16th May 2021, 5:46 pm | #181 | |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,298
|
Re: The Transam Triton Personal Computer
Quote:
|
|
16th May 2021, 8:21 pm | #182 |
Pentode
Join Date: Mar 2021
Location: Liphook, Hampshire, UK
Posts: 125
|
Re: The Transam Triton Personal Computer
So should I leave the encoder output pull down resistors going to -12v as they were in the original design or change the to pull down to 0v ?
|
16th May 2021, 9:32 pm | #183 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,298
|
Re: The Transam Triton Personal Computer
I took a look at the TI data sheet for 74ls244.
Input low current is max of 0.2mA at 0.4v. So that would estimate 0.54v max input voltage with 2.7k pull down resistors to 0v. This is worst case spec so I think you are safe with 2.7 k pull down to 0 v. The spec also shows Vik of -1.5v at 18mA, but is not really clear if this is ok for every input simultaneously. 2.7k pull down to -12v would be approx 3mA for each input protection diode. As you already had a problem with failure of the 74LS244, and we know from the PET thread that these seem to have reliability issues, then maybe the safest option is to pull down to 0v instead of -12v. |
16th May 2021, 10:04 pm | #184 |
Pentode
Join Date: Mar 2021
Location: Liphook, Hampshire, UK
Posts: 125
|
Re: The Transam Triton Personal Computer
Would the out put of the encoder ic look like this ?
From Wikipedia, The p-type MOSFETs are arranged in a so-called "pull-up network" (PUN) between the logic gate output and positive supply voltage, while a resistor is placed between the logic gate output and the negative supply voltage. The circuit is designed such that if the desired output is high, then the PUN will be active, creating a current path between the positive supply and the output. |
16th May 2021, 10:12 pm | #185 | |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,444
|
Re: The Transam Triton Personal Computer
Quote:
(Although, if keyboard encoder IC had required pull-down to -12V / Outputs did go down to that, then could also inserted a 1N4148 diode in line with each output and fit pull-downs to 0V on 74LS244 inputs, to ensure there was no negative voltages fed to it). And you could always fit a 74HC244 in place if the 74LS244, if low-voltage input-level was a bit too close to 74LS244's threshold - Although in practice that will actually be somewhere in the middle of the spec'd input-low and input-high levels so should be OK for a one-off at room temperature and worst-case figures won't need to be worked-to. Last edited by ortek_service; 16th May 2021 at 10:40 pm. |
|
16th May 2021, 10:40 pm | #186 | |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,444
|
Re: The Transam Triton Personal Computer
Quote:
- Although may well have been designated as Vss on PMOS devices, which can be confusing when you're only used to dealing with NMOS / CMOS and TTL (Which being bipolar junctions based, tends to use Vcc and Gnd as Vss isn't really applicable on non-MOS devices). And it's even more confusing, when circuits are drawn with most-negative supply at the top rather than the bottom, as the conventional is to have the highest voltages at the top and lowest at the bottom. However, I think the IC has probably got open-drain outputs, if it requires the external resistors. - So maybe just the PMOS FET with source connected to +5V 'Positive Ground' / maybe Vss on this. But not really sure where the 0V pin comes into this, and maybe the outputs do actually go from +5V to 0V without any external resistors at all. And the resistors were only fitted to make outputs go down to -12V, bypassing any active-load FET's on outputs to 0V. However, if there were internal active-loads, to replace conventional pull-down resistors, they probably would have gone to 12V as it seems 5V supply difference wouldn't have been enough to operate PMOS stages sufficiently. Last edited by ortek_service; 16th May 2021 at 10:48 pm. |
|
17th May 2021, 1:46 am | #187 |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,444
|
Re: The Transam Triton Personal Computer
Further to the above, I found this useful webpage showing PMOS stages with active PMOS loads, rather than fixed resistors - Although they too have used positive ground and put negative rail at the top.
https://www.electrical4u.com/nmos-and-pmos-logic/ - Plus there's a few errors: Circuit (b) on first diagram should say PMOS, not NMOS And on PMOS diagram Circuit (a) should have -Vdd, as (circuit (b), - Not +Vdd |
17th May 2021, 1:38 pm | #188 | |
Pentode
Join Date: Mar 2021
Location: Liphook, Hampshire, UK
Posts: 125
|
Re: The Transam Triton Personal Computer
Quote:
so it is strange that the pull down risistors on mine go to -12v, still can not find any info on the encoder chip SW-20457k. |
|
17th May 2021, 10:38 pm | #189 | |
Pentode
Join Date: Mar 2021
Location: Liphook, Hampshire, UK
Posts: 125
|
Re: The Transam Triton Personal Computer
Quote:
I expecting my parts from Hall sensors from DSM components tomorrow based on the tracking information, but now taking a closer look at the data sheet for the mh3sd2 mh3ss2 these devices pull up. My Microswitch keyboard sensors are Output: A = sink level, B = sink pulse so pull down so now I will have to invert the output from the new devices for it to work. Also now I don't have enough new devises as I was only planing to replace the faulty ones, I don't think it will be possible to mix the pull up and pull down devices ? |
|
18th May 2021, 1:01 am | #190 | ||
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,444
|
Re: The Transam Triton Personal Computer
Quote:
- And looks like it was quite-useful in drawing yours up with similar naming conventions. The write-up mentions a 'MOS' device, and this usually refers to a CMOS device these days. But it seems like they meant PMOS (maybe the main-version back than), as it also has a -12V supply to the IC. Plus I also noticed that this schematic has this IC with a different '14026' (in-house?) part number (as well as another '14027'? custom IC) I think that particular board used 7404 Inverters, as they needed the opposite polarity for compatibility with what they were interfacing this to. (A Radio-shack ASCII-encoded keyboard PCB, I'd seen, had both positive and negative and negative going outputs, for most versatility) Although these will also provide some buffering - but only on sinking (at least 16mA), whereas encoder IC's outputs are probably better than 7404's only -0.4mA sourcing. However, they have used rather lower than should be required 300R pull-down resistors, for some reason (seems a bit unnecessarily wasteful on current, but maybe only for short durations). If the IC's internal PMOS output transistors are open-drain, then it shouldn't really matter what voltage you pull these down to - as long as it's less than +5V but not below max -Vds rating of these MOSFETs. And I guess whatever your keyboard came from, did originally require -12V 'Logic-low' level Last edited by ortek_service; 18th May 2021 at 1:13 am. |
||
18th May 2021, 1:29 am | #191 |
Octode
Join Date: Mar 2020
Location: Kitchener, Ontario, Canada
Posts: 1,298
|
Re: The Transam Triton Personal Computer
From badly faded memory the pull down resistor for a normal TTL input was always 470 ohm for a single input. This was why designs tried to avoid pull down resistors, and the habit stuck for LSTTL. So 300 ohms is not that much lower than expected but still a surprise that PMOS could drive that load.
I’m guessing that before NMOS was introduced PMOS was just MOS. |
18th May 2021, 3:41 am | #192 | ||
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,444
|
Re: The Transam Triton Personal Computer
Quote:
Unfortunately, I don't think these pull-up type Hall-sensors will work with your keyboard's encoder IC, as it looks like that IC must have internal pull-ups to work with original pull-down type. So even changing all to this type is unlikely to work. It the outputs on these Hall-sensors had been analogue-switches then there could have been a fairly simple dodge of connecting their +V to 0V and their ground to a -5V rail, generated off the -12V. (although some re-connection of their supply pins would be needed) So it looks like the only option is to add some open-collector / drain inverters to the 2 outputs of each of the replacements. And might be able to do this with a single built-in resistors NPN 'digital transistor' - aka Bias Resistor Transistor (BRT) / Resistor Equipped Transistor (RET) - in a suitable small Surface Mount package, that could be fitted to the modules to save having to cut tracks on the keyboard PCB's, as presumably you'd want to avoid doing that. Something like a Nexperia PEMH9 (SOT666 0.5mm pitch) / PUMH9 (SOT363 0.65mm pitch / PIMH9 (SOT457 0.95mm pitch) - Datasheet: http://www.farnell.com/datasheets/1794538.pdf https://www.nexperia.com/products/bi...H13_PUMH13-374 https://uk.farnell.com/nexperia/pumh...38599?st=pumh9 https://uk.farnell.com/nexperia/pemh...3?st=PEMH9,115 etc. You can get thin self-adhesive rows of track strips boards, to mount these onto: 0.65mm pitch: http://www.farnell.com/datasheets/1909561.pdf - Although Farnell no longer stock this (but RS do). But quite expensive if you only need some small bits, so could probably find a cheaper one / a suitable package breakout board |
||
18th May 2021, 5:50 am | #193 | |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,444
|
Re: The Transam Triton Personal Computer
Quote:
(I must have spent too many past decades designing with 74HC etc). It seems this is 1.6mA on 74xx vs still 0.4mA on 74LSxx (So rather higher than the Input-high current of < 40uA), and just as imbalanced as TTL's Max Output sink & source currents It seems CMOS is actually surprisingly-old, being invented as long ago as 1963: https://www.computerhistory.org/siliconengine/timeline/ - a year after DTL-logic, but a couple of years before the DIL-package! But it wasn't until 1968 that the 4000(A) CMOS logic was released. Only a year after PMOS logic, according to: https://en.wikipedia.org/wiki/Logic_...S_(CMOS)_logic And although CMOS allowed much more complex MSI / LSI circuits in the same size than TTL, it seems it was a bit too slow so presumably why PMOS was initially used. Then (from around mid 70's?) NMOS replaced that, and tends to be faster / lower-voltage and around 3x smaller than PMOS for same Rds(on)), with it being used for a long time in LSI's etc. until High-Speed CMOS was developed in the early 80's. So it seems both CMOS & PMOS were around quite a while before these Keyboard encoder IC's. Last edited by ortek_service; 18th May 2021 at 5:57 am. |
|
18th May 2021, 2:22 pm | #194 | |||
Pentode
Join Date: Mar 2021
Location: Liphook, Hampshire, UK
Posts: 125
|
Re: The Transam Triton Personal Computer
Quote:
Yes I did use the circuit diagram of the Diablo 1620 Hy TERM as a blueprint to draw up my keyboard circuit diagram, I thought this is a good technical manual. The https://www.computerhistory.org/siliconengine/timeline/ is a good find its amazing how far we have moved forward from the 1970 devices. Its also interesting how many vintage devices like the Diablo used intel 8080 processors like the Triton single board computer ,the wiki page on the 8080 states the following:- "The 8080 was designed for almost any application except a complete computer system. Hewlett Packard developed the HP 2640 series of smart terminals around the 8080. The HP 2647 is a terminal which runs the programming language BASIC on the 8080." I worked on a number of Centronics printers in the early 80's and they used intel too. (Centronics model 704 8080A, model 352 8085A). The 8080 appeared in 1974 and is NMOS implementation its forerunner 8008 was PMOS. Like ortek_service post "Vintage ASCII-Encoded Keyboard projects - Using standard logic IC's / just +5V" there is a whole bunch of vintage Keyboards out there, must also be the same for printers, I had a Epson MX80FT connected to my Triton which was good at the time given the display on the Triton was on 64 chrs by 16 lines. |
|||
19th May 2021, 12:25 am | #195 | ||
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,444
|
Re: The Transam Triton Personal Computer
Quote:
Plus the 8080 was really Intel's forerunner Microprocessor to the 8086 / 8088 etc - even if they did require a new CP/M-86. So I'm not sure why they thought the 8080 wasn't designed for computers, when Intel etc. hadn't really got anything better at that point. And the rival 6502 / 6800 ones were maybe much simpler,with few registers / instructions - Although that did make these able to run these instructions mostly at the full main clock speed, and did inspire Acorn to invent the ARM rather than going down Intel's ever-more complex & power-hungry 80x86 CISC path. Quote:
I do recall the MX80 - which always seemed to be much less-common than Epson's infamous FX80, RX80, then LX80 (I've still got) etc. Luckily most printers were either Centronics or RS232 / both. And were often switchable between Epson & IBM compatibility, although both early & later ones had more proprietary interfaces often being directly-controlled by the computer. But I think old printers are the poor relation to vintage computers, and not really achieved any great collectability status - Maybe because there's so much better ones for printing these days, if required. Although that does probably mean they can be picked-up quite cheap - assuming they haven't mostly all been thrown-out so now making these much rarer. And getting replacement ribbons etc. that still work might be challenging (although you can re-ink these to a certain extent) |
||
19th May 2021, 2:00 pm | #196 | |
Pentode
Join Date: Mar 2021
Location: Liphook, Hampshire, UK
Posts: 125
|
Re: The Transam Triton Personal Computer
Quote:
The other component I expected was a MH3SS2 which is a source level hall sensor by Tesla, but I did not read the DMS web site correctly and ordered 18 of these Schmitt trigger which is MH3ST2 not as I thought MH3SS2 from the data sheet MH1ST1 and MH3ST2 its just a single Schmitt trigger. I connected one up as per the diagram in the middle of the data sheet and the thing just heated up so that was wrong. I assumed the package diagram in the data sheet is looking at it from the bottom, this is true for the other sensor (MH3SD2). So there is no hall sensor so is it strange to have a single Schmitt trigger in this package ? DMS components don't do the part I wanted (MH3SS2) but I do have enough of the original level sensors to get a working keyboard just missing two so no break and rpt keys. The is one slight problem the new sensors are just slightly to fat to fit the key switch modules so may be some filing !? |
|
19th May 2021, 11:10 pm | #197 |
Pentode
Join Date: Mar 2021
Location: Liphook, Hampshire, UK
Posts: 125
|
Re: The Transam Triton Personal Computer
I posted a new user group news letter on my site triton-docs/user-group the reason I was doing more logic check on the Triton PCB and found an extra IC see included photo.
On Page 3 the newsletter talks about "Stop that Flashing" the circuit in on page 4 which includes an option for two graphics roms and double your tape speed. Thats my extra IC and its installed with a change the line which comes from IC53 in the newsletter is from pin 12 on mine its pin 9. I know some people had problems getting the Invaders game to run on real HW. Its run on the emulator, from the vids I have seen, it looks like a VDU issue this mode is to stop the snow effect when writing continuously to the VDU map memory with machine code programs. Did any on else implement this mod ? |
20th May 2021, 6:02 am | #198 | |
Octode
Join Date: May 2018
Location: Northampton, Northamptonshire, UK.
Posts: 1,444
|
Re: The Transam Triton Personal Computer
Quote:
- Although at least this did summarise most quite well: https://telcontar.net/KBK/TESLA/Hall_ICs There only really 3 pages (158-160) covering all of these in their databook, with index at: http://www.teslakatalog.cz/io.php With multiple-types being grouped together on each page, without clearly distinguishing the differences. So basically all of these are: http://www.teslakatalog.cz/io0/158.gif: (IO-52 (/51) 13mm? Height package) MH1SD1 www.teslakatalog.cz/MH1SD1.html (158.gif) - IO-52 Dual emitter-follower O/P's (20-1000us Pulse-Type) MH1SS1 www.teslakatalog.cz/MH1SS1.html (158.gif) - IO-52 (51) Dual emitter-follower O/P's (Level-Switched Type) http://www.teslakatalog.cz/io0/159.gif: (IO-53 (Low 7mm Height package) Also: https://www.bucek.name/pdf/mh3sd2,mh3ss2.pdf MH3SD2 www.teslakatalog.cz/MH3SD2.html (159.gif) - Dual emitter-follower O/P's (20-1000us Pulse-Type) MH3SS2 www.teslakatalog.cz/MH3SS2.html (159.gif) - Dual emitter-follower O/P's (Level-Switched Type) http://www.teslakatalog.cz/io0/160.gif: (IO-51 (13mm) / IO-54 / IO-53 (7mm) package) MH1ST1 www.teslakatalog.cz/MH1ST1.html (160.gif) - IO-51 Single emitter-follower O/P's. Hall-Probe? + Ext. I/P. Schmitt?? MH3ST2 www.teslakatalog.cz/MH3ST2.html (160.gif) - IO-54 Single emitter-follower O/P's. NO Hall-probe. Ext I/P. Schmitt Flip-Flop MAF100 www.teslakatalog.cz/MAF100.html (160.gif) - IO-53 Dual 'Semiconductor Generator' Hall-effect? raw sensor O/P's?? For some reason, they chose to use different pin numbers on the MHxSTx circuitry and the package. But should still end-up with supply at end of it. Whereas the MAF100 (raw Hall IC sensor only?) has one in a different position. I think the MH3ST2 is designed to be used with an external hall-sensor or maybe a mechanical contacts key / reed switch. But the MH1ST1 appears to have a built-in Hall? sensor. And you can buy these from Romania for not too much each / postage, via eBay-UK. (Although are higher MH1xx 13mm rather than the MH3xxx 7mm low profile package) As you need to invert the output's, then you don't actually need 2-output ones. Although very little room if only got 7mm inside switches (I thought the original substrates looked longer?) I did find very-small 1 x 1 x 0.5mm package ones, with 'Latched' output (appears to be a fixed-length pulse of various orderable values). https://www.ablic.com/en/doc/datashe...cs/S5725_E.pdf And output goes low (like most types) - Although only 1 output, so actually more awkward to add extra NPN transistors to make 2 outputs - But could probably make a small PCB to contain everything in the space of an original, making a completely pin-compatible replacement. And I thought maybe they do non-latched types as well - But not compared magnetic sensitivities / worked-out if polarity might cause issues. However, there's a large comparison table here: https://www.ablic.com/en/semicon/pro...ism-sensor-ic/ - that shows whether they are omnipolar / bipolar etc. (But not whether 'latched' pulsed type?) Last edited by ortek_service; 20th May 2021 at 6:20 am. |
|
20th May 2021, 1:29 pm | #199 | |
Pentode
Join Date: Mar 2021
Location: Liphook, Hampshire, UK
Posts: 125
|
Re: The Transam Triton Personal Computer
Quote:
Thanks that a lot of good information and this site http://www.teslakatalog.cz/io.php is a good find, I was looking for information on these chips MH103 and MH113 from Tesla, some information here telcontar.net TESLA encoders and DSM components have them MH103, keyboard encoder, DIL40 and MH103A, this might be a good replacement for my encoder and work with the MH3SD2 and MH3SS2 but is not listed. I connected a MH3ST2 the other way round to the data sheet and you are correct the 0V and supply are at either end. and it works. |
|
20th May 2021, 7:10 pm | #200 |
Pentode
Join Date: Mar 2021
Location: Liphook, Hampshire, UK
Posts: 125
|
Re: The Transam Triton Personal Computer
Maybe julie_m is on to something with Enhancing PETs and other Vintage Computers given the mission to get my Triton going what about my plan B just need to interface keyboard to the Raspberry PI and some work on the woodwork job done. (see photos)
|