An H8 Badge for VCFMW

[Lee Hart]

Let's start a new thread, as we were wandering off-topic in the VCF-East thread.

I'm thinking of making a wearable "badge", intended as an inexpensive kit for people to build at VCFMW in September. It would introduce people to the Heathkit world of kit-building. Here's a "straw man" description.

- 80C85 CPU (since this is a *vintage* computer show)

- 32k RAM at 8-40k

- 32k EPROM, 0-8k with PAM8 (or hex version?), 40-64k with Benton Harbor BASIC

- bit-banged serial port using the 80C85's SID and SOD pins.

- 4x4 keypad using tactile switches

- nine 0.3" 7-segment LED displays.

- a PCB styled like a half-size H8 front panel (about 6"w x 3"h)

- powered by USB-serial adapter, or battery operated with three AAA cells.

- a manual to imitate the Heathkit manuals (probably online to download).

- target parts cost around $20.

I've been scribbling circuits, and think I can do it with about six chips.

- 80C85. OKI seems to be the best source of these on ebay.

- 27C256 EPROM (cheap and easy to get).

- CY7C199 32k "skinny" RAM (because I use them in my other kits).

- 74HC151 address decoder (a 1-chip way to get the funny address map).

- 82C55 to provide three 8-bit I/O port (LED anodes, LED cathodes, read keypad).

- 74LS145 to select LED cathodes and keypad rows.

PAM-8 and BHBASIC would have to patched to run in this minimalist H8 environment, as the keyboard and display mapping won't be the same. Likewise for the bit-banged serial port. I'm guessing BASIC could be stored in high EPROM, and block-moved into RAM to execute. There should be lots of room in high EPROM for other programs as well (like the iconic "your H8 is up and running" demo).

This is just my "brainstorming" so far. I'm not a software guy, so I'll certainly need help to pull this off. Can you help, or think of any improvements? Please let me know what you think!

Lee Hart

Virus-free.www.avast.com

[Joseph Travis]

It's very likely that only the PAM8 monitor ROM will have to be modified for I/O as BHBASIC would be doing it's console and keypad I/O through the ROM.  We'll have to verify that though.  Seems to me there are a couple qwerks about BHBASIC when it starts up... I'll look into it.

Joe

--
You received this message because you are subscribed to the Google Groups "SEBHC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sebhc+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAF__8hd%2BHYa23X8u%3DetJKCFzCZvR_rN8%3DTCAM%2BYj_0fWrExsbQ%40mail.gmail.com.

[Douglas Miller]

I'll throw in my "free" software advice. The 8085 SID/SOD as serial does not make a good console. The challenges of receiving characters asynchronously (unexpectedly, as required for typing) requires near-constant CPU monitoring of SID. If it is only used for output, say a printer, that is easier to manage (XON/XOFF handshake is  still a challenge).

[Randy E (OoMOR)]

Lee/All,

Would these badge kits be available to those of us that can't make it to VCFMW?

Thanks, Randy

[Lee Hart]

Doug: I agree that SID by itself is not a responsive serial input. The program would have to be looping on a RIM instruction to keep reading SID to detect a START bit.

However, the serial input can also be wired to trigger an interrupt. Josh Bensadon did this on the software for my Z80 Membership Card and 8080  Altaid kits. The program can then be doing anything. When a serial character comes in, its START bit sets SID low *and* triggers an interrupt. The interrupt handler reads the bits on SID (ignoring further interrupts). When finished, the interrupt handler puts the character in a buffer, enables interrupts, and returns to the main program.

Randy: Yes, of course they would be available to anyone that wants one. Like Norberto etc. do with their boards, I expect that we'd try to get some idea of how many PCBs and parts to order. They'd get sold on a first-come first-serve basis. If there was enough interest, we could make additional batches.

Virus-free.www.avast.com

--

You received this message because you are subscribed to the Google Groups "SEBHC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sebhc+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sebhc/7ce72d58-2d8f-4654-939d-42701944d88dn%40googlegroups.com.

[Douglas Miller]

I think the results will still be unsatisfactory, especially in an H8 environment where you have interrupts occurring frequently (2mS clock) and that could delay the response to the START bit resulting in framing errors. Plus, you can't be sending characters and still receive characters at the same time (full duplex not supported, due to 100% CPU consumption required each for sending and receiving), further degrading the experience. I urge caution against expecting SID/SOD to work well as a general purpose serial port.

[Lee Hart]

Doug, take a look at Josh Bensadon's code for my 4 MHz Z80 Membership Card. It's at http://sunrise-ev.com/photos/z80/ZMCv15.zip. He implemented a full duplex interrupt-driven bit-banger 9600 baud serial interface, while also running a 2 msec clock interrupt. It can send and receive at the same time without missing characters.

The serial input interrupt has the highest priority. When a START bit arrives, it stops the 2 msec clock, and reads the serial input character. If a serial output character is already in progress when the START bit arrives, it finishes sending it while reading the serial input character. It works quite well as a console for running CP/M.

Josh wrote similar code for my Altaid 8800 with its 2 MHz 8080 CPU. It's at http://sunrise-ev.com/photos/8080/altaid05.asm. It also has a 2 msec clock and serial input interupts. The code is a little less ambitious due to the lower performance of the 8080, but it still works well enough for running CP/M.

Given that we're talking about a "toy" H8 and not the real thing, I think it can work well enough. At least I'm willing to give it a try. :-)

Virus-free.www.avast.com

--

You received this message because you are subscribed to the Google Groups "SEBHC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sebhc+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sebhc/4b888c16-6215-49eb-b84f-ef3a54a3c607n%40googlegroups.com.

[Douglas Miller]

Hmm, the ZMCV15 code uses an 8250 UART, but does bit-bang an SPI port for SDCard access (not the same thing). It does do some bit-banging using a spare bit on the 7-segment display, but that leverages the display refresh code and hardware. The altaid05 code looks to be half-duplex. I have my doubts, but best of luck.

[Joseph Travis]

I remember Josh's work on bit banging the serial I/O for the Z80 Membership Card and believe the serial I/O via SID/SOD will work fine but, you'll probably want to select the CPU XTAL frequency to accommodate the various baud rates.  I also started thinking about the "your H8 iS UP And running" demo and believe it uses direct I/O with the front panel and would have to be rewritten for the H8 Badge.  I still have to investigate using BH BASIC (tape version).

Joe

On Sat, May 2, 2026 at 3:17 PM Douglas Miller <durga...@gmail.com> wrote:

Hmm, the ZMCV15 code uses an 8250 UART, but does bit-bang an SPI port for SDCard access (not the same thing). It does do some bit-banging using a spare bit on the 7-segment display, but that leverages the display refresh code and hardware. The altaid05 code looks to be half-duplex. I have my doubts, but best of luck.

--

You received this message because you are subscribed to the Google Groups "SEBHC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sebhc+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sebhc/6a2bc767-bff6-4679-9d6e-dad0cd6157d6n%40googlegroups.com.

[Bob Bell]

I have been following this thread with interest; I always wanted an H8, but could not afford one.

IMO, you'll bypass the headache with SID/SOD for serial I/O by just including a hardware U/ART, such as the TR1602 or similar.

I used this U/ART on my SERCON board a few years ago. No initialization needed.  It just works the way you set it's configuration pins.

Bob Bell

[Douglas Miller]

Seems to me that a very simple H8-like machine would be nice. Display and keypad to simulate the front panel, minimal RAM and ROM, no HDOS or CP/M support. The 8085 SID/SOD could be implemented as a download/upload channel, something like the cassette interface or to load/save Intel HEX format programs. An expansion bus then to add additional hardware in the future. Connecting the SID/SOD to a USB-Serial dongle (also supplying power) would then allow connecting to a PC for download/upload. I'm not sure how BHBasic would work over that serial connection, but with an expansion bus you could add a regular UART. Of course, the SID/SOD interface would never work with existing software that expects an 8250 UART.

[Lee Hart]

Here's a (preliminary) schematic of what I have in mind. Eight ICs on a 3" x 6" PCB. I tried, but couldn't get it any less.

- 2 MHz 80C85, 32k RAM, 32k EPROM, add. latch, addr.decoder, 82C55 I/O ports, digit select, and 4053 for "glue" logic.

- Memory map has 0-2k and 40-64k ROM, 2k-40k RAM.

- RAM is battery backed up.

- The 9-digit display is similar to the H8 (same segment and digit assignments).

- I kept the RUN and PWR LEDs; the rest are not relevant on an 80C85.

- The keyboard setup is quite different. It's multiplexed with the display, but the same 0 + RST keys will reset the 8085.

- I used an 82C55 to get three parallel ports for the I/O.

- SID and SOD for a serial port. SID is tied to RST7.5 for fast START bit detection.

 There are a few "Hartian" design choices to save parts that bear explaining.

I used a 4053 triple SPDT analog switch to implement the needed "glue" logic. U8A is wired to chip-select the 82C55 when IO/M=1 and A7=0. U8B copies the RST-/ key to be in series with the 0-key (so both keys have to be pressed to Reset). U8C is wired as an inverter, so the high "idle" state of a TTL serial input sets interrupt RST7.5 low.

The "beeper" is just a piezo disk on an output bit. Software will have to toggle it high/low to make a sound. This should be possible if it's inverted each time the next LED row is selected every 2 mSec.

The "timer" is just an RC network using the RST5.5 interrupt. When the next LED row is selected, also set PB6 high. C5 is charged by R8 until RST5.5 reaches its trigger voltage. The interrupt sets PB6 low (to discharge C5), then selects the next LED row and sets PB6 high again. 

The battery backup diodes switch the ground end of the RAM. That's so when VCC goes to zero, the RAM's VCC=0 and its VSS is the negative battery voltage. That makes all the RAM's inputs (especially /WE) all at VCC (i.e. "high").

I'm not planning to run HDOS or CP/M on it. Just the monitor, (hopefully) BHBASIC, and maybe other programs from the cassette days. There's lots of room in the ROM to store various demo programs that can be block-moved from high ROM into RAM to run. There are four unused I/O bits, which could conceivably be used to bank-switch a larger RAM or ROM.

The interrupt setup is completely different, but the source for PAM8 is available so it could be modified to run. BHBASIC is more of a problem; it's probably doing direct I/O and I don't think the source is available.

Parts cost looks like about $20-$25, and I think it will all fit on a 3" x 6" PCB.See anything wrong?

See anything wrong? Or anything that could be improved?

Virus-free.www.avast.com

On Sun, May 3, 2026 at 12:44 PM Douglas Miller <durga...@gmail.com> wrote:

Seems to me that a very simple H8-like machine would be nice. Display and keypad to simulate the front panel, minimal RAM and ROM, no HDOS or CP/M support. The 8085 SID/SOD could be implemented as a download/upload channel, something like the cassette interface or to load/save Intel HEX format programs. An expansion bus then to add additional hardware in the future. Connecting the SID/SOD to a USB-Serial dongle (also supplying power) would then allow connecting to a PC for download/upload. I'm not sure how BHBasic would work over that serial connection, but with an expansion bus you could add a regular UART. Of course, the SID/SOD interface would never work with existing software that expects an 8250 UART.

--

You received this message because you are subscribed to the Google Groups "SEBHC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sebhc+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sebhc/e43bde98-c5f7-4280-bc95-dc2afdd6783en%40googlegroups.com.

[Joseph Travis]

Lee,

Norberto used a small 2KHz buzzer on his Super 89 board design that doesn't require an AC signal to emanate sound. It's available from Jameco and DigiKey: AdaFruit Industries part number 1536 for about $1.25.

Joe

To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAF__8hdDCRmODKGjYqNheY6%3DcHVnKSj_cMdjA9zj8P7QgPCkRQ%40mail.gmail.com.

[Eric Mack]

It’s time to downsize. This system includes software manuals, H14, DC Hayes Modem, and more. Most impressively, it includes Norberto’s CF Memory board.

 

Complete Heathkit H8 Computer. Working, with Software and Peripherals & Modem | eBay

 

Eric Mack

[Douglas Miller]

It might be nice if the ROM could optionally be a 28C256, rather than only 27C256.

[Lee Hart]

I thought of that. But I was thinking it might be more entertaining to have direct control of the speaker bit so you could bit-bang it to play chiptunes. :-)

Virus-free.www.avast.com

To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAGQDgBDnssepjkPJBGdpN8Q%2BwkM2XjMR%3D%3D_7d0%3D6AWxX5K-2Aw%40mail.gmail.com.

[Lee Hart]

I guess I'm "committed". :-) I ordered a batch of OKI 80C85 CPUs, 3x7 segment LEDs, and 32k RAMs for the H8 badge today. An updated schematic is attached. I replaced the triple analog switch with a quad 74HC132 NAND gate. I swapped 82C55 ports A and B so the display row/column address and bit assignments are the same as the H8. The extra gate also provided a 2 msec clock. Only 8 chips!

Software for it will be "entertaining". I can't find an online source for PAM8. Does anyone have it? I do have source for XCON8 but it's a lot more complicated and has a lot of code that would be useless in this project.

Joe Travis also noted that BHBASIC source isn't available, and it has embedded I/O code that won't be present in the badge. Anyone have any ideas for workarounds?

Virus-free.www.avast.com

[Douglas Miller]

I'd ask again that you add jumpers so that the 28C256 can be used instead of the 27C256.

I'll also echo what Joe said. Bit-banging the speaker to implement the H8 "horn" is going to seriously cripple software. The H8 turns on/off the "horn" in the 2mS interrupt handler, and so it takes no overhead - does not steal cycles away from anything else. bit-banging both speaker and serial port is going to make for some challenges.

[Joseph Travis]

Attached is a reconstruction of the XCON8 ROM Monitor source code for the H8.  There are some good notes in the file.

Joe

To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAF__8hcF0%2B2jJaZqweobHk8Vs-kj5DBpQvuebcx7tHc34A0%2BcA%40mail.gmail.com.

[glenn.f...@gmail.com]

David Troendle and I did the work to recreate HDOS source in ASM form. It lives on bitbucket (David’s preferred repository):

https://bitbucket.org/HeathkitGuy/sebhcsoftware/src/master/HDOS/HDOS2.0/

 

I believe for what’s posted there we assembled the code and compared to the original ABS files and found either no differences or differences only in uninitialized storage.

 

We also did BH BASIC but I don’t see it posted on the site. Perhaps we never came full circle and completed that. I have a work in progress version on my computer. The files are dated June 2019 so it’s been a while!...

 

• Glenn

To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAF__8hcF0%2B2jJaZqweobHk8Vs-kj5DBpQvuebcx7tHc34A0%2BcA%40mail.gmail.com.

[glenn.f...@gmail.com]

If you do the bit bang on the speaker make it compatible with the mod that was popular back in the early days. Basically instead of gating the 1Hz tone the mod lets you gate the speaker directly. Of course this takes all of the cpu when you do it. The programs are on HUG volume 1

 

 

 

From: se...@googlegroups.com <se...@googlegroups.com> On Behalf Of Lee Hart
Sent: Tuesday, May 5, 2026 4:19 PM
To: se...@googlegroups.com
Subject: Re: [sebhc] Re: An H8 Badge for VCFMW

 

I thought of that. But I was thinking it might be more entertaining to have direct control of the speaker bit so you could bit-bang it to play chiptunes. :-)

 

Virus-free.www.avast.com

To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAF__8he4peVBAHwYoWsBT2%3DGDgPEaODsGFoOr184UbZna%2BphiA%40mail.gmail.com.

[Eric Mack]

Guys, now that my H8 system is sold and shipped, I have two floppy drives and a drawer full of new 5.25" and 3.5" floppies all DSHD. Any Interest? As for the drives, I have one Tandon TM-100-2A and One Toshiba MKM 0202A 001 drives.

 

 

Eric Mack

[Lee Hart]

Consarn that H8! I just realized that its LED segments have *three* inversions, not two. A LOW data bit lights a segment. My current design has a HIGH light a segment.

This inversion seems like a biggie. I can add an octal inverter; but every extra part makes it bigger, more expensive, and will take longer to build (creeping featuritis). On the other hand, every deviation from the original means more H8 software won't work without modifications.

It seems like the only programs that will run without patches are ones that strictly call routines in the ROM (so the ROM entry points can be written to work the same as an H8). Any thoughts?

Virus-free.www.avast.com

To view this discussion visit https://groups.google.com/d/msgid/sebhc/01f801dcdcf2%240f530e50%242df92af0%24%40gmail.com.

[Lee Hart]

Thanks Glenn. I'll look up the speaker mod and see what it did, and whether it would work on the mini H8.

Hopefully BHBASIC would only need patching in its I/O routines. Please let me know if you find your source!  

As for HDOS running on this mini H8: Man, that's a stretch! I suppose I could use a flash ROM, like the Microchip 512k SST39SF040 (about $4 from Newark). It could simulate a disk drive. Jon Chapman used it on his Glitch Works 8085 SBC, which runs CP/M with it.

Virus-free.www.avast.com

To view this discussion visit https://groups.google.com/d/msgid/sebhc/01e901dcdcf1%24b6400430%2422c00c90%24%40gmail.com.

[Douglas Miller]

My 2 cents is to keep it simple and don't try to make it run HDOS or CP/M - at least not without an expansion bus/board. But allowing a 28C256 instead of 27C256 is important. The 7-segment inversion is problematic, although if existing code depends on the ROM to do the actual refresh then the ROM could invert the data before sending to the 8255. The keypad scanning is not software compatible anyway, and I don't believe the ROM provides any general purpose "keypad input" routine.

Providing a way to download programs will be valuable, although the bit-bang SID input will need to ensure it can process each character in the limited time available between the last bit being detected and the next START bit - less than 1.5x bit rate (slower baud means more time to process).

[Dave McGuire]

Eric, I'd be interested in picking those up from you for LSSM, if
they're still available.

-Dave

--
Dave McGuire, AK4HZ
New Kensington, PA

[glenn.f...@gmail.com]

So as I said, I do have the source for BH BASIC (110.06.00) – as distributed with HDOS 2, it may still have some typos from the OCR process but could presumably be repaired, but it makes extensive use of HDOS (SYSCALL) and not just for I/O. making it a ROMable or stand-alone version would be non trivial.

 

IMO more useful would be an earlier version from the “Tape OS” era. We have those in executable form and that may be enough?  My understanding is the source was available for this way back when but I’m not aware of any accessible copy on any of our servers (though I recall someone here saying they had it).

 

I agree that running HDOS on this system would be a stretch. Stick to the early era of software…

To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAF__8hfFeZPOSMpr1g5eK%3DU6yPqyO%3DO%2BJkpB2hzrN9aQpJtAQw%40mail.gmail.com.

[Eric Mack]

Hi Dave,
I'm giving them to Norberto

Eric Mack
Founder, Speaker, Consultant
Click here to schedule an appointment.

Eric...@IntentionallyProductive.com
(661) 242-8410
linkedin.com/in/EricMack
Deliver your best with calm and clarity.

-----Original Message-----
From: se...@googlegroups.com <se...@googlegroups.com> On Behalf Of Dave McGuire
Sent: Wednesday, May 6, 2026 1:07 PM
To: se...@googlegroups.com

--
You received this message because you are subscribed to the Google Groups "SEBHC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sebhc+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sebhc/9b80866f-bb3f-4f8d-b987-367dbc3fd66a%40neurotica.com.

[Lee Hart]

> I'd ask again that you add jumpers so that the 28C256 can be used instead of the 27C256.

Douglas, I like this idea! But the 28C256 is an expensive part (Digikey lists it for $11.62). I wonder if a flash ROM might be a better choice? For example, Digikey has the Microchip SST39SF020A-70-4C-PHE (a (256k x 8 32-pin DIP) for $3.36. The 4 unused bits of the 82C55 could be used for the extra address lines (A15-A18).

> I do have the source for BH BASIC (110.06.00) – as distributed with HDOS 2... IMO more useful would be an earlier version from the “Tape OS” era. 

Herb Johnson sent me a link to a PDF of this BASIC source. I looked, and agree that it would be a big problem to use it without HDOS. The earlier tape-based 4k BHBASIC should be easier to adapt, if the source can be found.

PS: Today's little additions to the schematic were to add pullup resistors to /WR /RD and IO/M as these pins get tri-stated during Reset.

Virus-free.www.avast.com

To view this discussion visit https://groups.google.com/d/msgid/sebhc/047501dcdd98%240ff7d560%242fe78020%24%40gmail.com.

[Douglas Miller]

Oh, I see, eliminate the 27C256 entirely and replace it with a larger flash device. I see that the DIP32 versions of SST39SF020/SST39SF040 are discontinued (or soon will be), but I guess one can get a PLCC adapter and use PLCC32 ones.

[Lee Hart]

I'm not sure I can get the 040 chips, or what they'd cost. But it's possible. :-)

The memory mapper provides 32k of ROM and 32k of RAM. There are 4 unused outputs (PC0-PC3), so that could address 16 banks of 32k = 512k. But I'd kind of like to keep a spare output bit to change to an org.0 memory map. Maybe PA4 could be used for that, since there is no single-step circuit?

I'd rather avoid surface-mount parts if we expect hobbyists to build it.

Virus-free.www.avast.com

On Wed, May 6, 2026 at 9:10 PM Douglas Miller <durga...@gmail.com> wrote:

Oh, I see, eliminate the 27C256 entirely and replace it with a larger flash device. I see that the DIP32 versions of SST39SF020/SST39SF040 are discontinued (or soon will be), but I guess one can get a PLCC adapter and use PLCC32 ones.

--

You received this message because you are subscribed to the Google Groups "SEBHC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sebhc+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sebhc/1149b5a2-2a53-48c1-a354-61717848f107n%40googlegroups.com.

[Douglas Miller]

PLCC are not surface mount, but their sockets do have pins in a denser arrangement. Not that bad to solder, IMHO. The 040 parts are possibly cheaper than 020. In keeping with most storage, the larger devices quickly become the normal and thus cheapest, and both have been around for some time now.

The problem with bank selecting is that you need to provide some mapping where the code being executed at the time banks are switching does not get switched. And ORG0 introduces a new complexity to the mapping. At the very least, you probably want to use a GAL/CPLD for the memory mapping/address decoding, but not sure what size you'll need to support the mappings you need.

[Lee Hart]

Hi Douglas. Sorry for the pause. I've had some health issues. But I've also had time to think. :-)

This mini-H8 is intended to introduce people to Heath and kit-building, so KISS is the name of the game. I'd like to get at least a basic "toy" ready for VCFMW, but allow for expansion. Through-hole is desirable, as even a PLCC socket is difficult to solder. But Flash ROMs are still available in DIPs. The 040 seems to no longer be available, but the 010 and 020 aren't too expensive or hard to find (yet). They could be a cheap / easy way to provide "disk" storage.

Here's my thought: As you suggest, provide ROM socket jumper options for a 32k 27C256 EPROM (cheapest base model), 64k 27C512 EPROM (not much more) or 128k SST39SF010 flash ROM (for a write-able "disk"; $2.89 from Digikey, but last time buy is 7/1/26).

Then I can provide two memory maps:

1. ROM at 0-8k, RAM at 8-40k, ROM at 40-64k (the setup in my last schematic). Similar to H8 at power-up or Reset. The upper ROM could imitate a disk drive. It could theoretically run HDOS, though I'm not enough of a programmer to do it.
2. Org.0 with RAM at 0-32k, ROM at 32-64k. This could theoretically run CP/M. I can use bit 4 in the digit select port which is currently unused. (It was the H8 INT20 single-step interrupt). Looking back at my schematic, bit 4 can be tied to the D0 input of address decoder U5. A resistor can pull it high on power-up or Reset to put ROM at 0-8k, and bit 4 can be set low to put RAM at 0-8k the D0 input of the U5 address decoder. The unused bits PC0-PC3 can be the extra address lines for larger ROMs.

Yes, bank switching like this can really be messy software-wise! But I've done it on my Z80 and 8080 membership cards. The program that's switching banks just has to be in RAM when you "blow up" the ROM!

I'll post a new schematic when I get this entered.

Virus-free.www.avast.com

On Wed, May 6, 2026 at 9:47 PM Douglas Miller <durga...@gmail.com> wrote:

PLCC are not surface mount, but their sockets do have pins in a denser arrangement. Not that bad to solder, IMHO. The 040 parts are possibly cheaper than 020. In keeping with most storage, the larger devices quickly become the normal and thus cheapest, and both have been around for some time now.

The problem with bank selecting is that you need to provide some mapping where the code being executed at the time banks are switching does not get switched. And ORG0 introduces a new complexity to the mapping. At the very least, you probably want to use a GAL/CPLD for the memory mapping/address decoding, but not sure what size you'll need to support the mappings you need.

--

You received this message because you are subscribed to the Google Groups "SEBHC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sebhc+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sebhc/d8e18a1a-590c-4fa4-94ed-f5520e272f1fn%40googlegroups.com.

[Douglas Miller]

Best of luck. I think some of those things are more complicated than they seem. If it were me, I'd focus on making the best platform for your original intent - a miniature bare-bones H8 with a way to be able to upload/download small programs from a PC (as if it were a cassette). I think HDOS, ORG0, and CP/M just unnecessarily complicate the whole thing. In my opinion.

[Lee Hart]

I agree. The rev. version is sure to be bare-bones, as that's all I'm likely to be able to do myself in time for VCFMW.

But jumper options and bigger parts cost nothing if not populated.

Virus-free.www.avast.com

On Mon, May 18, 2026 at 6:31 PM Douglas Miller <durga...@gmail.com> wrote:

Best of luck. I think some of those things are more complicated than they seem. If it were me, I'd focus on making the best platform for your original intent - a miniature bare-bones H8 with a way to be able to upload/download small programs from a PC (as if it were a cassette). I think HDOS, ORG0, and CP/M just unnecessarily complicate the whole thing. In my opinion.

--

You received this message because you are subscribed to the Google Groups "SEBHC" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sebhc+un...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/sebhc/7cd92dcf-ce17-4897-8ac3-a20e4519ae54n%40googlegroups.com.

[Lee Hart]

Hi all,

I've been having some health problems, which slowed me down. But here is an updated schematic for the mini H8.

- added jumper options so it can use anything from a 27C256 32K EPROM to a 39SF040 512K flash ROM.

- added battery backup for the RAM.

- org.0 switching so the first 8K can be RAM or ROM (so could theoretically run HDOS or CP/M).

- a 2 mSec clock on RST7.5 (the only 8085 edge-triggered input).

- serial using SID and SOD pins.

- serial input also triggers a RST6.5 interrupt to avoid missing Start bits.

Though drastically simplified (only 8 chips), I *think* it should be able to accurately emulate a real H8. But since the I/O is different, I'm sure a lot of H8 software will need patches to run on it. Does anyone know how many programs talk directly to I/O, versus ones that stick to monitor calls?

[Bob Bell]

Lee,

Sorry to hear of your health problems and pray you are on the mend.  I have been there, done that several times now too.

I have been mostly lurking on this thread since you started it.  As a builder of many Heathkits, most still functioning here in my office/shop, I always wanted an H8, but never had the $$ at the time.  So I am keenly interested in this project and will most likely buy from you when they are ready, assuming the plan is to make bare boards available to the community like many do over on the S100Computers group.

My question to you as I look over your schematic: What EDA software are you using?  Is it KiCAD?

I really appreciate your logical layout of the schematic.  It is very readable, well documented, and easy to follow unlike the schematics for some of the modern S-100 boards that have been created in the past 20 years where components are just plastered on the sheet with connections by netname only and little explanation of function.  That method leaves one wondering how it works.

I posted a comment some time ago in regards to the serial I/O.  I still think a real SIO chip would be welcome by many.  I know you want to keep the first version very simple, but maybe just add a header with all the bus signals, so perhaps a daughter board could be added in the future to accommodate the SIO and other add-ons.

Thanks for what seems like a very worthwhile project!

Bob Bell

To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAF__8hc_E4A7oZF-FgaVfeTiRV4WtS8sW_ACXFDxfAYu5YFH3w%40mail.gmail.com.

[Joseph Travis]

Hi Lee,

I've given this some more thought over the weekend.  Due to the I/O differences between this project and the original H8, it looks like the monitor ROM will be a complete rewrite.  We can double up on the SID/SOD ports, using it for both console and tape I/O.  The unfortunate thing is that many H8 tape programs won't work because they use direct I/O but, you'll be able to create, dump and load new programs the same way it was done on the H8.

Are you going to implement a SD interface similar to the Z80 Altoids computer?  I'm thinking that we may be able to boot CP/M with a proper BIOS and possibly HDOS 3.0 as well (not HDOS 2.0 or 1.5).  HDOS 3.0 has an external console driver (TT.DVD) that could be rewritten for SID/SOD and the SY.DVD would have to be written for the SD card.  Of course the monitor ROM will need to have the necessary code to boot from the SD.

That's my thoughts for the moment.  BTW - I hope that you are feeling better soon!  Oh, just one more thought... In order to reduce / minimize confusion, you might want to change the name of this project to H8.5 which gives a nod to both the H8 and the 8085 CPU?  'Just a(nother) thought.

BR,

Joe

To view this discussion visit https://groups.google.com/d/msgid/sebhc/06f885c2-6af9-4ba6-b875-8d894a089dd2%40gmail.com.

[Lee Hart]

I've made a few small additions to my mini H8 schematic.

- I changed the SPST power/standby switch to a SPDT switch, so the Standby position grounds the RESET line to the 8085. This should stop it from going "nuts" and writing nonsense to memory as it shuts down.

- I added a jumper option so memory from 0-8K can be set for either RAM or ROM at power-on or RESET. It obviously needs to be set to ROM initially. But since RAM is battery backed up, a program in RAM (like CP/M) could be executed immediately. This also provides a path to initialize a new flash ROM without a programmer.

Bob, I use OrCAD SDT386 for my schematics and PCB layouts. It's an old DOS program from 1996. It's powerful and extremely fast on modern hardware. I particularly like that it is keyboard-driven (no mouse and menus needed).

Joe, I've started looking at the XCON8 source, and I agree it needs to be extensively rewritten. But I still don't know what entry points in it should be kept the same for external programs. I guess time will tell.

I can include an expansion connector to add a UART or SD-card. But it's so small that H8 or even H89 I/O cards would dwarf it. They would probably be on a new PCB. I could rework my Z80-SIO card to work.

I haven't really settled on a name for this little guy. I've been calling it a "mini H8", but that can change. Any ideas? "H8-5" sounds like an accessory card for the H8.

Now... Onward to see if I can layout this little darling!

Virus-free.www.avast.com

To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAGQDgBC0rgNcnK-G%2BPqcnyD3Bv%2B%3DaLyLBRczKRYt3-hLsyb%2B8A%40mail.gmail.com.

[Joseph Travis]

Hi Lee,

I believe the entry points you are looking for are actually part of the H17 boot ROM.  They should be referenced in the ASM / ACM files included with HDOS.

BR,

Joe

To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAF__8hdfhk-Ls8AnMManWNEG1eCMHK%2BHKzGbrU6KBPTxLmQbrg%40mail.gmail.com.

[Douglas Miller]

There are definitely some "entry points" in the main ROM that were used (I won't comment on that practice, but you can guess my thoughts). I tried to preserve those in my new H8/H89 monitor ROM, they are marked by conditionals that check the addresses. I'm by no means certain I've got them all, but so far there haven't been any complaints. See https://github.com/durgadas311/MmsCpm3/blob/master/rom/newmon/src/h8core.asm

On Wednesday, June 3, 2026 at 7:22:24 PM UTC-5 Lee wrote:

...

Joe, I've started looking at the XCON8 source, and I agree it needs to be extensively rewritten. But I still don't know what entry points in it should be kept the same for external programs. I guess time will tell.

... 

[Glenn Roberts]

Right. MTR.ACM as I recall? Perhaps another?

To view this discussion visit https://groups.google.com/d/msgid/sebhc/CAGQDgBCv-HmdT0MPri_zkKnD-h%2BAW9n6d%2BehBkY9Mi6OFq%3DS4g%40mail.gmail.com.