Image Map Image Map
Page 4 of 5 FirstFirst 12345 LastLast
Results 31 to 40 of 47

Thread: Exidy Sorcerer error

  1. #31

    Default

    When I got my Sorcerer II it was reporting strange RAM amounts (said the stack started at FFFF). It turned out not to be a RAM issue but rather a bad capacitor. You might want to check all those caps.

  2. #32
    Join Date
    May 2018
    Location
    Melbourne, Australia
    Posts
    82

    Default

    Quote Originally Posted by whizzi View Post


    Nope.
    Well now we're getting somewhere.
    That is definitive.
    You can't write to those RAM locations (and may not necessarily really be reading them either).
    However you must have some working RAM towards the top of the address space for the MWA (Monitor Work Area) or you wouldn't even be able to enter monitor commands into the buffer to be parsed.

    I will have to study the schematics over the weekend to see what could causing these symptoms (address decoding?) and develop some tests.
    Tempests idea to check the caps is a good one.

  3. #33

    Default

    This won't help Whizzi but this is the point I break out either a full set of replacement RAM chips or my trusty 4116 RAM tester that I made out of a beat-up ZX Spectrum board. Rule the RAM out once and for all. I do wonder why half the RAM decouplers are tantalums and the other half look like film caps.
    www.binarydinosaurs.co.uk - UK home computer history
    Where RIFA capacitors come to die
    facebook.com/binarydinosaurs

  4. #34

    Default

    What I noticed is that every first byte of a page of memory is 00 and the rest is FF, maybe that can give a clue what is happening.
    Like: addresslines 0-7 at logic '0' give '00' at the data bus or such.

  5. #35

    Default

    I don't know if this is normal (it's a video), but I made it a few days ago.



    Fact is, the machine responds exactly the same as with memory..
    I have a few computers and only have 612m2 for computers.

  6. #36

    Default

    OK, how is it doing THAT with no RAM on board? The only other RAM chips I can see in the schematic are the 6x 2114 for screen RAM. I don't suppose the pair of 74LS157 address decoders (8B, 8C) and the 74LS241 driver (4D) are socketed are they?
    www.binarydinosaurs.co.uk - UK home computer history
    Where RIFA capacitors come to die
    facebook.com/binarydinosaurs

  7. #37
    Join Date
    May 2018
    Location
    Melbourne, Australia
    Posts
    82

    Default

    Quote Originally Posted by Witchy View Post
    OK, how is it doing THAT with no RAM on board?
    Is there a RAM board in your S-100 expansion unit?

  8. #38
    Join Date
    May 2018
    Location
    Melbourne, Australia
    Posts
    82

    Default

    Quote Originally Posted by Witchy View Post
    OK, how is it doing THAT with no RAM on board? The only other RAM chips I can see in the schematic are the 6x 2114 for screen RAM.
    It is perplexing. Could the design really be that elegant that in the complete absence of processor RAM the system is able to run by employing the screen RAM.?
    While pondering I came across this post from jonb on an Acorn forum in which he speculates that "When there's no DRAM, the monitor uses the screen ram for its stack". https://stardot.org.uk/forums/viewto...p?f=45&t=13868

    Looking at the memory map in the Technical Manual and Monitor source code in the Software Manual shows the during a cold start the cold start routine begins at the bottom of the memory map 0000 and works its way upwards testing for the top of RAM.

    So what is reported in this screen shot below could literally be true.



    The very top of RAM is not 'screen RAM' per se as one usually thinks of it but is the user definable graphics RAM (as mentioned by Witchy at #11 above). This explains why you can make use of the inbuilt monitor commands but can't enter and execute a short arbitrary machine language program in RAM proper.

    Presumably when you "GO BC00" the BIOS in the controller PROM's dutifully starts trying to fill (non-existent) RAM from the CP/M boot tracks and hilarity ensues.

    In the stardot post jonb also mentions, like tempest, that he removed a shorted tantalum capacitor to get his RAM working.

  9. #39

    Default

    well, I just tried and well, I think most of the tantalums are shorting :/

    As well as these blue things, I assume they roughly to the same.
    short video

    So, I'm gonna order a whole bunch of them and just replace them all, which will probably take a while. Not sure if I can remove the really broken ones and try it then? It'd be nice if I can at least a little memory to work with for now..
    I have a few computers and only have 612m2 for computers.

  10. #40

    Default

    whizzi,
    By measuring across the caps that way you also have several parallel paths for each of the other RAM IC's, and anything else
    on the +5VDC bus.

    Here is a quick snapshot of the TRS-80 4P's RAM, with the second bank on the right. Notice that all CAPS for RAM are in parallel,
    along with numerous others, which are all over the motherboard.

    RAMCaps.jpg

    I'd suggest a better way to locate your short or current path would be to inject a low current by using 2 "D" cells in series with two
    3.9 OHM Series resistors. That should get you 3 VDC at ~380ma to use as current path to trace where the current is going on your
    motherboard. You should be able to take a Voltmeter and set it to millivolts and just trace along the 5VDC circuit trace and locate
    the lowest voltage along that path. It the voltage doesn't change, there is no current flowing on that PCB trace. Select a different one.
    That should get you to the problem quicker than the shotgun approach of mass replacement. If you remove all the RAM IC's it would
    even be a better test. Then you could use another method to test the RAM. I'll PM you the information on my HDRT.


    Larry

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •