Close
0%
0%

STEbus backplane

STEbus backplane

Similar projects worth following
An essential part of an STEbus system.

Any fool can invent a bus, and many have. 

It isn't as simple as most people think, because as wires get longer they start behaving like transmission lines. Signals take time to travel, and bounce back from the end of wires. Wires act like antennae, transmitting radiation and picking it up. 

I've pointed this out to people using stripboard as a bus, and because they don't want to take it on board they get defensive and say "well it works well enough for me so far". Well yes, it will work okay right up to a point when it doesn't, and then waste endless hours figuring out what is wrong and trying to fix it. In a badly designed bus, it can be prohibitively expensive to redesign boards properly and scrap the old boards.

A colleague of mine worked for a company that took over Acorn's Acorn-bus business. Systems worked on an office desk or hobby den, with maybe a CPU and a disk controller. But they could never be made to work reliably anywhere doing serious work, like a factory.  Eventually they realised it wasn't robust and they moved to the STEbus.

Well-designed termination is essential to stop problems like reflections, ringing, overshoot, undershoot, electromagnetic transmission, picking up interference and so on. 

It isn't difficult to understand, or implement, so there are no excuses for bad backplane design.

STE_backplane_5.pdf

First version of the STEbus backplane. The LTV1117 regulator does not work as it can't sink any current. It shall be replaced by a LT1118 or a REG1118 (both expensive and hard to find). However these regulators have a different pinout. I am modifying the design to use the simple transistor based regulator found on Arcom backplane

Adobe Portable Document Format - 373.94 kB - 03/23/2021 at 18:36

Preview

  • PCB layout

    Keith03/24/2021 at 18:27 0 comments

    The connector side is almost entirely covered with copper, which is the ground plane for all bus signals.

    This is essential, because it ensures every signal has identical controlled impedance.

    The STEbus specification states:


    7.6.2 Characteristic Impedance

    The IEEE Std 1000 Bus is designed to take into account the driving requirements of high-performance transmission-line backplanes. The transmission-line system, together with the specified maximum signal length, allow an accurate determination of the time required for a signal to be correctly received.

    Backplanes should be designed using only microstrips for the signal lines, and should provide an unloaded characteristic impedance of 60 Ω ±10% including the effects of plated through holes and connectors.

    Backplane signal tracks should have a constant width throughout the length of the backplane so as to keep the same characteristic impedance throughout its length.

    Groundplanes are required so as to form a well-defined transmission-line environment. All groundplanes shall be continuous, allowing breaks in the groundplane only around the holes where connector pins must pass through. Under no circumstances should slot lines be allowed to exist in the groundplane, whether in the horizontal, vertical, or any other direction relative to the signal lines.


    Microstrip is explained here: https://en.wikipedia.org/wiki/Microstrip

    The formula for calculating microstrip impedance is given in

    https://www.analog.com/media/en/training-seminars/tutorials/MT-094.pdf

    and there are online calculators, e.g. here: http://pwcircuits.co.uk/microstrip-line-calculator/

    https://resources.altium.com/p/fr4 says the dielectric constant of FR4 material is 3.3-4.8 (depends on weave style, resin content, and material composition).

  • Power supply connections

    Keith03/24/2021 at 18:17 0 comments

    The STEbus specification allows up to 4 amps per board on the 5V rail.

    A fully-populated STEbus backplane with 21 boards could thus take up to 84 amps.

    The Arcom backplanes had many M4 threaded power terminals for attaching heavy-duty power cables.

    Although it is unlikely that anyone will need so much current, it is wise to make sure power cables are thick and low-resistance to reduce power rail ripple from transient high currents.

  • Termination voltage regulator

    Keith03/24/2021 at 18:03 0 comments

    The bus termination circuit from the specification is simple but the 2.8V regulator must also sink current when required. Common voltage regulators are only designed to source current. Examination of a 5-slot backplane reveals a working circuit.

    Current flows through R3 to drive the NPN transistor TR1. The voltage reaches 2.8 volts, then the junction of the feedback resistor chain (R1, R2) reaches 0.7 volts and activates TR3 to drain drive current from the base of TR1 . TR3 also activates the PNP transistor TR2, to sink current from the 2.8 volt rail. Diode D1 helps stop both transistors conducting at the same time.

    The whole circuit consumes about 20 mA, unloaded.

    The termination regulator circuit should be useful for other buses.

    Termination can be built on boards with DIN41612 connectors, but this adds the cost of connectors and occupies one or two slots.

    Termination is recommended at both ends of a long bus (e.g. 10 to 21 slots), but a short bus (e.g. 5 slots) will work safely with termination at one end.

View all 3 project logs

Enjoy this project?

Share

Discussions

Trevor R.H. Clarke wrote 04/20/2023 at 16:34 point

It's this project still active? Do you have a functioning backplane? If so can you post Gerbers or board layouts?

  Are you sure? yes | no

Keith wrote 04/20/2023 at 17:00 point

I have a couple of functional backplanes. They were laid out with coloured tape so no digital description exists. I would like to create one so they can be remade on demand. My time is very limited and I would need to learn a PCB CAD entry system. 

  Are you sure? yes | no

Jon Mayo wrote 04/12/2021 at 04:10 point

A bad bus creates two problems that requires a re-spin. The host and one or more devices have to be redone if there is a major redesign. Sometimes people get lucky and can modify something like S-100 or ISA for several generations and hack around their problems.
I was only vaguely aware of STEbus because of VMEbus, sp thanks for sharing this. It looks more straight forward than I initially assumed.

  Are you sure? yes | no

Keith wrote 04/12/2021 at 10:47 point

S100 is basically an 8080 bus with multiple kludges to cope with other CPUs and bus widths. A good example of why bus standards should be designed first. ISA is basically an 8088/86 bus, and nobody had to kludge it for other CPUs.

  Are you sure? yes | no

Keith wrote 04/12/2021 at 10:54 point

VMEbus is certainly powerful but big and complicated. A double-height Eurocard would often take about a third of its area in logic to handle the bus. STEbus is like VMEbus stripped down to a practical level for most control systems, which only need an 8-bit data bus peripherals, and not masses of high-bandwidth memory.

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates