mcranton@npe.isnet.inmos.co.uk (01/17/91)
In article <1991Jan9.184528.10447@uvm.edu>, pegram@kira.UUCP (Robert B. Pegram) writes: > From article <CMM.0.88.663383801.jkr@skakke.ifi.uio.no>, > by jkr@ifi.uio.no (Kristian Rosenvold): > >> Ok, here's a couple of unusual RAM upgrade problems... > >> The first one's a brand new STE,- this is the absolutely latest > >> The machines work fine with 2 SIMM modules inserted. Insert 4 >> modules (of any type) and there are several different reactions. > > For testing at least, use four of the *same type* - do not mix 256K > simms with 1Meg simms, the STe will get very confused otherwise. > >> Upon booting the STE language disk the "POOL FIX INSTALLED" appears >> onscreen. Then after 0.5 second EVERY SECOND character fades out. A similar >> effect may sometimes be observed on the Desktop. About 50% of all software >> works while the extra 2 modules are inserted. It may seem to me >> like the CPU recognizes the RAM but that the VIDEO CHIP has >> problems with it. This is only a guess, I haven't been in there >> with a machine code monitor to check. I've tried shielding the modules >> from the switchmode power supply, but it doesn't seem to help. I've >> compared the Motherboard with a previous release STE that handles the SIMMs > > I can't really help on this one, the missing resistors may well be the > problem. On regular STs memory problems can run from missing > resistors and capacitors to a bad MMU. Certainly the strange screen > problems are due to bad high ram. Speculating wildly, it might even > be only 1 bank of the high memory ram. > >> properly, and I have found no (obvious) differences. Resistors R405 & R459 >> (Behind SIMM sockets) are missing on both versions of the board. >> Anyone have any suggestions ? >> >> The second one's more conventional: >> A 520STFM, again the latest model. Just to be on the safe side I >> soldered in SOCKETS instead of soldering the RAM straight in. >> I have not installed the condensators, JUST the sockets. > > And so's the answer 8-). PUT IN THOSE CAPS! They are what kill the > switching noise from the RAM. You should also check for a bad MMU, > and for all the necessary resistors and caps for the second bank, > Atari doesn't always put them in. > > To add to the above I have also had a couple of machines with etching > faults on the pcb. After all you can't test whats not there ?? > >> The problem is that the machine doesn't work. I get various amounts >> of bombs,- the interesting part is that in the black part of the >> bombs I get random 'sprinkles' of colored pixels that come & go. >> (Kinda like multi-colored "snow" from the old IBMs) >> (They seem to be generated by the video circuitry because the do >> not remain on screen) >> Due to the width of my sockets I had to solder 4 of >> the RAM chips directy onto the motherboard (this has always worked fine >> earlier..) The symptoms are much the same regardless of whether the 12 >> remaining chips are >> installed or not. I have a feeling this problem is noise-related, >> although that may only be a part of the problem. I do not think >> this lies in the custom chip sockets, although there's always a >> possibility... > >> K. Rosenvold, jkr@ifi.uio.no / ...!{uunet,mcvax,sunic}!ifi.uio.no!jkr >> Short signatures R cute. > > Bob Pegram > > pegram@griffin.uvm.edu > or > uunet!uvm-gen!pegram