-most jumpers disconnected after doing this, as the Magnum manual shows. The wait state jumper should also be set appropriately (0ws seems to work fine for me).

[wongck]

That's no problem -- why would anyone want less RAM?

Yeap... the more the merrier..... just like money, girls and booze, but I guess RAM will be way way behind these in priority.

[wongck]

sorry double post.... can't delete....
removed the text...

[wongck]

After inserting the first SIMM you'll have the opportunity to check out if the plastic stayed in the right place. If your technique is good, go for the second SIMM -- on that one you won't be able to check your work after the SIMM is inserted because a socket on the board blocks the view. If you do a good job, the end result will be solid as the SIMMs will hold the plastic firmly in place.

Hey... I did something similar on the Intel CPU to overclock it back in the days of P3 or P4... can't remember which one.
I used nail varnish to isolate the pins. It worked for me then.

[dross]

Ack, looks like I spoke a little too soon. I'm getting system instability. Bumping the board down to 1 wait state to see if it goes away.

[dross]

Ok, things seem to be stable at 1ws.

I'd be curious to hear if anyone is operating at 0ws without any issues. Without removing any parts, it's not obvious where the 68030 lives exactly -- on a daughter board or otherwise. I wonder if it could be the RAM, not the TT itself.

[majere]

I did it, I'm up and running with 256MB. Without harming my Magnum. =)

You need to just cover pin #68 (Presence Detect 2 shown in the diagram linked below) on the SIMMs. To do so, what I did was cut two little slivers of plastic from a silver anti-static bag. I folded the pieces in half and stuck them into the SIMM slot. You can cover a few extra pins on the right (69, 70, 71) without a problem, so my sliver of plastic is a bit wider than one pin just to make it easier to position within the slot. Just make sure not to cover pins to the _left_ though (pins 67 and lower).

Reference diagram:
http://www.interfacebus.com/Memory_Modu ... inOut.html

After inserting the first SIMM you'll have the opportunity to check out if the plastic stayed in the right place. If your technique is good, go for the second SIMM -- on that one you won't be able to check your work after the SIMM is inserted because a socket on the board blocks the view. If you do a good job, the end result will be solid as the SIMMs will hold the plastic firmly in place.

Upon boot you'll see the TT-RAM test takes quite some time due to the large memory size. The memory size won't be reported properly on the boot screen -- this is normal. Once you get into the OS though you can check the RAM and observe 256MB installed.

You'll want to leave all four left-most jumpers disconnected after doing this, as the Magnum manual shows. The wait state jumper should also be set appropriately (0ws seems to work fine for me).

Great! Good job

Can you put image of this?

[dross]

There's not much to see... the little piece of anti-static plastic is so small you will just hardly see it.

I may still have issues though. I did some exhaustive memory testing last night and hit a lock up after many successful tests. It's possible that I would see the same thing even with my old memory board. Or it may be that the issue is so obscure I'll never see it during daily use. Time will tell...

[simbo]

now this is what i wanted to see

good info and problems solved

i know ram can be an issue and detection is a know problem

so ,.., let see
ill look at the datasheet of the simm and see what resistors to change on it

ok the pic at the top has no pd resistors fitted on this side you show

the pads are on the right of the pic just below the last ic on the right
the two sets of pads are unpopulated

perhaps the ones you need to adjust are on the other side

you need to update tos to 3.06 v3 is very bad with many bugs

[dross]

Ok, 256MB just turned out to be unstable. So I removed the small plastic bits I had inserted into the SIMM slots, swapped the SIMMs between the two slots, and set the jumpers back to 128MB. Now I've been running without a freeze for over a day under memory stress testing. Previously I was able to go maybe six hours max before a freeze.

Moral of the story: don't stick pieces of plastic in your SIMM slots. =)
At this point I'd be willing to buy some different SIMMS with a high likelyhood of working at 256MB, but I don't think I'll attempt to modify hardware.

Regarding tos 3.06 -- I'm using NetBSD exclusively as the OS (See http://netbsd-ataritt.wikia.com/) so I doubt that newer tos ROMs would help me out.

[simbo]

seems no movement not so good
however if anyone needs a spare board i have one as needed
no probs
but look to ram issue with size
and read all posts before order ram
you MUST make sure
even i get it wrong
however i just use a pair of 32
seems fine
but i also replaced the battery and shorted the RCLR pin to ground
of the nvram
now my wait state issue is gone.,..!!!
TT perfecto....!
seems the nvram stores lots of debug data for the chipset as well as its settings
replace the battery and short the nvram {power to machine off } to ground
then reboot etc

much like the falcon certain nvram settings can really screw up a machine

above all a failed battery does not mean all settings in nvram are gone

just frozen

so reset the nvram chip after replacing the cell

method

replace the cell and with power to the machine off
short the nvrams RCLR pin 21 or 22 usualy pin 21
to ground for 5 seconds

then power on and reset any nvram settings you need to change

then power off and reboot

i solved a wait state prob the cart showed by doing this
i had no cell connected before and had issues

before you change ANYTHING do this

[dross]

I've replaced the battery in the past and the machine was off when doing so. Are you saying that in order to clear the NVRAM it's necessary to short a certain pin on the motherboard?

What revision is your mainboard? Mine is a rev H. I'm not sure if it should be capable of running at 0ws or not.

In other news... I found that even with 1ws and 128MB, the board reliably caused system lockups after about 6 hours or so of heavy load. My old 16MB Atari board runs rocks solid, so for now I put it back in.

[simbo]

to reset most of these nvram there is a RCLR pin
usualy pin 21 or 22

yes
with a battery connected to it
and the machine off

short the rclr pin to ground for 5 seconds

resets the nvram 50byte block to 1

so,, factory set with a new battery

its daft i feel to leave stuff floating around in nvram

as the TT stores debug settings
i dont know what the 50 bytes do etc

so better to set them too 1

cured my issues with the chipset and the cart now reports fine.,..

[simbo]

ive seen 4 different nvram used in many TT

so depends what chip is there where the RCLR pin is

i think a key set addresses this pin

some people report key actions dont reset nvram when fitted to falcons
as they should

sometimes a hard reset is the only way with a fixed chip

i think there is a key action to do this in the TT

like ctrl etc..etc

advice??

[dross]

Ok, I'll have to try that tonight and report back.

What's "the cart?" A debug cartridge?

[simbo]

yes
however the key stroke will do the same thing

but you must have a charged full battery connected

see what number your nvram is
first

then work foward

slowly ...!

and slow down with precious TT