Image Map Image Map
Page 1 of 2 12 LastLast
Results 1 to 10 of 11

Thread: Baud rate generator issue. What's safe in a TRS-80 Model 1 RS232 Card?

  1. #1

    Default Baud rate generator issue. What's safe in a TRS-80 Model 1 RS232 Card?

    Hi Guys,

    Does anyone know about baud rate generator ICs vintage late 70's/early 80s?

    Here is the issue. I have an RS232 card for a TRS-80 Model 1 that doesn't work (it's never worked). It has a faulty baud rate generator, which incidently gets very hot.

    The baud generator IC in the card is COM5016, COM5016 is also compatible with a BR1941 and a COM8116. However the TRS-80 RS232 schematic for this Model 1 board specifies a BR2941? The difference between a COM5016/BR1941/COM8116 and a BR2941 is with pin 10. In the BR 2941 spec sheet it says this is for -5v. The Model 1 RS232 card certainly has -5 volts on pin 10. In fact BR 2941 is cited as being a low cost version of BR1941/COM 5016/COM8116, but one which requires a -5V power supply (obviously on pin 10).

    However in the BR1941/COM 5016/COM8116 pin 10 is not -5 volts, rather it is listed as a 1/4 crystal/clock frequency reference output (I'm not sure what this means). As I said, this pin 10 difference is the only one between those sets of compatible ICs above and the BR 2941.

    Now we get to the TRS-80 Model III/4 RS232 card. Here BR1941 (and presumably COM5016 or COM8116) is cited as the correct baud generator to use. However, in the Model 4 technical reference manual there is a big fat warning to do with pin 10. it says of pin 10 - *INTERNALLY BONDED. DO NOT CONNECT ANYTHING TO THIS PIN. The chip diagram itself simply says "nc".

    My initial conclusion regarding my non-working Model 1 RS-232 board? The wrong IC (i.e. COM5016 instead of BR2941) was used as the Model I baud generator and the -5V delivered through pin 10 zapped the chip!

    Seemed logical to me BUT now I think I'm wrong. Why? Because I have seen THREE other Model 1 boards and they all have either COM 5016, BR1941 or COM8116???

    So I'm thinking (despite the warning in the Model 4 tech manual) that perhaps -5v is ok on pin 10 of COM 5016/BR1941/COM8116 after all?? Maybe they can be used in place of a BR2941 on the Model 1 board? Surely FOUR Model 1 boards can't ALL have the wrong IC?? Maybe the Model 4 RS232 circuitry is just wired differently hence the warning.

    Ok, stay with me. Nearly there. One final thing. How do I know my COM5016 is faulty? My Kaypro uses a compatible COM8116 baud generator to drive the keyboard. I swapped the suspect chip from the M1 RS-232 board into the Kaypro and the keyboard stopped working.

    Isn't the case closed then? Not quite. Despite the seeming abundance of BR1941/COM 5016 baud rate generators in Model 1 cards, I don't want to drop my Kaypro COM8116 in my Model 1 card as a substitute (to make sure the RS232 card now works) only to have it fried by -5v! I want to be absolutely sure it's safe.

    In summary...
    • A Model 1 RS232 card has a COM5016 (compatible with BR1941/COM8116) baud rate generator which is faulty
    • Three other Model 1 RS232 cards have this same IC or compatible
    • The circuit diagram specifies a different IC though, the BR2941, similar but with -5 volts on pin 10 (The card definitely IS delivering -5v on pin 10).
    • The model 4 tech manual warns against connecting anything to pin 10 of a COM5016/BR1941/COM8116 (the Model 1 RS232 manual doesn't mention these chips)
    My question.

    Despite the Model 4 tech manual warning, can COM5016/BR1941/COM8116 baud rate generators be used in the Model 1 RS232 board? In other words, does having -5V on pin 10 not matter to them? Their abundance in other Model 1 RS232 cards would suggest so. However, I don't want to risk zapping my borrowed Kaypro II one to test the M1 RS232 card unless I'm 100% sure.

    What do you think?

    Tez
    Last edited by tezza; October 26th, 2011 at 02:11 AM.
    ------------------------------------------------
    My vintage collection: https://classic-computers.org.nz/collection/
    My vintage activities blog: https://www.classic-computers.org.nz/blog/
    Twitter: @classiccomputNZ ; YouTube Videos: (click here)


  2. #2
    Join Date
    Sep 2011
    Location
    Melbourne, Australia
    Posts
    948

    Default

    Hi Tez, not sure why you didn't post this in the TRS-80 forum so it took me longer to find.... anyway turns out there are two types of M1 RS232, early and late, and the early one has 3 more components than the late one.

    My early one which dates to late 1978 is pale green in colour and has a BR2941 on it. To the right of U12 are 2 resistors (R10 and R11) and a capacitor (C9).

    My late one which is from 1981 or so is dark green in colour and has a BR1941 on it. To the right of U12 are just empty solder pads where R10, R11 and C9 would have lived.

    On the underside of both boards you can follow the traces of those three discrete components back to Pin 10 of U10 (the baud rate generator)... on the early board they are hooked up, on the late board they aren't.

    I bet your RS232 board has that components in place and are feeding +5V to Pin 10 of U10. Carefully unsolder one end of each of R10, R11 and C9 and power the RS232 back up and see if the BRG gets hot and report back.

    Cheers,

    Ian.

  3. #3

    Default

    Hiya Ian


    Quote Originally Posted by TRS-Ian View Post
    Hi Tez, not sure why you didn't post this in the TRS-80 forum so it took me longer to find.... .
    Yea, I did think about it but (1) I wasn't sure how many hard-core TRS-80 hardware accessory guys would be lurking there and (2) The baud generator chip is used in lots of vintage computers and I wanted to guage the opinon of hardware hackers as to whether or not -5V would be damaging on pin 10 of COM5016 and compatibles, when docs would indicated it wasn't designed for it.

    Quote Originally Posted by TRS-Ian View Post
    ...anyway turns out there are two types of M1 RS232, early and late, and the early one has 3 more components than the late one.

    My early one which dates to late 1978 is pale green in colour and has a BR2941 on it. To the right of U12 are 2 resistors (R10 and R11) and a capacitor (C9).

    My late one which is from 1981 or so is dark green in colour and has a BR1941 on it. To the right of U12 are just empty solder pads where R10, R11 and C9 would have lived.

    On the underside of both boards you can follow the traces of those three discrete components back to Pin 10 of U10 (the baud rate generator)... on the early board they are hooked up, on the late board they aren't. I bet your RS232 board has that components in place and are feeding +5V to Pin 10 of U10.
    Looks like you might have solved the mystery. The card is indeed an earlier one. It has those components and - 5V is indeed being delivered to pin 10!

    Quote Originally Posted by TRS-Ian View Post
    Carefully unsolder one end of each of R10, R11 and C9 and power the RS232 back up and see if the BRG gets hot and report back.
    I'm sure it won't but I know for a fact that COM5016 is damaged now anyway.

    Well..interesting! One of the photos of other boards I've seen has a heat sink on it. I wonder if again this was a case of the wrong chip in an earlier version of the card. The owner noticed the heat and added a heat sink, not realising the chip itself was incorrect? I suspect COM5016/BR1941/COM8116 baud generators are a lot more common than the -5V, low-powered BR2941s.

    Here is a couple of questions that you or others may or may not know the answer to.

    (1) Would -5V going through a COM5016/BR1941/COM8116 (when clearly it is not suppose to) kill the chip immediately or just shorten it's life through overheating? (I suspect the latter).

    (2) Would removing the -5V supply (i.e. unsoldering R10, R11 and C9 or alternately, just bending out pin 10 so it's not connecting) make it safe to use a COM5016/BR1941/COM8116 in this earlier card and would the card work ok?

    Much obliged for the help.

    Tez
    ------------------------------------------------
    My vintage collection: https://classic-computers.org.nz/collection/
    My vintage activities blog: https://www.classic-computers.org.nz/blog/
    Twitter: @classiccomputNZ ; YouTube Videos: (click here)


  4. #4
    Join Date
    Sep 2011
    Location
    Melbourne, Australia
    Posts
    948

    Default

    Hi Tez, I'm thinking in the best australian tradition, if the 1941 is already damaged it can't be hurt any further than it already is, so bend pin 10 out and see if it still works.

    Over my career in electronics I've (inadvertently) done some of the most abominable things to ICs and while some do die, the majority seem to handle it and continue working when you rectify the problem. Maybe you'll be lucky and the 1941 will be in that majority.

    Alternately put the 1941 from the m1 RS232 into your M3/4 RS232 board and run the usual RS232 tests (null model adaptor to another TRS-80 and stepping up through the baud rates usually does the trick).

    Looking at the design of the M1 Rs232 board I see some cleverness on Tandy's part... they produce them with the 1941 and leave out R10 R11, C9, but if the supplier runs out of 1941s and can only supply 2941s, then Tandy can continue production just by adding the three extra components.

    What do you think?

    Ian.

  5. #5

    Default

    Thanks Ian,

    I'm not worried about damaging the existing COM5016. I'm pretty sure that is toast as it didn't work when I tried it in the Kaypro. It will be easy enough to replace.

    What I don't want is to zap another one. I'm assuming it will be safe to plug a replacement compatible chip if I disconnect the -5V circuitry so I'll do that and try it out with my COM8116 from the Kaypro. The COM8116 is compatible with the COM5016. Worth a go I guess, so long as I don't damage anything.

    As it's the earlier board that has the B2941, my feeling is that Tandy initially designed the board to use this chip as (from the docs) it is a low power version of the B1941 and draws less current. This may have been a consideration in the early EIs where adding the board may have already stretched a fully committed IE PSU. Then they probably discovered that (particularly perhaps in newer versions of the IE) power issues were not critical so why not use the standard version of the chip. It probably cost less, was a lot more available and they could cut down the components required as they didn't need to supply -5v for it.

    In other words, save money.

    Tez
    ------------------------------------------------
    My vintage collection: https://classic-computers.org.nz/collection/
    My vintage activities blog: https://www.classic-computers.org.nz/blog/
    Twitter: @classiccomputNZ ; YouTube Videos: (click here)


  6. #6
    Join Date
    May 2006
    Location
    Hamilton, Ontario, Canada
    Posts
    3,748

    Default

    Well, when we had to replace the BRG on an early card with a later card version, we just cut the trace to pin 10.
    Legacy Computers and Parts

    Sales of, parts for, and repairs to, Vintage and Legacy computers.

  7. #7

    Default

    Quote Originally Posted by Druid6900 View Post
    Well, when we had to replace the BRG on an early card with a later card version, we just cut the trace to pin 10.
    And this was all that was needed? Ok, that's good to know.

    Tez
    ------------------------------------------------
    My vintage collection: https://classic-computers.org.nz/collection/
    My vintage activities blog: https://www.classic-computers.org.nz/blog/
    Twitter: @classiccomputNZ ; YouTube Videos: (click here)


  8. #8

    Default

    Problem now solved. As usual I wrote up the experience in a (repair)-blog article.

    Thanks to Ian and Druid for comments.

    Tez
    ------------------------------------------------
    My vintage collection: https://classic-computers.org.nz/collection/
    My vintage activities blog: https://www.classic-computers.org.nz/blog/
    Twitter: @classiccomputNZ ; YouTube Videos: (click here)


  9. #9
    Join Date
    Feb 2009
    Location
    Southern California, USA
    Posts
    2,495

    Default

    Tez,
    Nice write-up. It will help others with COM problems on early TRS machines.

  10. #10
    Join Date
    May 2006
    Location
    Hamilton, Ontario, Canada
    Posts
    3,748

    Default

    BTW, one of the biggest problems with the RS-232 board was heat from the E/I and the board itself warping the board and breaking the connection to the connector fingers.

    What I discovered was that, if you took a thin piece of lexon or plexiglass and cut a piece to cover the connector area, you could, using the board as a template, drill a couple of holes in the plastic piece where the screws went, place it on top of the board, drop the screws (with a lockwasher on each one) through the plastic and the board and into the connector holes.

    Then tighten it down enough to get a positive on the connector and framing test, then torque it down until the washers locked. I never had a Model I E/I come back with an intermittent RS-232 board connector problem again.
    Legacy Computers and Parts

    Sales of, parts for, and repairs to, Vintage and Legacy computers.

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
  •