Image Map Image Map
Page 24 of 25 FirstFirst ... 14202122232425 LastLast
Results 231 to 240 of 247

Thread: I wish to create a new DMA/RAM expansion card for the Tandy 1000 line.

  1. #231
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    1,268

    Default

    I seriously don't get why it's giving you so much grief. The RAM portion has been working perfectly (other than the possibly nebulous results from Check-It 3) ever since I sorted the buffer mistake. I'm kind of tempted to build up another board and see if there's some kind of divine intervention that's allowing Patient Zero to function. (But, man, do I *not* want to solder another surface mount DS1315.)

  2. #232
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    1,268

    Default

    FWIW, I just performed some thread necromancy on a topic that I believe was quoted earlier in this thread. TL;DR, according to Topbench a RAM card based on this brain-dead SRAM design may be measurably faster than the original card for CPU/Memory bound operations.. Obviously there are probably holes in this statement one can drive a freight train through.

  3. #233
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    1,268

    Default

    Quote Originally Posted by blackepyon View Post
    Na, the 8237A was for my board later. I've got a few other projects I want to finish up first before I can get back to that (including TTL a monitor with a bad vertical circuit that I need to fix and put back together), but it looks like Eudimorphodon's board is ready to go... For him at least. Still trying to figure out why the heck his board isn't working for me. The logic chips TEST good, but still locking up after it detects 640K, regardless of whether I have the UMB pages enabled. Gonna try just swapping them out with fresh ones anyways just to be sure. If not, I'm gonna map the thing out end to end on my logic analyzer and try to figure out what's going on.
    I was thinking about this a little today, and just to sanity check: your HX still works with other RAM expansions, right?

    A theory that brewed in my head is I was wondering if it's possible there is some sort of conflict or corruption going on with the configuration EEPROM in your HX, given how the machine hangs after the 640k display but before booting? I don't know what precisely the HX stores in there, if it might retain some kind of data about installed cards or devices that *something* about this card results in a freakout?

    That seems incredibly unlikely given your success with the proto-board RAM experiment, obviously. Just a straw that as of yet remained ungrasped.

  4. #234
    Join Date
    Feb 2017
    Location
    Chilliwack, BC, Canada
    Posts
    377

    Default

    It works with the original DMA/RAM expansion board, and with the 3-chip solution we worked out.
    My vintage systems: Tandy 1000 HX, Tandy 1100FD, Tandy 1000 RSX, and some random Pentium in a Hewitt Rand chassis...

    Some people keep a classic car in their garage. Some people keep vintage computers. The latter hobby is cheaper, usually takes less space, and is less likely to lead to a fatal accident.

  5. #235
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    1,268

    Default

    So that almost certainly has nothing to do with it... I do have two stupid ideas.

    1: if you have the board fully occupied with ICs... Does it make any difference if you sever the connection between the c000 output from the 74138 (pull the pin out of the socket or whatever) and adding a pull up all the downstream circuitry? (I probably should have put a ROM disable jumper on it.)

    2: try a smaller value resistor on the E000 pull-up, or just tie it (temporarily) to 5v?

    Suggestion #1 is simply a crazy idea that it might be getting confused by something it's seeing on the bus when scanning for option ROMs, and #2 is to absolutely rule out we're stomping on the ROM DOS.

    (Edit: Actually, an idea might be run it without the 74138 installed and the three outputs C, D, and E000 pulled up. That leaves all the base 384k decoding and address inversion circuitry as is but should hard disable everything in the expansion space.)
    Last edited by Eudimorphodon; August 17th, 2019 at 09:38 PM.

  6. #236
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    1,268

    Default

    ... although, I definitely want to make clear that if you're sick of it and want to take a break to work on your stuff that's awesome, don't feel obligated to bang your head on this wall. I really appreciate you looking at it but it's not like you're being paid.

    I got an offer a few days ago for a spare HX motherboard to play with, if that pans out my plan is to try to construct a "HackinTandy HX" for bench testing. I am just flabbergasted trying to figure out what might be the problem. (I've kind of been slacking off from the circuit designing I should be doing because I'm having so much fun wasting my time trying old games I remember loving on my EX now that it has the RAM to handle them, and so far as I can tell it's working flawlessly.)

  7. #237
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    1,268

    Default

    Quote Originally Posted by blackepyon View Post
    It works with the original DMA/RAM expansion board, and with the 3-chip solution we worked out.
    This is another shot in the dark, but... I've been banging on KiCad today, and I noticed while cleaning up traces on a potential Rev 2 of the board that the MEMR trace takes a *really* torturous path around the pads of IC U3 (the 740 to get to the RAM/ROM sockets. (Kicad doesn't have an autorouter but it does have a "drag and choose a path" function that sometimes makes really poor decisions, and I guess I missed what a mess this was.) The board's working for me but maaaybe there could be a noise problem?

    (I fixed the trace so it now substantially cleaner and shares routing with the buffer direction fix.)

    In other news I took the plunge and ordered some cheap 16552 UARTs. (They're pulls so I hope they work.) I have a Kicad component symbol created for the part that looks correct and I think I've wrapped my head around what needs to be changed to make a 16552 work instead of a 16550. I think I'm going to temporarily can my ambition of getting the CF and Serial on the same boards; instead I'm just going to finish a brain-dead three chip layout to hold the 40 pin->CF sled, send that to China, and sketch up a stand-alone serial schematic. I'll save several headaches laying things out, can always try to merge them later.

  8. #238
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    1,268

    Default

    Layout finished for a CF-only daughtercard. (The "easy" part.)

    Tandy-CF-Port.jpg

    I'm really glad I double-checked everything before I got too deep into laying the traces; I discovered that one of the two styles of 74LS139 symbols in Kicad has the input labels for each multiplexer reversed. I suppose I should file a bug for that, because it very nearly scr3wed me. (Not that I can be sure the rest of the layout is going to work, but I want the mistakes to be my fault.) If I don't find any other boners in the final check it's off to PCBway.

  9. #239
    Join Date
    Feb 2017
    Location
    Chilliwack, BC, Canada
    Posts
    377

    Default

    I like the graphic

    I also want to know what the heck is going on. Sorry, I haven't gotten back to it, I've been busy the last few days. I'm also working on building a better breadboard rig (I'll post pics when I'm done that, and you'll see why) which will also include soldered 40-pin headers for my logic analyzer pods so I can more easily figure out what the bus is doing at any given time.

    Setting up an HX motherboard to play with would rule out for sure whether this is the HX that's doing it, and not just something on my end. At one point, over a decade ago, I saw an EX at one of my local thrift stores. I didn't pick it up, because I already had my HX and didn't see any point - a decision that I'm still kicking myself in the butt over, because it would be useful not just for testing, but because there are actually layout differences (as I was discussing with dJOS in his SmartWatch thread) between the EX and HX, and I've never had the opportunity to do a proper comparison between them.

    Since it looks like you got room on your board, I've got an idea for the CF card LED so you can can see CF card activity without any destructive case mods. I set up a '00 (an '02 or '05 work too, depending on what you've got free) as a set of inverters in series with each other, with the first input connected to the *DASP signal. First output is active high, second output is active low, and they're both at TTL voltage with enough current to properly drive a 20mA LED at full brightness, regardless of whether or not the CF sled has an LED onboard. Replace the power LED above the keyboard with a two-colour LED (common cathode), run a cable to the card, and have the green lead go to the second gate, the red to the first (with 470R resistors inline, of course). This will make the LED green when the CF card is disconnected or inactive (assuming you use gates with internal pullups, not "open collectors"), and red when there's activity. I tried just keeping the green active all the time, but the red was being washed out, so having it flip colours works better. I tested out on my own board, and I'm quite happy with the results.
    My vintage systems: Tandy 1000 HX, Tandy 1100FD, Tandy 1000 RSX, and some random Pentium in a Hewitt Rand chassis...

    Some people keep a classic car in their garage. Some people keep vintage computers. The latter hobby is cheaper, usually takes less space, and is less likely to lead to a fatal accident.

  10. #240
    Join Date
    May 2011
    Location
    Outer Mongolia
    Posts
    1,268

    Default

    Quote Originally Posted by blackepyon View Post
    I like the graphic
    Heh. Yeah, that's all of five minute's work there. Figured a card for these old dinosaurs deserved appropriate branding.

    Anyway, yeah, no problem you've been busy. I was a little worried I was coming across as pestering. Definitely keeping my fingers crossed that the HX motherboard thing works out, because if nothing else it would save a little wear on my EX's motherboard connector yanking boards on and off it if I had a test mule. I still just can't fathom what could be different between the machines other than the E-page being occupied, and the E-disable jumper *seems* to work fine on my machine, so... yeah. Gremlins. Definitely gremlins.

    The LED thing definitely sounds like a neat idea. When you say the inverters are in "series with each other" you've got it set up so you're driving the input of the second *and* the RED led with the inverted *DASP output, and essentially using the second inverter as a driver for the green one (instead of trying to drive it directly *and* the red inverter straight from the card? I don't know if I want to try to incorporate it into my rev 1 there, because I'm kind of sick of messing with the board. I realized the next day that I totally forgot there was that "gutter zone" you need to watch out for on the PLUS cards to make sure they don't interfere with the port area of cards above them, and thus I had to move the 74245 down about a quarter of an inch. Fixed version:

    Tandy-CF-Port.jpg

    I haven't sent it to PCBway yet, because, well... in a demonic fit of banging my head against the wall I completed a serial card layout last night:

    Tandy-Serial.jpg

    The board is set up to use a 16552. Channel 1 is hardwired to IRQ4, Channel 2 to IRQ3, but you can choose one each COM1/3 and COM 2/4 for the respective channels. I used the decoding from the serial portion of Sergey's Serial/FDC combo card, with the addition of a 7408 AND gate in front of CS on the 16552 so it's pulled low when *either* Com1 or Com2 is active. (My interpretation of the datasheet is CS is low for both serial ports, and CHSL high == channel 1 and CHSL low == channel2.)

    Kind of nervous because I feel like I'm a little in uncharted territory with the 16552. I'm surprised more people aren't using it because it seems to be explicitly designed to live on PC serial cards. I've double-checked everything three times now (getting burned by that bad KiCad symbol wasn't a confidence builder), I guess eventually I need to pull the trigger.

Tags for this Thread

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
  •