UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Specific Vintage Equipment > Vintage Computers

Notices

Vintage Computers Any vintage computer systems, calculators, video games etc., but with an emphasis on 1980s and earlier equipment.

Closed Thread
 
Thread Tools
Old 30th Aug 2022, 10:13 pm   #1
Realtime
Hexode
 
Join Date: Jan 2021
Location: Ashford, Kent, UK
Posts: 318
Smile Yet another MK14 vdu card

Greetings.
I‘ve been posting a couple of messages on the introduction section, so thought I’d get down to the nitty gritty. I built a MK14 issue VI PCB about 2 years ago and have been creating add-ons for it. I’ve sold a few Tape Interface and Extension RAM modules to forum members and as a result had some interesting discussions about other extras to build. In particular, email chats with Coolsnaz2 fired me up to start creating a new, more compact VDU card based around a Xilinx CPLD. It’s been progressing well and I thought I would give a blow by blow account of progress in the coming weeks.
As I’m new to posting I’m not sure if I can attach photos and documents as yet, but the module is designed using DesignSpark PCB (I find it easier than Kicad for some reason) and I have a 3D view to share. Prototype PCBs are on order. Also been working on the VHDL using an FPGA development board to start with.

Part 2 of the project is a plug in replacement for the DM8678 character generator, again based around a CPLD. The aim here is to keep the pin out and function identical, just in a bigger package that would plug into a DM8678 socket on existing VDU boards. I’ll kick off a separate thread for that one in the near future.
Realtime is offline  
Old 31st Aug 2022, 10:03 am   #2
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Yet another MK14 vdu card

Hi Realtime, and welcome.

Perhaps it was also you who made the matching keypad frames / keypad overlay for the issue VI as well? If so we have been admirers of your work. As a new forum user your posts will be held for moderation for the first few posts but other than that, I think you can post images from the start so give it a try.

It sounds as though the design is fairly far advanced if you already have PCBs on the way, otherwise I would have suggested you include RAM at 0200-07FF on the VDU card itself because the original VDU was never really usable when it chewed up 512 bytes of the 640 bytes on a fully populated MK14.

I appreciate that you already did produce a memory expansion in the form of a daughterboard which goes in one of the 40-pin sockets with the original chip on top, but if this will be aimed specifically at the issue VI then it would be more elegant for the extra RAM and its associated address decoding to be on the VDU card.
SiriusHardware is offline  
Old 31st Aug 2022, 1:06 pm   #3
Realtime
Hexode
 
Join Date: Jan 2021
Location: Ashford, Kent, UK
Posts: 318
Default Re: Yet another MK14 vdu card

Yes, it was me that did the 3D printed keyboard overlay for the Issue VI. Just recently added the Issue 5 overlay as well thanks to Coolsnaz2 providing dimensions and doing the trial fits. On the subject of the Issue VI overlay I just updated the font to be far closer to the MK14 original. So if any forum members bought one with the old font and want an updated laminated insert PM me and I can get it sent out for the price of an envelope and stamp.

So, yes the VDU design has progressed quite well. Having followed the forums for a while I was aware of a number of shortfalls with the original so tried to take some into account. It does include an 8K SRAM on board covering the usual 200-7FF but also allows an additional 4K page controlled by A12 of the 8060 address bus, which obviously needs to be paged by the SW. We'll see how usable that is in due course. It also allows all the control lines from the INS8154 to be set using dip switches instead, so useful for testing but also if you don't have the I/O chip fitted.

This evening I'll sort out some images and more detail on the card's function. Also got some photos of the FPGA development card outputting characters from the DM8678 so the firmware is fairly mature too. It's been difficult getting 64 way edge connectors so I've got a 84 way to cut down, but also a few 64's on order from China (of course
Realtime is offline  
Old 31st Aug 2022, 2:18 pm   #4
Timbucus
Octode
 
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
Default Re: Yet another MK14 vdu card

Yes welcome - I was one of the ones who bought the RAM Expansion/Tape Interface and a the 3D printed Keyboard overlays - thanks for helping keep the MK14 alive!

Interesting approach with the CPLD and the intention to have replacement DM8678's - a nice compliment to the software approach on Karen Orton's video card and one that can hopefully be very close electrically.

Do you have a chip fab and can you make 74L86 - quite why they made that one pin incompatible escapes me... - as the biggest block to discrete replicas...

You can use the "go advanced" button to upload photos and of course the URL link in either mode to link to hi res ones hosted elsewhere if you wish.
Timbucus is offline  
Old 31st Aug 2022, 7:13 pm   #5
Realtime
Hexode
 
Join Date: Jan 2021
Location: Ashford, Kent, UK
Posts: 318
Default Re: Yet another MK14 VDU card.

Here’s some info on the new VDU module and a few photos.

I wanted to do a fairly traditional interface to the MK14, while adding some features and flexibility. I decided a Xilinx 5V tolerant CPLD (XC9572XL) would be used to contain all the VDU logic, include on-board memory that fills the gaps in the normal MK14 memory map and host the DM8678 character generator. I've used a variant of this CPLD on my ZX80 Slow Mode adapter module so was confident it would have enough resources.
Photo 1 shows an FPGA development module hooked up to an actual DM8678 to do some confidence testing. What I had forgotten is that my 8678 is a BWF version with a 7x9 font (not the CAB version with a 5x7 font). So the first mod was to add a switch for either font format and trick the memory addressing into adding additional font lines for each character, without affecting the character addressing to the MK14 or the video timing.
Photo 2 shows the full screen 7x9 display and photo 3 a comparison with 5x7 mode (but still using the BWF chip of course so character symbols are cropped). So looking good, just a few video blanking timings to sort out. Obviously 7x9 mode results in only 23 character rows instead of 32 due to the extra display lines per character and I allowed for 2 blank lines between characters to make it more readable. There’s no connection to the MK14 (don't have a connector !!) at this point but firmware simulation shows that the memory access / nENIN timing should be good – time will tell.

The graphics mode currently just takes a value from some on-board switches and repeats the data pattern across the screen, but demonstrates basic operation and timing.

Photo 4 shows a 3D view of the finished module which is about 100x50mm. I’ve allowed for 2 x 10 way DIP switches. SW2 connects to the MK14’s INS8154 Port B and allows either 8154 control of the VDU signals or to be isolated from the 8154. When isolated (or if an 8154 isn’t fitted) SW1 allows for manual control of the signals with a pull up or switch to GND. SW1 also controls the 5x7 / 7x9 mode and allows different TOPPAGE control options. Photo 5 shows what the PCB looks like. The CPLD is a 64 QFP package so a bit fiddly to solder but I’ve done finer pitch parts in the past by hand

Hoping the boards will be here by the weekend and then it's full steam ahead.

Quote:
Originally Posted by Timbucus View Post
Yes welcome - I was one of the ones who bought the RAM Expansion/Tape Interface and a the 3D printed Keyboard overlays
Thanks Timbucus (and all the other forum members) for supporting me and future circuit designs! I generally figure that if I'm willing to commit to a PCB then there will be enough interest out there to order 10 off blank PCBs. The Tape Interface has now passed 15 modules shipped .

You must have one of the early font legends for the Keypad so if you want the updated one let me know.
Attached Thumbnails
Click image for larger version

Name:	1 - FPGA Prototype.jpg
Views:	122
Size:	38.6 KB
ID:	264021   Click image for larger version

Name:	2 - DM8678BWF Mode.jpg
Views:	143
Size:	97.3 KB
ID:	264022   Click image for larger version

Name:	3 - 5x7 vs 7x9 modes.jpg
Views:	129
Size:	76.3 KB
ID:	264023   Click image for larger version

Name:	5 - 3D View annotated_.PNG
Views:	125
Size:	132.9 KB
ID:	264024   Click image for larger version

Name:	7 - PCB in production.PNG
Views:	109
Size:	181.9 KB
ID:	264025  

Realtime is offline  
Old 1st Sep 2022, 5:54 pm   #6
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Yet another MK14 vdu card

Good work so far, and we look forward to seeing the updates as things go along. All of this CPLD type stuff is completely over my head, I know it has been around for a while but I've never had the opportunity or the toolchain to delve into it so the apparent ability to recreate whole systems, even whole computers on a single chip is a completely dark art to me.

One thing I do notice is that the edge connector mounting holes are oriented for a through hole connector, so presumably the intention is that the board will stand vertically rather like a ZX RAM pack or similar?
SiriusHardware is offline  
Old 1st Sep 2022, 6:56 pm   #7
Realtime
Hexode
 
Join Date: Jan 2021
Location: Ashford, Kent, UK
Posts: 318
Default Re: Yet another MK14 vdu card

The PCBs have arrived. Much excitement in the house (well, I was excited ). Next step is to fit the CPLD to the VDU card and enough components to test the JTAG programming chain. If that works then tomorrow I'll be populating the rest of the card. The DM8678 PCB is put to one side for now - I'll start a new thread for that.

On CPLDs / FPGAs: it's really not that much different to learning a language like C. You can achieve an awful lot with a basic knowledge of the language (VHDL in my case). And as with PICs etc, there are a range of fairly cheap development modules out there that make getting started so much easier. Happy to talk tools and devices that I'm using if you want more.

Yes, good spot. The PCB will be at right angles to the MK14. This is mainly because that's the only connector I could source, but I think if you could find a right angled version of the 64 way connector it should be the same pad pitch and then sit horizontal.
Attached Thumbnails
Click image for larger version

Name:	PCBs 01_09_2022.jpg
Views:	70
Size:	30.3 KB
ID:	264084  
Realtime is offline  
Old 1st Sep 2022, 7:50 pm   #8
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Yet another MK14 vdu card

For our sins, we have a multitude of (well, at least two) different methods of connecting peripherals to our MK14s here. I am old school like you and use push-on edge connectors despite their tendency to skid sideways and connect the wrong things to the wrong places, there being no 'key slot' on the MK14 rear edge connector. That's mainly because I have a real MK14 as well as an issue VI replica, and I like to have the option to connect any extras I make or obtain to that as well, but I wouldn't want to be soldering anything to the edge connector on that.

Tim and others have soldered a DIN female 'card rack' connector to their issue VI rear edge as they are easier to still find, and a lot better at staying connected in the hoped for manner. On their Ortonviews, etc, they have soldered a corresponding DIN male card rack connector. This will unfortunately make it difficult for them to evaluate your project without making some kind of adaptor / changer.

To try to please everyone perhaps the most universal kind of connection to provide on any future revision would be double sided edge contact fingers like the ones on the MK14 rear edge, because you can (with some ingenuity) fit a wide range of double row connectors to those by sliding the connector over the edge with one row of pins lying along the top contact fingers and one row of pins lying along the lower contact fingers and soldering them on.

That would be some way into the future however, we are all very keen to see this up and running.
SiriusHardware is offline  
Old 1st Sep 2022, 9:01 pm   #9
Timbucus
Octode
 
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
Default Re: Yet another MK14 vdu card

Indeed I did that with the DIN connector as of course the SOC VDU was always pictured with one on (as a Eurocard design) so it seemed natural to do that on the MK14 but, as you say not a good idea for £800-£100 vintage computers to have one.

One solution could be to have that hole footprint available as well so it can be fitted as an alternative or as you say a small conversion PCB should not be hard to fabricate as of course I will want one of these as well.
Timbucus is offline  
Old 1st Sep 2022, 9:55 pm   #10
Realtime
Hexode
 
Join Date: Jan 2021
Location: Ashford, Kent, UK
Posts: 318
Default Re: Yet another MK14 vdu card

Quote:
Originally Posted by Timbucus View Post
Do you have a chip fab and can you make 74L86 - quite why they made that one pin incompatible escapes me... - as the biggest block to discrete replicas...


It should be possible to come up with a replacement for it in the same footprint. The propagation delay is a bit of a problem though. Modern 74LS86 parts have about 10ns delay. The 74L86 had about 50ns. Whether that's important depends on how marginal the timing in the original circuits was. As you know, many short cuts were taken on early computer boards and reliance on propagation delay was probably one of those. Does anyone have a full datasheet for the L86? I only found a pin diagram and output equivalent circuit.
Realtime is offline  
Old 1st Sep 2022, 10:33 pm   #11
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Yet another MK14 vdu card

Quote:
The propagation delay is a bit of a problem though.
We think that expansive delay is the sole reason that device was chosen over the obviously more standard 74LS86, it is there specifically to introduce that delay. Starting from a position of knowing absolutely nothing, I don't know if you could synthesise one by making the equivalent of a 74LS86 followed by a chain of 4x LS buffers on its output to force the required 50ns propagation delay.
SiriusHardware is offline  
Old 1st Sep 2022, 11:32 pm   #12
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,265
Default Re: Yet another MK14 vdu card

Found a link to a spec for the 74L86
http://www.ralphselectronics.com/Pro...I-DM74L86N.PDF

Looks to be same pinout as CD4070B, but at 5v those are even slower than the 74L86
https://www.ti.com/lit/ds/symlink/cd4070b.pdf
Mark1960 is offline  
Old 2nd Sep 2022, 12:44 am   #13
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,265
Default Re: Yet another MK14 vdu card

Quote:
Originally Posted by SiriusHardware View Post
Quote:
The propagation delay is a bit of a problem though.
We think that expansive delay is the sole reason that device was chosen over the obviously more standard 74LS86, it is there specifically to introduce that delay. Starting from a position of knowing absolutely nothing, I don't know if you could synthesise one by making the equivalent of a 74LS86 followed by a chain of 4x LS buffers on its output to force the required 50ns propagation delay.
CPLD and FPGA synthesis would almost certainly optimise away any chain of gates that are an attempt at a delay. It would probably be better to add something similar to the ‘74 circuit that we added to orton view to delay NENIN after NADS.

One nice thing about CPLD is you can change the logic to experiment without modifying the pcb, just connect NADS to a spare pin of the CPLD in case you need it later.

Do the xilink tools support schematic capture, or do you have to use VHDL?
Mark1960 is offline  
Old 2nd Sep 2022, 8:21 am   #14
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Yet another MK14 vdu card

The info about the CD4070B is quite interesting, unfortunately I don't have one here to try as a replacement for the 74L86 in my SOC VDU - I don't know if Tim does? If it worked that would be quite a breakthrough for people trying to put together a clone of the original SOC VDU, although obviously not relevant to Realtime's project here.
SiriusHardware is offline  
Old 2nd Sep 2022, 1:25 pm   #15
Timbucus
Octode
 
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
Default Re: Yet another MK14 vdu card

I have one 4070 in the parts I have put together for my SOFTY according to my inventory spreadsheet. Not sure if it is a B I will dig it out and take a look. My usual suppliers seem to have a mix of 4070 but not B; specifically 'not mentioned', BE and BCN so may be a try and see.
Timbucus is offline  
Old 2nd Sep 2022, 2:35 pm   #16
Mark1960
Octode
 
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,265
Default Re: Yet another MK14 vdu card

I think the pinout between variations of the 4070 would be the same. Its a common pinout for 4000 series gates. I didn’t choose B for any reason other than the first 4070 spec found in a search. I think all standard cmos will be too slow at 5v supply and watch out for any inputs driven by ttl input levels.
Mark1960 is offline  
Old 3rd Sep 2022, 4:28 pm   #17
Realtime
Hexode
 
Join Date: Jan 2021
Location: Ashford, Kent, UK
Posts: 318
Default Re: Yet another MK14 vdu card

Here's a couple of photos of the built VDU module. I've got to say that it's all going remarkably well. In truth, that's probably because I did a lot of simulation of the VHDL whilst waiting for the PCBs to be delivered. Apart from a few errors in the positioning of the video there wasn't much to change. I did have problems with NENIN and nRAMCS which I'll cover in another message.

Incidentaly, in the first photo I've got one of Coolsnaz2's fast loader M5Stick-C devices connected. Absolutely brilliant piece of kit. It's attached to my slow MK14 tape interface with a switch to be able to select which one is connected to the SIN and SB lines.

In the second photo you can see that T1 (the video output buffer transistor) is fitted to the reverse of the board. Why? Well, that's another case of changing the pin out of a completely standardised component. I laid out the PCB for a 2N2222A. I purchased P2N2222's. Turns out this part has the emitter and collector swapped over. Why would you do that . Simple fix to either rotate it of change board sides but it took me a while to figure it out.

Another thing I had not taken into account was the fact that the HMI-6561B RAMS on the MK VI are latched and so not compatible with the original MK14 VDUs approach of keeping nRDS active for the whole video line. As a result only the first character gets accessed and then repeated across the line. Somewhere in the back of my mind I knew I had read something about that on an earlier forum post. Sure enough it was confirmed. So the VHDL has been updated to gate nRDS and pulse it just before the DM8678 loads the data. It now works for both latched and standard non-latched RAM.

I have uploaded a short video to YouTube for your amusement

https://youtu.be/PLZfzgN6mwM

The memory at 0x200 is being written to by the Tape Interface. I use a 5" portable TV. Cost about £25. it's very compact and doesn't need much bench space.

There are still some minor things to fix regarding video positioning and start of blanking. Also I will create a proper back porch for the video so that black video on a white background works correctly (it doesn't currently). A few tweaks needed on the graphics mode before I post a video of that.

I haven't tested the INS8154 control of the board yet - it's all set up via the DIP switches right now.
Attached Thumbnails
Click image for larger version

Name:	MK14 with VDU card.jpg
Views:	97
Size:	56.0 KB
ID:	264178   Click image for larger version

Name:	MK14 with VDU card2.jpg
Views:	97
Size:	106.1 KB
ID:	264179  
Realtime is offline  
Old 4th Sep 2022, 11:58 am   #18
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,485
Default Re: Yet another MK14 vdu card

You have been busy!

Yes, that strange quirk of the IM6561 RAMs not liking the constant-low read signal from the SOC VDU came to light when Tim tried to use his Czech SOC replica with an issue VI fitted with IM6561s. Karen arranged for Ortonview to 'flap' the read signal as well, once we had made her aware of it. In truth there was an easy workaround, if the VDU also included RAM at 0200-07FF as yours does, and as Slothie's PCB for Ortonview does - just use that RAM area as the screen RAM, which you would be more likely to do because it's better for the screen RAM to be 512 consecutive bytes anyway.

I think that because people are very likely to be using this with a modern screen, the video signal does need to be properly formed with a 'back porch' so that the display can correctly measure the black level and set the overall brightness of the screen accordingly. You are no doubt aware of this being a problem with ZX81s with early ULAs which did not generate a back porch, although the final version ZX81 ULA did generate a back porch.

Regarding transistors and pinouts, I share your pain, I never understood why it was necessary for there to be standard and 'L' versions of common BCxxx transistors, the 'L' having a different pinout.

Enjoyed the video - I thinks it's probably time we all started writing some new software to run on the MK14 VDU. I remember a contemporary review of the MK14 and VDU suggesting that it would be useful to be able to keep notes on tape and load them in from tape whenever you wanted to read them. That was about as ambitious as it got, back in the day.
SiriusHardware is offline  
Old 4th Sep 2022, 6:15 pm   #19
Realtime
Hexode
 
Join Date: Jan 2021
Location: Ashford, Kent, UK
Posts: 318
Default Re: Yet another MK14 vdu card

A bit more progress. I had to do a little happy dance when I saw this

https://www.youtube.com/watch?v=Z01KtR6k57Y

On the subject of SW, I've started to write a simple Invaders game - I expect someone has already done that but it seemed like a fairly straightforward game to try. The last time I coded in assembler was Z80 in the early 80's and I certainly don't consider myself a programmer. It's taking me a while to get my head around the 8060 but the attached photo shows the start up screen created by a bunch of nested loops. No game action though. It's on the back burner while I continue with the VDU card.
Attached Thumbnails
Click image for larger version

Name:	Invader Splash.jpg
Views:	86
Size:	71.1 KB
ID:	264250  
Realtime is offline  
Old 4th Sep 2022, 6:37 pm   #20
Timbucus
Octode
 
Join Date: Mar 2019
Location: Barry, Vale of Glamorgan, Wales, UK.
Posts: 1,362
Default Re: Yet another MK14 vdu card

No wow nice start - the only game we know of other than TANKS is Pong and we do not have a copy - I have asked a few times on the forum of the guy who wrote it but, it seems to fall on deaf ears.
Timbucus is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 12:10 pm.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.