1802 Membership Card, Engineering notes


Last updated Feb 1 2017. Edited by Herb Johnson, (c) Herb Johnson, except for content written by Lee Hart and others. Contact Herb at www.retrotechnology.com, an email address is on that page.

These are general engineering notes, or links to notes, for the 1802 Membership Card. Check specific version Web pages for notes specific to a version of the design. Many of these notes are taken from previous versions Web pages. See the the hardware notes page for other hardware notes. - Herb Johnson

General documentation, discussion, changes

Check the Membership Card Home Web page. There's a link there to the current version Web page, and links to pages on hardware, software, development, history and more.

See the the hardware notes page for links to other hardware notes.

See the list of links-to 1802 Membership Card revisions, and look at notes and links for each revision. Those will cover hardware changes and possible upgrades of previous revisions. The current production version Web page has similar information.

30 years ago, Lee developed an 1802 single board computer called BASYS. Look at the BASYS manual for hardware interface suggestions for the Membership Card.

Background on power and standby

standby mode and RAM

Rev H2 and earlier:

You don't have to remove the Front Panel board to go into standby (minimal power consumption, supercap saves RAM contents). Power connector P4 [for Rev H2 front-panel and earlier] has 4 pins:

pin 1 = positive supply (3-6v)
pin 2 = Standby/Run; tie to pin 1 to run, leave it open for standby
pin 3 = LEDs; tie to pin 4 to enable LEDs, leave it open to disable
pin 4 = negative supply (3-6v)

So, for minimum standby power, open the connections between 1-2 and 3-4 with a switch. Also: DC power at 1v is enough to maintain memory contents, but stops the clock and other logic from drawing power. - Lee Hart

Rev H2 standby w/ Rev I front panel

The Rev I front-panel has a 6-pin power, serial, and active-low /ON signal. Removing DC power and asserting /ON to ground puts the CPU board in idle and powers RAM and CPU with the supercap. Details on the rev I support page.

Lee Hart, Jan 31 2017: The base 1802MC (rev.H2 CPU + rev.I Front Panel, with Sony CXK58256 RAM) draws 15-20 uA in standby at 4v. It's hard to get steady readings because the supercapacitor takes *hours* to stabilize, and modern digital meters have significant voltage drops. The supercapacitor loses about 0.8v per day, and held the RAM contents for at least 2 days.

Above about 4v, the 1N5231B 5.1v zener diode [across the power input] begins to conduct, and at 5.0v they draw 5-25ma (depending on the part). This causes a pretty fast voltage drop from 5v to 4.5v. It gradually tapers off from there. There are better zeners; but not for 10 cents.

All versions; unexpected power consumption in manual LOAD mode

The 1802 has an undocumented "feature" in Load mode. After the first LOAD cycle (first DMA-IN cycle, triggered by pressing the IN button), the 1802 leaves the data bus floating. This allows the memory or I/O chips to put data on the bus.

But right after CLEAR, and before the first LOAD pulse, the 1802 *grounds* the data bus! This will result in high power consumption if any of the 8 data input switches are set to fight it. So measure your power consumption in LOAD *after* the first IN pulse.

The lowest power will generally be measured with the clock stopped (the RUN line low). But, that assumes the 1802 was stopped at a point where it is not driving the bus low, and thus fighting the memory or I/O logic.

- Lee Hart, Nov 5 2012 (and reported previously)

Engineering Data on current consumption

The COSMAC 1802 is a CMOS product, which can consume millwatts of power if run at lower voltage and clock speed. Almost no power is consumed when the processor clock is stopped; maximum power is consumed at fastest clock speed. It's the RAM which consumes much of the power; RAM output enable time is most of that. RAM enable varies by CPU version among the 1802 Membership Card designs. Other current consumers include the 5V Zener and LED's. Here's some notes about speed and current consumption, accumulated over several versions of the CPU design. Also check the Web support page for the CPU version of interest. - Herb Johnson

Current consumption of .3-wide SRAM by brand

Current consumption of .3-wide SRAM by brand for Rev G & H, was tested in 2015 and 2017. See this Tech Note for details.

Rev F & G current vs RAM and changes to RAM select, Feb 2014

1. Rev.F at 5v, 1.782 MHz, U2 = Hitachi HM62256 32k RAM, U8 = none, Q1 shortens U2 chip-select time to 1 clock cycle. Running program 3 in the manual (echo the switch settings in LEDs; all LEDs off): Icc = 4.22ma

2. Rev.G, same conditions: Icc = 7.40ma

3. Rev.G, same but U8 = .3-inch-wide Cypress CY7C199 32k RAM (not used by program) Icc = 7.41ma so not accessing U8 does not affect power

4. Rev.G, same as 3 but using U8 RAM Icc = 46.4ma so accessing U8 increases power about 6:1

5. same as 4, using WAIT to stop program at various points: Icc = 4.78ma if it stops when U8 is *not* chip selected, Icc = 28.31ma if it stops when U8 *is* chip-selected.

Conclusions:

- Power consumption increased slightly (with Hitachi HM62256 CMOS RAM) by chip-selecting U2 for an entire 8 clock cycles instead of just 1.

- Cypress CY7C199 increases power consumption about 6:1 when active. Has no effect when not active. It apparently draws a high "access" current (~100ma) when first chip-selected, which drops to a lower "hold" current (~25ma) when chip selected but all inputs are not changing. - Lee Hart

Sept 5 2010: Rev B Power consumption vs clock speed

Current consumption running a program with the pot at min/max frequencies, with a Hitachi HM62256P-12 32k byte RAM. Current declines as speed declines.

VCC    ICC at clock(min)    ICC at clock(max)
---    ----------------    ----------------
3v    .12ma    13.5 KHz    .39ma    280 KHz
4v    .25ma    9.24 KHz    .71ma    350 KHz
5v    .45ma    7.45 KHz    1.06ma    341 KHz

[with power connector removed and processor idle,] I've had it save a program for 10 minutes [by the supercapacitor].

Zener diode D11 and excessive DC supply voltage

The part supplied is a 1N5231B 5.1v 5% 500mw zener diode. It's fine for general use, and does a good job of holding the max supply voltage under 5.5v to protect the supercap C5.

But it draws more supply current. I also got some low-power 1N4625 5.1v 250mw zeners. They draw about 1/10th the current at any given voltage. This is good for micropower applications, but a higher voltage supply can easily overwhelm its ability to keep C5 under 5.5v. Here are the supply currents of these two zeners (average of 5 parts of each):

VCC =     3v        4v     5v     5.5v    6v
-----     --        --     --     ----    --
1N5231    1.4ua    26ua    6ma    44ma    86ma
1N4625    0.35ua    8ua    .12ua  1.8ma   30ma

LEDs and current vs supply voltage

I've tried several different colors of LEDs. Whites cost the most, and need the most voltage to work; but are the brightest for a given current. Reds are the opposite; cheaper, lower brightness, lowest operating voltage. I'm supplying red LEDs and R11=1k with the kits. Here is the additional supply current per LED, and perceived brightnesses at different VCC voltages for the two extremes. - Lee Hart

         R11    2v    3v    4v    5v
         ---    --    --    --    --
white   3.3k    0ma     .12ma    .38ma     .68ma
                off    dim      normal    bright
          1k    0ma     .27ma   1.1ma     2.05ma
                off    normal   bright    very bright
red     3.3k    .13ma   .43ma    .73ma    1.04ma
              very dim  dim     dim       normal
          1k    .39ma  1.35ma   2.33ma    3.31ma
                dim     normal  bright    very bright 

My Membership Card has clear (instead of red) LEDs. Those are blue LEDs. I also changed the 1k series resistor network to 3.3k, as they are *way* too bright at 1k. :-) - Lee Hart, Mar 2012

Additional note: blue LEDs with a 3.3K resistor and 4.5V power, draw about 1ma each, and are very bright. - Herb Johnson.

Engineering Q&A notes

modern RAM, NVRAM, etc.

Q: ....consider trying a NVRAM from Simtek (now Cypress Semiconductor) STK12C68 Series 64K-bit (8K-bit x 8) 5 V 300 mil DIP-28 AutoStore NVSRAM 45 ns. Or, a Dallas Semiconductor DS1225 [with mechanical mods].

A: The Membership Card has a 600-mil wide memory socket, but an adapter board might hold a 300-mil chip.... The DS1225 has a coin cell stuck on top of it, which adds to the height, and there is none to spare. An adapter board could accept modern surface mount chips, one RAM and one flash ROM. This would give you non-volatile storage with cheap modern chips. I didn't do this as I was deliberately restricting myself to "old school" chips and construction techniques.

One other point: Consider that an ordinary alkaline AA cell will deliver 250uA for over a year. That's enough to *run* the Membership Card! Are you sure you need nonvolatility longer than that? - Lee Hart, Oct 11 2010 on cosmacelf

data available signal

Q: Is there a way to signal external hardware when data is available at the output port?

A: The Q line can be used. Or use one of the 8 output bits as a strobe. [TPB] could be done, [or] use INT or EF1-EF3, for example. - Lee Hart, Oct 23 2010 on cosmacelf


This page and edited content is copyright Herb Johnson (c) 2017. Contact Herb at www.retrotechnology.com, an email address is available on that page..