Image Map Image Map
Page 2 of 3 FirstFirst 123 LastLast
Results 11 to 20 of 24

Thread: Circuit design questions

  1. #11

    Default

    Quote Originally Posted by Agent Orange View Post
    Hugo,

    Back in the late 70's while still on active duty, I had a Coleco Pong game. Would that be any relation to the Atari version.

    Tom
    When ARCADE pong came out in 1972 (the large cabinet version with about 65 TTL IC's) the design was quickly knocked off by Midway, Allied and others. Clone PCB's even appeared in the UK. Atari didn't patent it. So if your game was a large pcb with that many, or about that TTL's it would likely have been an Atari knock off in that era. Atari published the circuit, but no analysis of how it worked, so I wrote one and even got to interview the designer:

    http://worldphaco.com/uploads/LAWN_TENNIS.pdf

    Shortly after the entire design was put on a single IC, so it could be used with an RF modulator for home pong. But the design of the circuitry in this LSI chip was not as good as the Arcade version with the many TTL's, it had different sized paddles & scores and the range of ball motion wasn't as good. But it was good enough for home Pong, but those IC's have never really impressed me as much as the original design. This had 42 states of ball motion, but they were reduced a little due to a pcb design blunder on all the original Pong E Syzygy boards, that swapped the least significant bit of each player's paddle data.(It took over 40 years to discover this and alert the designer to it).It caused some interesting effects in game play. It is called "The Ghost in the Machine Bug" its explained in the article/link I posted. By the time Atari went to a 4 player "Pong Doubles" board, the bug was inadvertently eliminated in that version of the pcb.

    I'm very indebted to the original TTL design, because it inspired my interest in TTL logic, when I first saw the pcb as a teenager in the 1970's and I became determined to understand how it worked and build my own TTL games and other circuits. Mr Alcorn used most of the TTL techniques available at the time to realize the game. Hidden in there is a very unusual implementation of the XOR function.

  2. #12

    Default

    For unused inputs, I typically use 5K( actually 4.7K ) or 10K. The reason for using a higher value is that if there is a high load on an output, you don't want the input pin driving the power rail. It is OK the share a single resistor with many inputs but it should be wired close to that chip's +5v connections.
    For open collector, you need to both understand the amount of load, the maximum drive current and the speed required of that circuit. To low a value is just heating the board.
    Dwight

  3. #13
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    34,147
    Blog Entries
    18

    Default

    The problem with using a high-value pullup is that it does render the input more susceptible to noise pickup, so I aim for 2.2K (easily recognizable; red-red-red). This conforms to what I've seen in commercial designs as well.

    Open-collector/drain output pullups have to be taken on a case-by-case basis, particularly when driving long lines. There, you're dealing with transmission-line issues.
    Last edited by Chuck(G); March 26th, 2020 at 08:34 AM.

  4. #14

    Default

    Quote Originally Posted by Dwight Elvey View Post
    For unused inputs, I typically use 5K( actually 4.7K ) or 10K. The reason for using a higher value is that if there is a high load on an output, you don't want the input pin driving the power rail. It is OK the share a single resistor with many inputs but it should be wired close to that chip's +5v connections.

    Dwight
    One interesting thing pointed out by Horowitz & Hill (in their book The Art of Electronics, Cambridge University press), is that if you forget to wire up a cmos IC's ground pin, the chip then becomes powered by any of the input pins that are high. So all can seem to be working well, until all of the input pins happen to be logic 0 and the chip is unpowered !

  5. #15
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    34,147
    Blog Entries
    18

    Default

    The worse situation is where an unterminated CMOS input wanders somewhere between logic high and low levels. This was warned against with the 4000 series CMOS--there was a possibility that both transistors in a totem pole output could be set to conduct, creating a more-or-less short-circuit between Vdd and Vss. Outcome: tears and wailing!

  6. #16

    Default

    Quote Originally Posted by Chuck(G) View Post
    The worse situation is where an unterminated CMOS input wanders somewhere between logic high and low levels.
    This actually happens quite a lot with cmos failure modes, but the cause is that an output of a cmos gate driving an input goes totally open circuit. Even looking at the scope the technician can be fooled if that input is supposed to be logic low, because the scope probe discharges the floating gate. Later, with the scope not there the gate acquires charges and goes to logic high, perhaps disabling some circuit or causing some other effect. This sort of thing used to happen quite a lot on vintage Umatic and other VCR players that were largely 4000 cmos based, with control pcb's with dozens of 4000 cmos IC's prior to microprocessor based mechanism controllers.

    Of course quite a few people bias cmos inverter gates with a 1 meg resistor from input to output, into a Class A amplifier configuration and the output sits mid rail to make a quick and cheap high gain AC amp, so most of the time, the output devices are ok like that, in other words there is the equivalent of a bias gap that causes cross over distortion in the output stage.

  7. #17
    Join Date
    Jan 2007
    Location
    Pacific Northwest, USA
    Posts
    34,147
    Blog Entries
    18

    Default

    One frustrating thing that I've found with old 70s-era 4000-series CMOS devices is that if a device failure occurs, it's almost always (in my experience) with the PMOS totem-pole output transistor going short--I don't think I've ever seen a device fail because the NMOS transistor failed. In theory, there should be an equal likelihood of failure, but it must be some issue with the fab process and P-type CMOS.

  8. #18

    Default

    Quote Originally Posted by Hugo Holden View Post
    One interesting thing pointed out by Horowitz & Hill (in their book The Art of Electronics, Cambridge University press), is that if you forget to wire up a cmos IC's ground pin, the chip then becomes powered by any of the input pins that are high. So all can seem to be working well, until all of the input pins happen to be logic 0 and the chip is unpowered !
    I was working for a company called speed call ( before some reading this were born ) and we made touchtone decoders for taxi cabs ( and cement trucks ). I was probing one board for some failure when I came across a signal that was about 1/2 volt lower than the rail. Just like Hugo said. We'd been selling they boards for several years and they worked fine. There was never an operational combination that at least one pin wasn't pulled high.
    That was where I learned to put my finger in my mouth before testing to see if some component was hot. It took an almost permanent TO-03 can burn on the index finger before I learned proper order of events.
    Dwight

  9. #19
    Join Date
    Dec 2013
    Location
    Near Milwaukee Wisconsin
    Posts
    1,254

    Default

    This is an interesting topic. I have corrected (changed) a number of pull up situations. The machine operates the same, but maybe it will weed out some erratic situations later. By the way the machine is still running just fine. I do have a question regarding Open Collector, wire or'ing. I have a few places where I used wire or'ing. The question is how many pull up resistors are needed? One one each open collector? Or should there be only one pull up resistor be placed at the output connection of the wired or? Mike

  10. #20

    Default

    Quote Originally Posted by Mike_Z View Post
    This is an interesting topic. I have corrected (changed) a number of pull up situations. The machine operates the same, but maybe it will weed out some erratic situations later. By the way the machine is still running just fine. I do have a question regarding Open Collector, wire or'ing. I have a few places where I used wire or'ing. The question is how many pull up resistors are needed? One one each open collector? Or should there be only one pull up resistor be placed at the output connection of the wired or? Mike
    I really depends on the length between drivers and receivers. Technically it is a transmission line problem. Still, in most cases, it is just replacing the typical pull up of the the circuit designed logic type because the distances are small and we are just dealing with what a proper sized value would be. In those cases, look at what the typical pullup current of that family output of those devices are. Use a single resistor that would provide that much current at 1/2 the desired voltage swing ( like for ttl it is 0 to 3 volts so (5 - 1.5)/current. ). For short runs that works well enough for wire or'ing.
    For transmission line it is usually termination at the load. There would be a resistor at the destination of the characteristic impedance of the line. For most wiring, it is around 100 to 150 ohms. Coaxial cables are around 50 to 75 ohms. With transmission line it is all about reflections and edge rates. There are other types of termination such are source termination as well. Each has their purpose.
    Dwight

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
  •