29th Jun 2020, 9:55 pm | #321 |
Pentode
Join Date: Jul 2017
Location: Toulon, France
Posts: 239
|
Re: Recreating the Bywood Scrumpi
I also plan to make a box with my 3d printer for this little jewel
|
29th Jun 2020, 9:56 pm | #322 |
Pentode
Join Date: Jul 2017
Location: Toulon, France
Posts: 239
|
Re: Recreating the Bywood Scrumpi
all files will be available after TIM validation
|
29th Jun 2020, 10:12 pm | #323 |
Octode
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
|
Re: Recreating the Bywood Scrumpi
No pressure again then... Will be nice to have a PMOS based one finally
the SCRUMPI 2 surprise sounds interesting - I must get back on the video card engineering so we are ahead of the curve on the 3 as I hold out little hope of ever finding a real one. I hope it is only a backing box you plan to make Phil - the PCB is too pretty to hide so needs a perspex front... in fact it needs a viewing hole on the back as well for the artwork on there... Head over to the below thread to see how I plan to mount mine probably behind clear perspex... who knows maybe KarenO will get there first... https://www.vintage-radio.net/forum/...10#post1264810 Mark - on my Vero model I put a small wire link through the tab holes and soldered them as they have no support if the board is not mounted on anything and entering a lot of bytes gives them a real pounding! |
30th Jun 2020, 1:25 am | #324 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,265
|
Re: Recreating the Bywood Scrumpi
Hi Tim
Vero probably doesn’t give as much support as FR4 and you don’t have the through hole plating for electrical contact. Mine was using a Z80, but similar principle to scrumpi, processor generates address and switches force data to the memory, toggle switch to either step by forcing a nop or single step the code. Space was a bit tight in 100mm x 100 mm, but using smd leds and resistors on back of the board. Definitely agree the sc/mp should be on display, especially in gold and ceramic. I think I may try and order one delivered to my Mum so I don’t need to worry about shipping, she already has a bunch of mk14 displays waiting for me. Is it still possible to get the ram in ceramic? I think I have a couple of 2102s but not enough for a set of eight. I might have enough of the ttl in ceramic, although they tend to be a bit ugly. I was researching clear solder mask, wondering if it might give a more retro look than green for technology of that age, but it doesn’t seem common. Maybe bare FR4 with HASL is more authentic. |
30th Jun 2020, 10:45 am | #325 |
Octode
Join Date: Apr 2018
Location: Newbury, Berkshire, UK.
Posts: 1,287
|
Re: Recreating the Bywood Scrumpi
The SCRUMPI was probably bare SRBP rather than FR4..'.
|
30th Jun 2020, 2:49 pm | #326 |
Rest in Peace
Join Date: Jul 2011
Location: Bridgnorth, Shropshire, UK.
Posts: 787
|
Re: Recreating the Bywood Scrumpi
TimB,
I have just read your document 'The soul of an old machine' and I have to say, what a wonderful service to the community you are providing in pulling together all of this research. There's plenty in there that I didn't know about and much to take in. Thanks for an absorbing read |
30th Jun 2020, 4:37 pm | #327 |
Pentode
Join Date: Jul 2017
Location: Toulon, France
Posts: 239
|
Re: Recreating the Bywood Scrumpi
as promised the copper side photos of the true scrumpi
of our Canadian sunmoonstar mike friend I succeeded in contacting him by the National Astronomical Association of Canada (and yes adorable). he always has his scrumpi and it works mike is super friendly |
30th Jun 2020, 6:08 pm | #328 | |
Octode
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
|
Re: Recreating the Bywood Scrumpi
Quote:
Glad you found it useful as well, I have a stack more on SC/MP in general and the later SCRUMPI devices some of which is scattered in this thread that I hope to build into the next document... |
|
30th Jun 2020, 6:10 pm | #329 | |
Octode
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
|
Re: Recreating the Bywood Scrumpi
Quote:
Let us hope we also get a response from The Man in York now... |
|
30th Jun 2020, 6:23 pm | #330 | |
Octode
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
|
Re: Recreating the Bywood Scrumpi
Quote:
Sounds like an interesting Z80 project - is it written up anywhere? I did not get ceramic ones but, they are the very grey plastic so look pretty ancient. You only need two for SCRUMPI 1, 6 for SCRUMPI 2 and as you say 8 for SCRUMPI 3 - I managed to get 6. When Martin did his recreated MK14 he just left the mask off - comes with its own issues but, does look very good. |
|
30th Jun 2020, 6:57 pm | #331 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,484
|
Re: Recreating the Bywood Scrumpi
The original boards don't have any solder mask, the tracks are just tinned overall on both sides. The PCB itself is that semi-transparent fibre glass stuff. The Scrumpi PCB in #328 above looks of similar construction.
|
30th Jun 2020, 8:35 pm | #332 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,265
|
Re: Recreating the Bywood Scrumpi
|
30th Jun 2020, 9:16 pm | #333 | |
Octode
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
|
Re: Recreating the Bywood Scrumpi
Quote:
https://groups.google.com/d/topic/rc...VEc/discussion |
|
1st Jul 2020, 10:28 pm | #334 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,484
|
Re: Recreating the Bywood Scrumpi
With regard to making a Pi-Uploader for Scrumpi, it looks as though - if you turn off all special functions of the GPIO pins from inside Raspi-Config - there are 26 GPIO pins which you can use for parallel data and address output plus a load / write strobe.
As with Arduino, the only tiresome thing is that the GPIO port pins are not conveniently arranged in nice sequential groups of 8 parallel port pins, so to output an 8-bit data value or 12-bit address value to a group of GPIO pins you have to do a bit-scan of the variable holding the value you want to write out, set high the GPIO pins which represent any bits found high in the variable, and set low the GPIO pins which represent any bits found low in the variable - this is assuming that the optocoupler LEDs will have a common +3V supply and the GPIO port pins will be driven low to light the LEDs. GPIO pin low = LED Lit = Opto output turned on = data or address line pulled low. GPIO pin high = LED not lit = opto output not turned on = address or data line pulled or floating high. (Wired with a common +Ve LED supply and each cathode individually driven low to light the LEDs, the optos have a non-inverting action). With regard to conversion of SEND14 for this purpose, you'd intervene at the point where the script is about to send the address as four individual characters and instead write the raw address out to the GPIO ports as 12 or 16 parallel bits, and whenever the script is about to send the data as two individual characters, instead write the raw data out to another 8 GIO ports as 8 parallel bits. Where the script currently sends 'Mem' to enter the code and advance to the next address, make that pulse the GPIO pin whose optocoupler is connected across the load / latch input on the SCRUMPI instead. All of the foregoing assumes that the input switches on the SCRUMPI pull their respective lines down to 0V when closed and allow them to rise high when open. To allow the uploader to work, all input switches would have to be set to the open position and the optocoupler outputs connected across them. |
2nd Jul 2020, 6:08 pm | #335 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,265
|
Re: Recreating the Bywood Scrumpi
Another option for download would be similar method that I have used on z80. In bootstrap mode every read cycle is from ft245, while write cycles still write to ram. Then you send a series of machine code instructions to write a bootstrap program to ram, a jump instruction to the address of the bootloader, and then an instruction to disable the bootstrap mode. For the scmp one of the flag outputs would enable bootstrap mode. This method uses wait input driven by RX empty from the ft245. This method is very fast, 512k in less than 30 seconds on a z80 at 12MHz.
One advantage is there is no need to drive the address lines. A more retro method for the scmp might be to use a 6402 uart instead of the ft245. Or any input that does not need initialisation from the scmp. |
2nd Jul 2020, 6:21 pm | #336 |
Dekatron
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,484
|
Re: Recreating the Bywood Scrumpi
It has been pointed out as well by Slothie that the SC/MP is able to go hands-off on the buses so an external controller could take over the system, flood the RAM with program code and then hand back to the SC/MP. Any such system would need connections to the system buses though.
|
2nd Jul 2020, 7:50 pm | #337 |
Octode
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
|
Re: Recreating the Bywood Scrumpi
Actually I did start out adding a few extra Opto's to my board (which proved a worthwhile exercise as I got rid of the multiple cables and just modified send14 to vary the line layouts to work on my JMP, Slothie and Spartan machines in SW) with and some more GPIO pins to do exactly as Sirius describes. But, then I realised that the setup with a matrix press would have to do something like use one of the switches as a Ground and activate two to connect it with any of the data lines that were needed pulled low - but, then the Opto is the wrong way around as you only have four rows that can act as a source to one of the (on mine now) 10 columns that could be connected to a ground - I need to ground the pulled high data lines 8 of them of course and also control the LOAD and STEP as a minimum. So it would be a custom board for the SCRUMPI.
I even came up with a way of working around the CPU step (which I worked out for keying in longer programs) so "sendY" would have to understand instructions to hold the just written bytes to zero/or 0x08 to prevent jumps happening as the CPU will or could (on conditionals) execute them - this of course risks the chips with the high current draw of shorting the RAM Data lines to zero... which I have been doing briefly to allow quicker data entry... That is why I discounted stepping forced NOP's for every byte as that would definitely be a strain on the RAM's if they were not zeroed before. The Slothie method would need more work as well as it needs to put the whole chip in Tri-State as the default is the SCRUMPI holds on the Memory read so the CPU Data Bus is an input - It is really good but, as you say needs a lot of connections for the address as well - if you are doing that stitch an EPROM on and have some boot code for serial comms... which actually would be very contemporary and what happened on the SCRUMPI 2 and was recommended as an expansion for the 1. My current method then is the very short BOOT code I gave above (manually entered each time) and then sending the larger bulk over serial. My idea once that is as short as possible (which I think it now is) is to build a Diode boot board that can do as Mark says with a switch to override the RAM for the first bytes - I wanted to get it to 32 for simplicity so still studying that... Just because I have never had a machine with a Diode boot board and it would have been cheap enough to do BITD. |
2nd Jul 2020, 11:38 pm | #338 |
Octode
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
|
Re: Recreating the Bywood Scrumpi
Well you think you have found everything there is and suddenly lots crawls out of the woodwork... Phil and I may have some amazing news soon but, researching ETI again looking for their 550 VDU (anyone any ideas which issue that is? EDIT: It is Sept and Oct 1976...) I find this news item in ETI January 1977... sounds like some exciting addons coming for our little board.
Last edited by Timbucus; 2nd Jul 2020 at 11:52 pm. Reason: update to information |
3rd Jul 2020, 6:10 pm | #339 |
Pentode
Join Date: Jul 2017
Location: Toulon, France
Posts: 239
|
Re: Recreating the Bywood Scrumpi
pictures of the beast
the true |
3rd Jul 2020, 6:29 pm | #340 |
Pentode
Join Date: Jul 2017
Location: Toulon, France
Posts: 239
|
Re: Recreating the Bywood Scrumpi
see
|