Loading ...
Sorry, an error occurred while loading the content.

Re: [midatlanticretro] Godbout Terminator Card

Expand Messages
  • Systems Glitch
    ... I ve been keeping a running search for S-100 terminators on eBay for a while, and I ve found that almost all of them go for insanely high prices (and 100%
    Message 1 of 15 , Oct 5, 2011
    • 0 Attachment
      > These are Active Terminators, which provide a stable source voltage, much
      > more reliable than the common passive terminators cards of the day - much
      > like in a mini or mainframe. Herb has an extensive, concise writeup on his
      > website that explains all of this. I only have the Imsai terminator. The
      > Godbout were desired when you have a backplaned loaded with a dozen or more
      > cards to maintain signal integrity. The ebay seller is obviously taking.
      > advantage of the fact that its a Godbout design.

      I've been keeping a running search for S-100 terminators on eBay for a while, and I've found that almost all of them go for insanely high prices (and 100% go for out-of-my-budget prices). I'm guessing that people with more money than technical resources have been lead to believe that noisy backplanes are the result of all of their intermittent S-100 problems (this was suggested to me when I started working on my Cromemco Z-2D, which uses an unterminated backplane...backplane noise was /not/ the cause).

      Bill, it's not regulators failing because there are none! There's an op-amp that runs off of the unregulated bipolar supplies (it doesn't matter to the op-amp, it's output doesn't need to get anywhere near the supply voltages except if something fails) and generally a pair of driver transistors and a pair of output transistors. My Morrow/Thinker Toys Wunderbuss's active terminator apparently had the output transistors die, which killed one of the drivers and burned out the op-amp (a LM301). Since you've got the full supply potential at your disposal, failure of anything in the power output section seems to result in catastrophic failure.

      The driver and output transistor arrangement are basically current boosters for the op-amp. I designed a similar circuit in college to drive low-impedance headphones from bipolar JFET op-amps, which were very nice op-amps but lacked much current drive. With such a simple follower circuit, you can slew > 100 V/uS even with slow(ish) bipolars like the 2N4401/2N4403.

      Thanks,
      Jonathan
    • B Degnan
      ... OK ... OK. I will run some tests on MARCH s IMSAI s terminator tonight if I have time. bd
      Message 2 of 15 , Oct 5, 2011
      • 0 Attachment
        >
        >Bill, it's not regulators failing because there are none! There's an
        >op-amp that runs off of the unregulated bipolar supplies (it doesn't
        >matter to the op-amp, it's output doesn't need to get anywhere near
        >the supply voltages except if something fails) and generally a pair
        >of driver transistors and a pair of output transistors. My
        >Morrow/Thinker Toys Wunderbuss's active terminator apparently had
        >the output transistors die, which killed one of the drivers and
        >burned out the op-amp (a LM301). Since you've got the full supply
        >potential at your disposal, failure of anything in the power output
        >section seems to result in catastrophic failure.


        OK

        >The driver and output transistor arrangement are basically current
        >boosters for the op-amp. I designed a similar circuit in college to
        >drive low-impedance headphones from bipolar JFET op-amps, which were
        >very nice op-amps but lacked much current drive. With such a simple
        >follower circuit, you can slew > 100 V/uS even with slow(ish)
        >bipolars like the 2N4401/2N4403.

        OK. I will run some tests on MARCH's IMSAI's terminator tonight if I
        have time.
        bd
      • Dan Roganti
        okay ! back to the workbench ... The money issue is separate from reality. And it helps to define what kind of noise when stating a problem - noise exists --
        Message 3 of 15 , Oct 6, 2011
        • 0 Attachment
          okay ! back to the workbench

          On Wed, Oct 5, 2011 at 6:31 PM, Systems Glitch <systems.glitch@...> wrote:

          I'm guessing that people with more money than technical resources have been lead to believe that noisy backplanes are the result of all of their intermittent S-100 problems (this was suggested to me when I started working on my Cromemco Z-2D, which uses an unterminated backplane...backplane noise was /not/ the cause).


          The money issue is separate from reality.

          And it helps to define what kind of noise when stating a problem - noise exists -- as in RFI, EMI, Ground bounce, crosstalk, Reflections -- in many domains, regions, spectrums, etc. 

          As for the backplane again, when your're only using a handful of cards - maybe 5 at most depending on the length of the backplane - remember electrons travel at approx 1nsec in 6inches on a FR4 copper pcb. Then you can, 99% of the time, get away without having a need for a Terminator card. Providing you install the cards within distance. 

          It all happens to be a matter of Signal Integrity again. When using a backplane which is as long as an Altair or Imsai, about 24inches long and just as many slots. And you have to use almost more than half of the slots to build your system, as we always did when there were only 4K, 8K or 16K memory cards -- besides all of the I/O cards. When the backplane is loaded with all of the capacitance contributed by each card. And you're trying to get one lone I/O card to drive the bus. All the while the signal is pinging back and forth along the backplane as if it were some ruby laser trying to reach inversion to get stimulated emmission. You begin to approach the mysterious world in the Black Art of Signal Integrity. Yes, this already happens with a bus as <relatively> slow bus speed at 2mhz - it's been around ever since the 1940's Relay Telephone crossbar switches.

          The incident wave of a logic signal propagating down the backplane is preferable to have arriving at the receiving logic gate intact before the reflected wave bouncing off the ends of the backplane returns back to the input of the same card and basterdizes the signal to the point where it creates metastability problems throughout the circuit card. With the increased capacitance of the bus by the additional circuit cards only magnifies the problem. As the incident wave is stretched, thus expending more energy over a longer period of time, and allowing the reflection wave to have more energy upon it's return -- as if it were a tsunami of electrons. 


           

          Bill, it's not regulators failing because there are none! There's an op-amp that runs off of the unregulated bipolar supplies (it doesn't matter to the op-amp, it's output doesn't need to get anywhere near the supply voltages except if something fails) and generally a pair of driver transistors and a pair of output transistors.


          ok a little lesson,
          The Terminator card happens to be a >Voltage Regulator< - albeit, a discrete component Linear voltage regulator  -- e.g. seperate resistors,caps, transistors, Opamps can be wired into a Voltage Regualtor-- which provides a Voltage Reference of 2.6 volts for every signal on the backplane. Please read the schematics.

          Bill, was correct in one way, as there is a voltage regulator IC on the Godbout Terminator card. This provides the first stage of a stable voltage source. But it's less than half of the entire circuit.

          As with the remainder of the circuit, the final output stage of the Voltage Regulator, the driver transistors are there for the sole purpose of sinking or sourcing the large amount of current to maintain proper level on the bus when dealing the extra burden of circuit capacitance -- more C requires more I -- from the dozen or more cards that are installed. With only 3 or 4 cards installed in a system, there's hardly any capacitance to worry about - as a result very little energy to cause signal reflections on the bus which can impede the signal on any input.


           
          My Morrow/Thinker Toys Wunderbuss's active terminator apparently had the output transistors die, which killed one of the drivers and burned out the op-amp (a LM301). Since you've got the full supply potential at your disposal, failure of anything in the power output section seems to result in catastrophic failure.

          The driver and output transistor arrangement are basically current boosters for the op-amp. I designed a similar circuit in college to drive low-impedance headphones from bipolar JFET op-amps, which were very nice op-amps but lacked much current drive. With such a simple follower circuit, you can slew > 100 V/uS even with slow(ish) bipolars like the 2N4401/2N4403.

          This has nothing to do with amplifiers - as stated, it's a Reference Voltage. 

          If you want to learn about SI, I would recommend a very good informative book, by one the leading Guru's, Dr. Howard Johnson -- and don't squeal at the math.

          =Dan
           
        • B. Degnan
          Thanks for taking a few to explain all this, I understand conceptually better now. You should teach this stuff. Bill Degnan ... been ... working ... noise ...
          Message 4 of 15 , Oct 6, 2011
          • 0 Attachment
            Thanks for taking a few to explain all this, I understand conceptually
            better now. You should teach this stuff.

            Bill Degnan

            -------- Original Message --------
            > From: "Dan Roganti" <ragooman@...>
            > Sent: Thursday, October 06, 2011 1:12 PM
            > To: midatlanticretro@yahoogroups.com
            > Subject: Re: [midatlanticretro] Godbout Terminator Card
            >
            > okay ! back to the workbench
            >
            > On Wed, Oct 5, 2011 at 6:31 PM, Systems Glitch
            <systems.glitch@...>wrote:
            >
            > >
            > > I'm guessing that people with more money than technical resources have
            been
            > > lead to believe that noisy backplanes are the result of all of their
            > > intermittent S-100 problems (this was suggested to me when I started
            working
            > > on my Cromemco Z-2D, which uses an unterminated backplane...backplane
            noise
            > > was /not/ the cause).
            > >
            >
            >
            > The money issue is separate from reality.
            >
            > And it helps to define what kind of noise when stating a problem - noise
            > exists -- as in RFI, EMI, Ground bounce, crosstalk, Reflections -- in
            many
            > domains, regions, spectrums, etc.
            >
            > As for the backplane again, when your're only using a handful of cards -
            > maybe 5 at most depending on the length of the backplane - remember
            > electrons travel at approx 1nsec in 6inches on a FR4 copper pcb. Then
            you
            > can, 99% of the time, get away without having a need for a Terminator
            card.
            > Providing you install the cards within distance.
            >
            > It all happens to be a matter of Signal Integrity again. When using a
            > backplane which is as long as an Altair or Imsai, about 24inches long
            and
            > just as many slots. And you have to use almost more than half of the
            slots
            > to build your system, as we always did when there were only 4K, 8K or
            16K
            > memory cards -- besides all of the I/O cards. When the backplane is
            loaded
            > with all of the capacitance contributed by each card. And you're trying
            to
            > get one lone I/O card to drive the bus. All the while the signal is
            pinging
            > back and forth along the backplane as if it were some ruby laser trying
            to
            > reach inversion to get stimulated emmission. You begin to approach the
            > mysterious world in the Black Art of Signal Integrity. Yes, this already
            > happens with a bus as <relatively> slow bus speed at 2mhz - it's been
            around
            > ever since the 1940's Relay Telephone crossbar switches.
            >
            > The incident wave of a logic signal propagating down the backplane is
            > preferable to have arriving at the receiving logic gate intact before
            the
            > reflected wave bouncing off the ends of the backplane returns back to
            the
            > input of the same card and basterdizes the signal to the point where it
            > creates metastability problems throughout the circuit card. With the
            > increased capacitance of the bus by the additional circuit cards only
            > magnifies the problem. As the incident wave is stretched, thus expending
            > more energy over a longer period of time, and allowing the reflection
            wave
            > to have more energy upon it's return -- as if it were a tsunami of
            > electrons.
            >
            >
            >
            >
            > >
            > > Bill, it's not regulators failing because there are none! There's an
            op-amp
            > > that runs off of the unregulated bipolar supplies (it doesn't matter to
            the
            > > op-amp, it's output doesn't need to get anywhere near the supply
            voltages
            > > except if something fails) and generally a pair of driver transistors
            and a
            > > pair of output transistors.
            >
            >
            >
            > ok a little lesson,
            > The Terminator card happens to be a >Voltage Regulator< - albeit, a
            > discrete component Linear voltage regulator -- e.g. seperate
            > resistors,caps, transistors, Opamps can be wired into a Voltage
            Regualtor--
            > which provides a Voltage Reference of 2.6 volts for every signal on the
            > backplane. Please read the schematics.
            >
            > Bill, was correct in one way, as there is a voltage regulator IC on the
            > Godbout Terminator card. This provides the first stage of a stable
            voltage
            > source. But it's less than half of the entire circuit.
            >
            > As with the remainder of the circuit, the final output stage of the
            Voltage
            > Regulator, the driver transistors are there for the sole purpose of
            sinking
            > or sourcing the large amount of current to maintain proper level on the
            bus
            > when dealing the extra burden of circuit capacitance -- more C requires
            more
            > I -- from the dozen or more cards that are installed. With only 3 or 4
            cards
            > installed in a system, there's hardly any capacitance to worry about - as
            a
            > result very little energy to cause signal reflections on the bus which
            can
            > impede the signal on any input.
            >
            >
            >
            >
            > > My Morrow/Thinker Toys Wunderbuss's active terminator apparently had
            the
            > > output transistors die, which killed one of the drivers and burned out
            the
            > > op-amp (a LM301). Since you've got the full supply potential at your
            > > disposal, failure of anything in the power output section seems to
            result in
            > > catastrophic failure.
            > >
            > > The driver and output transistor arrangement are basically current
            boosters
            > > for the op-amp. I designed a similar circuit in college to drive
            > > low-impedance headphones from bipolar JFET op-amps, which were very
            nice
            > > op-amps but lacked much current drive. With such a simple follower
            circuit,
            > > you can slew > 100 V/uS even with slow(ish) bipolars like the
            2N4401/2N4403.
            > >
            >
            > This has nothing to do with amplifiers - as stated, it's a Reference
            > Voltage.
            >
            > If you want to learn about SI, I would recommend a very good informative
            > book, by one the leading Guru's, Dr. Howard Johnson -- and don't squeal
            at
            > the math.
            >
            > =Dan
          • Systems Glitch
            Much better (and accurate) explanation than mine. Turns out there s a fair number of differences between the Godbout terminator card and the built-in
            Message 5 of 15 , Oct 6, 2011
            • 0 Attachment
              Much better (and accurate) explanation than mine. Turns out there's a fair number of differences between the Godbout terminator card and the built-in Wunderbuss terminator...the biggest being the bipolar op-amp supply in the Wunderbuss.

              To clarify, I didn't mean to suggest that terminators aren't needed or that noise isn't a problem, I meant to suggest that I think the high price of S-100 terminator cards might be due to people assuming their S-100 systems are flakey due to a lack of termination. Magical fix-all mojo makes anything sell high on eBay -- just look at carbon comp resistors.

              I do disagree about this having nothing to do with amplifiers though...it's got everything to do with them! The op-amp and its support transistors make active linear regulation possible. While the circuit in the active terminator may have little to do with /audio/ amplification in its application, the basic op-amp and driver/output transistor arrangement is just a unity-gain (or close to it) voltage follower. The circuit used in the Wunderbuss and Godbout termination is almost identical to the small power output stage I designed (I used diodes to bias the output transistor bases, and of course smaller output transistors). In both cases, an op-amp with an output stage is being used to drive a low impedance load at unity gain. What difference does it make if the input signal is a reference voltage or an AC signal?

              Thanks,
              Jonathan
            • Dan Roganti
              ... yes, there s no one fits all design. The Wunderbuss design is much better, I wish I had one. ... yep, eBay is not a good measure to determine what
              Message 6 of 15 , Oct 6, 2011
              • 0 Attachment


                On Thu, Oct 6, 2011 at 11:31 PM, Systems Glitch <systems.glitch@...> wrote:
                Much better (and accurate) explanation than mine. Turns out there's a fair number of differences between the Godbout terminator card and the built-in Wunderbuss terminator...the biggest being the bipolar op-amp supply in the Wunderbuss.

                yes, there's no one fits all design. The Wunderbuss design is much better, I wish I had one.
                 

                To clarify, I didn't mean to suggest that terminators aren't needed or that noise isn't a problem, I meant to suggest that I think the high price of S-100 terminator cards might be due to people assuming their S-100 systems are flakey due to a lack of termination. Magical fix-all mojo makes anything sell high on eBay -- just look at carbon comp resistors.


                yep, eBay is not a good measure to determine what solutions are needed.

                 

                I do disagree about this having nothing to do with amplifiers though...it's got everything to do with them! The op-amp and its support transistors make active linear regulation possible. While the circuit in the active terminator may have little to do with /audio/ amplification in its application, the basic op-amp and driver/output transistor arrangement is just a unity-gain (or close to it) voltage follower. The circuit used in the Wunderbuss and Godbout termination is almost identical to the small power output stage I designed (I used diodes to bias the output transistor bases, and of course smaller output transistors). In both cases, an op-amp with an output stage is being used to drive a low impedance load at unity gain. What difference does it make if the input signal is a reference voltage or an AC signal?


                I think you were using a Class B amplifier - correct ?  Which is what I'm familiar with in amplifiers - it's a common arrangement. That type of circuit entails having the preamp stage --in this the Opamp --  feed directly into the base of the pre-driver transistors in each half of the push-pull power transistor output stage. And the common node of the push-pull stage provides the output to the speaker - which then creates a sinusoidal signal -- or distortion if you prefer ;)
                In case of the Terminator card, the Opamp provides the voltage reference,  the predrivers are biased by the power supply only. And the output transistors which are biased by the predrivers, also have a feedback via the resistors on the common node of the push-pull output stage. This is how it can sense the current draw from the bus (via the termination resistors) -- the transistor is a current operated device, and mosfets are a voltage operated device. Whenever there are more Logic 1's or 0's across the bus -- it's either demanding more current or surging more current, it will then cause them to sink or source the current accordingly -- rather than letting all that energy reflect back down the backplane. One analogy might be a TTL buffer gate - a rather large fat ass gate - which can handle a few Amps of current, whether it's output is a Logic 1 or Logic 0 and it'ss capable of maintaining the voltage at a stable level.

                I hope this helps explain the difference.
                =Dan

              • B Degnan
                I have a copy of the schematic if you need it bd
                Message 7 of 15 , Oct 7, 2011
                • 0 Attachment
                  I have a copy of the schematic if you need it
                  bd


                  At 01:02 AM 10/7/2011, you wrote:




                  On Thu, Oct 6, 2011 at 11:31 PM, Systems Glitch <systems.glitch@... > wrote:
                  Much better (and accurate) explanation than mine. Turns out there's a fair number of differences between the Godbout terminator card and the built-in Wunderbuss terminator...the biggest being the bipolar op-amp supply in the Wunderbuss.


                  yes, there's no one fits all design. The Wunderbuss design is much better, I wish I had one.
                   

                  To clarify, I didn't mean to suggest that terminators aren't needed or that noise isn't a problem, I meant to suggest that I think the high price of S-100 terminator cards might be due to people assuming their S-100 systems are flakey due to a lack of termination. Magical fix-all mojo makes anything sell high on eBay -- just look at carbon comp resistors.



                  yep, eBay is not a good measure to determine what solutions are needed.

                   

                  I do disagree about this having nothing to do with amplifiers though...it's got everything to do with them! The op-amp and its support transistors make active linear regulation possible. While the circuit in the active terminator may have little to do with /audio/ amplification in its application, the basic op-amp and driver/output transistor arrangement is just a unity-gain (or close to it) voltage follower. The circuit used in the Wunderbuss and Godbout termination is almost identical to the small power output stage I designed (I used diodes to bias the output transistor bases, and of course smaller output transistors). In both cases, an op-amp with an output stage is being used to drive a low impedance load at unity gain. What difference does it make if the input signal is a reference voltage or an AC signal?


                  I think you were using a Class B amplifier - correct ?  Which is what I'm familiar with in amplifiers - it's a common arrangement. That type of circuit entails having the preamp stage --in this the Opamp --  feed directly into the base of the pre-driver transistors in each half of the push-pull power transistor output stage. And the common node of the push-pull stage provides the output to the speaker - which then creates a sinusoidal signal -- or distortion if you prefer ;)
                  In case of the Terminator card, the Opamp provides the voltage reference,  the predrivers are biased by the power supply only. And the output transistors which are biased by the predrivers, also have a feedback via the resistors on the common node of the push-pull output stage. This is how it can sense the current draw from the bus (via the termination resistors) -- the transistor is a current operated device, and mosfets are a voltage operated device. Whenever there are more Logic 1's or 0's across the bus -- it's either demanding more current or surging more current, it will then cause them to sink or source the current accordingly -- rather than letting all that energy reflect back down the backplane. One analogy might be a TTL buffer gate - a rather large fat ass gate - which can handle a few Amps of current, whether it's output is a Logic 1 or Logic 0 and it'ss capable of maintaining the voltage at a stable level.

                  I hope this helps explain the difference.
                  =Dan



                Your message has been successfully submitted and would be delivered to recipients shortly.