H8-5 Cassette Rework Proposal.
norberto.collado koyado.com
We have seen several original H8-5 within our group not working properly with the 8251’s that are not from National Semiconductor INS8251. The INS8251 are very expensive and hard to get.
Attached is my proposal to fix this issue. Please review and let me know if any concerns.
I need to close on this to be able to release the new H8-Z5-4 board.
Thanks,
Norberto
Joseph Travis
--
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 on the web visit https://groups.google.com/d/msgid/sebhc/SN6PR01MB3855544B0516C27B8A73BDAEF7DE9%40SN6PR01MB3855.prod.exchangelabs.com.
dwight
To: se...@googlegroups.com <se...@googlegroups.com>
Subject: Re: [sebhc] H8-5 Cassette Rework Proposal.
norberto.collado koyado.com
As long you have an INS8251 in IC123 socket, the board will work fine to do the Dump cassette function. Otherwise, it doesn’t work.
Norberto
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/BYAPR01MB53199924631A218037684793A3DE9%40BYAPR01MB5319.prod.exchangelabs.com.
norberto.collado koyado.com
It won’t work as there is a filter to clean up the TxE glitches when the buffer is empty.
From: se...@googlegroups.com <se...@googlegroups.com>
On Behalf Of Joseph Travis
Sent: Wednesday, June 1, 2022 9:36 PM
To: se...@googlegroups.com
Subject: Re: [sebhc] H8-5 Cassette Rework Proposal.
How about this... NOR these 3 signals: !WR(10), !CS(11), C/!D(12) then OR the output with TxE(18). I believe that may kickstart the process.
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/CAGQDgBD6qSqYgWU%2B0%3Db_En2xEc%3DX759%2Bozh7tV%3DFx3yncdPXYQ%40mail.gmail.com.
norberto.collado koyado.com
I’m adding a gate to use either TxE or /DTR to the new H8-5 board.
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/SN6PR01MB3855A6B80B8BD20F776B25AEF7DE9%40SN6PR01MB3855.prod.exchangelabs.com.
Douglas Miller
My concern about a solution that involves software changes is
that it may not be isolated to the ROMs (which also includes the
v4 Z80 board ROM). I thought previous cassette discussions
revealed that BHBASIC, cassette monitors, and possibly even HDOS
or CP/M programs, would also need to change. Or has that been
debunked and is just stuck in my head? My impression was that the
ROMs did not export a well-defined API for the cassette, so most
software (that accessed the cassette) had to include their own
cassette routines.
--
glenn.f...@gmail.com
How does one identify the INS8251? I have several H8-5 boards and they have Intel branded chips – large ‘i’ followed by “P8251”. Are these INS? They all are Heath original parts (443-776) and work fine.
Surprising to me that there are variations in behavior across 8251s…
- Glenn
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/SN6PR01MB385562B3AEA30FF24D921C72F7DE9%40SN6PR01MB3855.prod.exchangelabs.com.
norberto.collado koyado.com
Interesting! Can you take a picture of such 8251?
This is what it works for me and nothing else:
Thanks,
Norberto
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/032d01d8767c%24bee57280%243cb05780%24%40gmail.com.
norberto.collado koyado.com
Yes, you are correct! I will leave it as original schematic.
From: se...@googlegroups.com <se...@googlegroups.com>
On Behalf Of Douglas Miller
Sent: Thursday, June 2, 2022 5:09 AM
To: se...@googlegroups.com
Subject: Re: [sebhc] H8-5 Cassette Rework Proposal.
My concern about a solution that involves software changes is that it may not be isolated to the ROMs (which also includes the v4 Z80 board ROM). I thought previous cassette discussions revealed that BHBASIC, cassette monitors, and possibly even HDOS or CP/M programs, would also need to change. Or has that been debunked and is just stuck in my head? My impression was that the ROMs did not export a well-defined API for the cassette, so most software (that accessed the cassette) had to include their own cassette routines.
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/bbccf1c4-71cf-7dde-8c7e-9d03098d2752%40gmail.com.
glenn.f...@gmail.com
I have three operational H8-5 boards. I am not the original owner of any of them – they came from systems I have acquired over the years. Two have the “intel” branded 8251 chips:
https://photos.app.goo.gl/nFUaW5dYttgkhawZ7
and the third has the (National Semiconductor) INS8251 chips. I can’t prove it but it is my belief that most of these chips are original to the boards (I do note one of them has a 1990 date, but the others are dated ’82).
I’ve used these boards interchangeable in multiple H8 systems and never noticed any issues. I have never used the boards with an actual cassette recorder – primarily have used them digitally (reading/saving .H8T files) using Dave Runkle’s daughterboard:
https://photos.app.goo.gl/xRHXp3VfkKBx88b7A
The one analog use case was when I played .WAV files from an iPad and saved the images as .H8T files. That worked fine.
Is there a specific test (one that doesn’t involve a physical tape recorder, since I don’t currently have a working one) that you’d like me to run on the Intel ones?
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/SN6PR01MB3855DF48DFF14E027630FCDAF7DE9%40SN6PR01MB3855.prod.exchangelabs.com.
norberto.collado koyado.com
Sent: Thursday, June 2, 2022 9:50:37 AM
To: se...@googlegroups.com <se...@googlegroups.com>
glenn.f...@gmail.com
I have three operational H8-5 boards numbered 1, 2 and 3
Boards 1 and 2 have the Intel chips; board 3 has the INS8251 chips
The CPU is an 8080 board with the original 444-13 ROM.
Board 2: Turned on and hit DUMP. After the 5-second timeout the display begins to show the countdown. This is with the intel 8251 installed in the “Astrorat” daughterboard. This happens in both modes (with and without the jumper installed, i.e. in “digital” and “analog” dump modes). Removed the 8251 from the daughterboard and installed it directly in the socket – same result – works properly as expected – LEDs begin the countdown after 5 seconds.
Board 1: nothing happens when I hit Dump. This board has had this issue before and I guess I never fixed it? the 5-second timer did not work as expected. I will have to revisit but I think I had diagnosed a bad chip. Would have been nice if I made a note of that ☹
Board3: works as expected. After 5 seconds the LEDs begin to count down.
I will take a look at Board 1 and see why it’s not working, but I don’t believe it’s an 8251 issue…
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/SN6PR01MB3855C832D8344A33DEE023D3F7DE9%40SN6PR01MB3855.prod.exchangelabs.com.
Douglas Miller
Based on no actual evidence, I would have thought that Intel was
the original designer of the chip, and that NS was an authorized
second-source. It would make sense then, based on that conjecture,
that both Intel and NS parts would work. I would then guess that
these newer parts are some sort of "re-imagining" of the
originals, and they didn't get it quite right. At least it makes a
good story.
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/03d301d876c6%247de59630%2479b0c290%24%40gmail.com.
norberto.collado koyado.com
So, what you are indicating is that the Intel P8251 works fine. The Intel P8251A that I have will not work. Neither the P82C51A.
Works for Glenn:
Doesn’t work for Norberto:
I also have a P8251 from AMD and it works.
Works for Norberto
In summary:
- Any P8251 will work fine.
- Any 8251A will not work with the H8-5 board.
- The INS8251N is also labeled as P8251 and that makes sense.
Thanks Glenn for adding more clarity at my end with your testing.
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/03d301d876c6%247de59630%2479b0c290%24%40gmail.com.
Peter Higgins
glenn.f...@gmail.com
All of my Intel chips are marked P8251 (not 8251A) and yes they pass the “DUMP” test.
There were at least three versions of the H8-5 board.
My boards 1 and 2 were part number 85-2109-1 and they apparently came with the Intel 8251 chips.
My board #3 is an 85-2110-2 and it contains the INS8251 chips.
I actually found a fourth board that I had torn down (it had been cannibalized by previous owner) it is an 85-2110-3 and it also has the INS chips.
Perhaps some design changes between the 2109 and 2110 versions necessitated the switch to INS versions of the chip?
I’ve found at least three sets of doc’s (595-2032, 595-2032-01, and 595-2032-03). They all refer to the part as Heath #443-776 and say “may be replaced by 8251”.
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/SN6PR01MB38554241E7B7D5EB988C5B44F7A19%40SN6PR01MB3855.prod.exchangelabs.com.
glenn.f...@gmail.com
p.s. just noticed the dates on the boards:
85-2109-1 3/16/77
85-2110-2 12/15/77 (silkscreen) 10/24/77 (solder mask)
85-2110-3 3/20/79 (silkscreen) 10/24/77 (solder mask)
Guessing there was probably also an 85-2110-1 dated 10/24/77 as the -2 and -3 versions keep the 10/24 date on the solder mask and update the date on the silkscreening.
norberto.collado koyado.com
I’m curious on why Heath did not use the /DTR signal instead of /TxE signal to set /CTS low. The 8251A is working as it should be per specs. TxE should wait for /CTS to go low and no the other way around.
As I have a GAL for I/O decode with spares lines, I will add to monitor for port 371Q to be asserted and then trigger a flip-flop to kickstart /TxE. Then when /TxE goes low it will clear/lock the flip-flop. When the transfer is completed, TxE goes high forcing /CTS to go high to re-enable the flip-flop. This will allow to use any 8251A, etc…
Per PAM-37 to enable TxE:
Workaround steps:
- GAL to monitor port 371Q (DUMP port)
- When port 371Q is detected, trigger a flip-flop (MOTOR-ON). /CTS is still high
- After 5 seconds, /CTS goes low. Then TxE goes low, clearing the flip-flop and preventing future triggers.
- When transfer is completed TxE goes high and that will re-enabled the flip-flop.
This H8-5 new board is become a challenge due to P8251 and P8251A variations.
I think this circuit with the flip-flop will clear up the issue:
Norberto
From: se...@googlegroups.com <se...@googlegroups.com>
On Behalf Of Peter Higgins
Sent: Thursday, June 2, 2022 6:34 PM
To: SEBHC <se...@googlegroups.com>
Subject: Re: [sebhc] H8-5 Cassette Rework Proposal.
Norby - you stated in a related thread that:
--
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 on the web visit https://groups.google.com/d/msgid/sebhc/923e378d-3604-4efd-89bf-680a56eb0e47n%40googlegroups.com.
Joseph Travis
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/SN6PR01MB3855C3D615780AA31208E91BF7A19%40SN6PR01MB3855.prod.exchangelabs.com.
norberto.collado koyado.com
Thanks to everyone for their feedback and hopefully the added changes will fix the issue on the new board.
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/042701d876ec%24c1a7f900%2444f7eb00%24%40gmail.com.
norberto.collado koyado.com
That is possible as the GAL is helping a lot as well. For the original H8-5 board I will encourage to find P8251 in eBay or other sources to avoid any additional changes.
This is what the GAL will do only on 371Q writes to enable the motor-on signal.
IO_CASS_L_371Q = !(IoEnable & A_371Q & !WR_L) ;
Thanks,
Norberto
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/CAGQDgBDg16G2faL%2BHOq%2Bw6CJKE72REKnOGpY4CYwaJL7bttucg%40mail.gmail.com.
Douglas Miller
You may also need the GAL to detect D0=1 (TxEnable) when
triggering the motor-on/delay. Otherwise, you start recording when
trying to read cassettes? (371Q also written, but D0=0).
To view this discussion on the web visit https://groups.google.com/d/msgid/sebhc/SN6PR01MB38555D3F84A0DB42FB4907B6F7A19%40SN6PR01MB3855.prod.exchangelabs.com.
Joe Travis N6YPC
norberto.collado koyado.com
Sent: Friday, June 10, 2022 8:30 AM