OT - Angelo's Blincolnlights 18!

[Andy]

Looks like a lot of fun!   Blincolnlights 18 | Hackaday.io

Once I figure out how to get a board made from the kicad I am going to give it a shot :-)

Andy

[pbi...@gmail.com]

Just zip the gerber directory and submit the zip file to JLCPCB.  5 boards for $15.30, plus shipping to your location.  In my case that adds $14.92 for “Global Standard Direct Line”.

--
You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/e9b69ec2-b99a-4114-b815-c7680eec2315n%40googlegroups.com.

[a...@papnet.eu]

It *is* a lot of fun. My first PCB! If there's enough interest I'd like to sell it as a kit.

Also want more simulators to use with it :)

[Lars Brinkhoff]

10/10 would buy^H^H^H get a free sample again.

[Andy]

ordered!

[Andy]

Hey Angelo - could you give me a few tips regarding the BOM?

1. What sort of Pi would fit the bill for this?

2. What spec 5mm LEDs (voltage?)

3. what 1N4148 diodes? - there seem to be a million different flavours on mouser

cheers!

Andy

[Clem Cole]

below.... 

On Mon, Jul 31, 2023 at 5:07 PM Andy <lordha...@gmail.com> wrote:

Hey Angelo - could you give me a few tips regarding the BOM?

I'll try to take a stab given his Schematic  [my PC Boards have been ordered]

1. What sort of Pi would fit the bill for this?

Anything that will run a current version of OpenSIMH and has a 40-pin header AND runs Raspian - assuming Oscar's code to drive the GPIO lines.  Anything after the Raspberry Pi 2 Model B should be fine.   I would recommend a RP 3 if you have one, but I have an RP2 driving a couple of Oscar's kits, and they are fine.

I'm experimenting with both a LePotato and BananaPI. However, Oscar's code will not work due to the difference between the 'official' GPIO library for Linux/Debian and what RPI provides. I am examining a more portable solution so people have more options. 

2. What spec 5mm LEDs (voltage?)

Given the use of a 390 ohm pull up and a UDN2981A [which is what is in the schematic and what Oscar uses], and a TD6283A driver [which Angelo mentions in his BOM], I don't think it should matter other than aesthetics.  Either chip should drive any reasonable colored LED.

3. what 1N4148 diodes? - there seem to be a million different flavours on mouser

PC Board shows standard thru-hole, not SMT.  Any generic 1N4148 or 1N914 should work just fine for what he is doing.   Just need small ones that you can bend the leads to fit in the holes.

ᐧ

[Andy]

thanks Clem!

So is this based on SIMH? From the github link I was not clear

cheers

Andy

[a...@papnet.eu]

No, not based on simh. The simulations are written in verilog currently and translated to c++ with verilator. This is somewhat nice because transcribing schematics is fairly natural with verilog (of course i wish i had actual TX-0 schematics, but that's a different story). Unfortunately however the simulation is much too slow this way! i get around 25μs per cycle on a raspi 3 for the pdp-1 when the real machine had 5μs cycle time. Also my IO is currently too fast (i was lazy)....so this is only a stopgap solution. The plan is to write another simulation straight in C (would be nice if it ran well on a raspi zero) and also hook up an fpga for super-duper-turbo-accurately timed simulations.

Cheers,

Angelo

[Lars Brinkhoff]

I'm also writing a simulation for Angelo's Blincolnligts-18: a PDP-9.
It's pretty much a hand-coded translation of the schematics into C,
and it does run reasonably fast even on a Raspberry Pi 3.

To reiterate, Angelo's board is similar to, but not exactly like
Oscar's. And there is so far no SIMH interfacing, but presumably that
could be done.

[Heinz-Bernd Eggenstein]

As for SIMH Interfacing:

I wonder whether it would make sense, for future blinkenlights projects, to delegate the low level interfacing of LEDs and switches to a microcontroller like the Raspberry Pico, sitting on the blinkenlights PCB, and let any entity that wants to interface with it (SIMH on a Raspi, FPGA emulators, whatetver) talk to that microntroller over a standard hardware interface like I²C or SPI with a simple API (to update  LED brightness values  and get notified when switches change).

The question is whether these inexpensive yet easy to program microcontrollers have enough computing power to handle all the low level tasks which would ideally include things like the "Incandescent Lamp Simulation" of the PiDP-8 (mimicking the after-glow of light bulbs with LEDs) and software-debouncing of switches , in addition to  the mere mapping of blinkenlight status to GPIO signals.

I have a feeling (not more than that at the moment) that this could make it easier to adapt legacy simulator software that was not written with the intention to drive "real" front panels.

Thoughts?

[Ian Schofield]

Good thought about off loading the panel drive to a pico. However, a Raspberry Pi has got enough ooomph with 4 cores to do all of the work. From the hardware point of view, simplifying the panel is preferable and, to this end, see Neopixel suggestion.

The attached (rubbish!) video is a test of a Neopixel array (16x10) driven by a pico (behind the display) running OS/8. The LEDS are driven by a continuous loop in the 2nd core at about 120Hz. (PC at the top, then MB, ACC, MQ). Larsen.PA.

A Pi4 will have no trouble with this. Equally, switches etc. can be managed the same way with a set of shift registers driven by 3 wires (sample,clock,output). This method does not require address decoding etc.

Quite rightly you point out the interface to Simh. For realtime displays of processor state etc., there are a number of issues. Given this, I do wonder if the simulation code needs a rewrite (as with the code in this Pico).

This is no easy task to end up with cycle exact timing......

Vide0:https://drive.google.com/file/d/1jIeQyJymHO2d2aFC1ZmT9LDCbOYxvcn6/view?usp=sharing

[Andy]

soooo thinking out loud here... while I wait for my components and the board to arrive, I was thinking it would be nice to make a cover for the panel. SOmething simple, like holes in plastic for the LEDs and switches, AltairDuino-style.  Now I was pretty awesome at uploading the Kicad files to JLCPCB.. .any thoughts on the level of effort to take that kicad and make a simple cover?  Presumably in some sort of plastic or aluminium using the same component positions from Angelo's board.  Having the switch names and LED numbers printed as well would be a bonus.

thoughts appreciated!

Andy

On Friday, July 21, 2023 at 9:10:14 AM UTC-4 pbi...@gmail.com wrote:

[Chris Smith]

Re: "A Pi4 will have no trouble with this. Equally, switches etc. can be managed the same way with a set of shift registers driven by 3 wires (sample,clock,output)."

Since we are talking about unloading operations here, why reinvent the wheel for the switches? There are USB arcade encoders which directly support between 14 and 20 switches (depending on model). The libraries are already there because these are used with a Pi4 for game emulators - and since those emulators often want two controllers, that already puts us in the 30 switch range with off the shelf plug in hardware.

The encoders are designed for game switches so they already need full switch independence and low latency. A normal keyboard encoder would manage the latency, but the typical matrix scanning might encounter limits trying to encode stateful front panel switches.

Between addressable LEDs for "output" and USB game encoders for "input" we might be close to avoiding the need for custom PCBs to create the workings of front panels.

... Chris 

--

You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/6a92b7f7-ce44-4413-96eb-2340a17a6d79n%40googlegroups.com.

[Heinz-Bernd Eggenstein]

Not sure I'd want to give up USB port(s) for reading the switches, tho. A couple of 74HC165 parallel-to-serial shift registers would be enough tho. And this would work for interfacing a blinkenlights PCB to any hardware, including FPGAs, not just SBCs like the Raspi.
hb

[pbi...@gmail.com]

You might want to study the design of the BlinkenBone, which accomplished this about almost a decade ago: http://retrocmp.com/projects/blinkenbone

 

See http://retrocmp.com/projects/blinkenbone/blinkenbone-physical-panels/170-blinkenbone-hardware-components

BlinkenBoard = 88 outputs, 40 inputs

Schematics and bus specification at the bottom of the page

 

From: 'Heinz-Bernd Eggenstein' via PiDP-10 <pid...@googlegroups.com>
Sent: Monday, August 7, 2023 7:43 AM
To: PiDP-10 <pid...@googlegroups.com>

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/3fca7c34-4801-4d63-8fcc-df7e66a20f2fn%40googlegroups.com.

[Henk Gooijen]

Have a look at MCP23017 and MCP23S17. They are 16 bit (2 8-bit ports), where each bit can be defined as an output or an input.

Interfacing is either I2C or SPI. I used the MCP for the RK05 (60% scaled to match with the PiDP-11/70). One MCP chip is enough for the lamps and switches (and solenoid!) on the RK05
On and off, I am now working on an RK11-C lamp panel (144 lamps, but IIRC, 8 not used).

[a...@papnet.eu]

I thought about this too, I'm just not sure if it's really worth it to spend time on polishing this thing, which was sort of meant as a prototype and a learning experience for me. Having a dedicated panel for the machine you *actually* want to simulate (rather than a compromise) might be a better thing.

So let me ask you this: which machines do you plan to use the panel with?

[Andy]

Yeah of course I am somewhat constrained by my lack of knowledge :-)  but I am quite intrigued with the thought of getting a TX0 and a PDP1-style of panel up and running.  The more the merrier!

[a...@papnet.eu]

Then I have good news for you: Oscar and I have started working on a PiDP-1. For at least myself (but possibly more people) I also want a proper TX-0 panel, but first the PDP-1.

[Ian Schofield]

Hi all,

Panel for the PDP1/TX-0 .... brilliant! And, most relevant to this discussion.......

 Just a few more comments here which are really aimed at some future project..

The Blinkenbone board is a very impressive bit of hardware design and I think would tick boxes for an industrial controller. However, the chip count is substantial and there is PLD as well.

 I think it is reasonable to go for minimum chip count and I do like the SPI and USB options suggested above. But, I can't find these chips in a DIL package ... soldering SMDs is a real pain unless you prepared to try gas mark 4 for 10 mins!!!!

 A 74HC165 is exactly what I had in mind for the switches. It think this would be OK for quadrature rotary encoders as well without too much software overhead (people do tend to spin these very fast!).

[Lars Brinkhoff]

Angelo wrote:

I also want a proper TX-0 panel

This is a proper TX-0 panel, right?

[Malcolm Ray]

Great news on the PiDP-1! I've been toying with making a PDP-7 panel, and it looks like the indicator lenses and conical toggle switches were similar on the two machines.

So, if the PiDP-1 is a reasonably faithful replica, some parts could be reused for a PiDP-7!

[a...@papnet.eu]

That is indeed a problem. We want the PiDP-1 to be affordable so we're using stock parts only, and the panel will just be a nice pretty PCB (Oscar is currently working on that). Flat top LEDs for the lamps and regular switches for the toggle and key switches. Maybe a future version could be more accurate and prettier.

[Malcolm Ray]

Oh well, that's still desirable!

--
You received this message because you are subscribed to the Google Groups "PiDP-10" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/55281e15-47e7-4b98-af78-9a936e68cf92n%40googlegroups.com.

[Drew Rogge]

Now we just need to find a supply of 10 inch vector scopes.

Many years ago I worked a machine shop with a number of NC machines that ran off paper tape. While we had a paper tape punch in the shop some times the network was down and we couldn't make the tapes locally. When this happened I had to go to the computer center, pick up a 9 track tape with the data on it. Then it was off to the basement of another building where there was a PDP-1. You would hang the tape and type MTP (Magtape to Papertape) on the Flex-O-Writer and the machine would spit out your tape.

However if instead you typed SPW the vectorscope next to machine would flash and you'd have a bright spot in the middle with two ships orbiting around it. Eight of the sense switcher were split into two sets of four, one set for each player. Once I found out about this I always made sure I brought someone with me when I needed to punch a tape.

On 8/9/23 12:33 PM, a...@papnet.eu wrote:
> That is indeed a problem. We want the PiDP-1 to be affordable so we're using stock parts only, and the panel will just be a nice pretty PCB (Oscar is currently working on that). Flat top LEDs for the lamps and regular switches for the toggle and key switches. Maybe a future version could be more accurate and prettier.
>
> On Wednesday, August 9, 2023 at 9:27:55 PM UTC+2 Sheepless wrote:
>
> Great news on the PiDP-1! I've been toying with making a PDP-7 panel, and it looks like the indicator lenses and conical toggle switches were similar on the two machines.
>
>

> On Mon, 2023-08-07 at 08:58 -0700, a...@papnet.eu wrote:
>> Then I have good news for you: Oscar and I have started working on a PiDP-1. For at least myself (but possibly more people) I also want a proper TX-0 panel, but first the PDP-1.
>>
>> On Monday, August 7, 2023 at 5:52:09 PM UTC+2 Andy wrote:
>>> Yeah of course I am somewhat constrained by my lack of knowledge :-)  but I am quite intrigued with the thought of getting a TX0 and a PDP1-style of panel up and running.  The more the merrier!
>>>
>>> On Monday, August 7, 2023 at 9:12:25 AM UTC-4 a...@papnet.eu wrote:
>>>> I thought about this too, I'm just not sure if it's really worth it to spend time on polishing this thing, which was sort of meant as a prototype and a learning experience for me. Having a dedicated panel for the machine you *actually* want to simulate (rather than a compromise) might be a better thing.
>>>> So let me ask you this: which machines do you plan to use the panel with?
>>>>
>>>> On Sunday, August 6, 2023 at 6:10:20 PM UTC+2 Andy wrote:
>>>>> soooo thinking out loud here... while I wait for my components and the board to arrive, I was thinking it would be nice to make a cover for the panel. SOmething simple, like holes in plastic for the LEDs and switches, AltairDuino-style.  Now I was pretty awesome at uploading the Kicad files to JLCPCB.. .any thoughts on the level of effort to take that kicad and make a simple cover?  Presumably in some sort of plastic or aluminium using the same component positions from Angelo's board.  Having the switch names and LED numbers printed as well would be a bonus.
>>>>>
>>>>> thoughts appreciated!
>>>>> Andy
>>>>>
>>>>> On Friday, July 21, 2023 at 9:10:14 AM UTC-4 pbi...@gmail.com wrote:
>>>>>>
>>>>>> Just zip the gerber directory and submit the zip file to JLCPCB.  5 boards for $15.30, plus shipping to your location.  In my case that adds $14.92 for “Global Standard Direct Line”.
>>>>>>

>>>>>> *From:* pid...@googlegroups.com <pid...@googlegroups.com> *On Behalf Of *Andy
>>>>>> *Sent:* Friday, July 21, 2023 8:28 AM
>>>>>> *To:* PiDP-10 <pid...@googlegroups.com>
>>>>>> *Subject:* [pidp-10] OT - Angelo's Blincolnlights 18!
>>>>>>
>>>>>> Looks like a lot of fun! Blincolnlights 18 | Hackaday.io <https://hackaday.io/project/191985-blincolnlights-18>

>>>>>>
>>>>>> Once I figure out how to get a board made from the kicad I am going to give it a shot :-)
>>>>>>
>>>>>> Andy
>>>>>>
>

> --
> You received this message because you are subscribed to the Google Groups "PiDP-10" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to pidp-10+u...@googlegroups.com.

> To view this discussion on the web visit https://groups.google.com/d/msgid/pidp-10/55281e15-47e7-4b98-af78-9a936e68cf92n%40googlegroups.com <https://groups.google.com/d/msgid/pidp-10/55281e15-47e7-4b98-af78-9a936e68cf92n%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
Drew Rogge
dr...@dasrogges.com

Phone: 8934OOO629OO4829631OOOOOOO

[Andy]

Hey Angelo - I finally got my board delivered and assembled... had a few dumb questions following the tutorial.  I cannot seem to get past the mkptyfl /tmp/fl step....

Anyway, any tips appreciated.  Below is what I did so far

cheers!

Andy

sudo apt-get install verilator

wget https://github.com/joan2937/pigpio/archive/master.zip

unzip master.zip

cd pigpio-master

sudo make install

          I downloaded blincolnlight-master.zip from github

          Unzipped it

          Went into tx0 folder

 make

cd ..

cd tools

make mkptyfl

make mkptyfio

         I can see mlptyfl file in the tools folder with x attributes

         But if I try to mkptyfl /tmp/fl  I get

Bash: mkptyfl: command not found

        I tried running the mkptyfl /tmp/fl from different folders but no luck.

On Friday, July 21, 2023 at 9:14:58 AM UTC-4 a...@papnet.eu wrote:

[a...@papnet.eu]

Oh just run it from the local directory as './mkptyfl blabla'

Also, this mkpty thing is a bit awkward. Originally I had my emulators open an (ascii) tty file so i could also attach a serial terminal easily but now that i included flexowriter and FIO-DEC conversion in these programs that's become a bit silly. Would probably be easier to just accept connections via telnet. Oh well, this will all be fleshed out for the PiDP-1, the whole thing is not particularly flexible and also incomplete right now.

But feel free to make suggestions.

I'm currently working on a whirlwind simulator for the panel.

Wonder if this is becoming a bit too off-topic for the pidp-10 group. but not sure where else to go.

[Andy]

Hey thanks Angelo!  Yep I am now doing  better than before - I now have a nice blank terminal screen as shown below.

But I may have other problems.  Hard to tell from the photo but most of my LEDs are very faintly powered up. And the power button doesnt do too much from what I can tell. I assume the ribbon cable is oriented correctly - I tried the other way round but the little TBD62783APG chip got very toasty, so I powered off.

And finally yeah I agree this is definitely a bit off topic. Apologies to all 😁  All I need is one "you can now order!" post from Oscar and I will cease and desist 🤣

[a...@papnet.eu]

Sorry for late reply. I can't tell what exactly you're connecting from that photo, but i'm not convinced it's right.

The 1-pin on the panel goes to the 1-pin on the pi.

I added a photo to the hackaday page showing how i imagined it:

https://hackaday.io/project/191985/gallery#9b3db6dc762bf76f264975b194282c57

Cheers

[Lars Brinkhoff]

Andy wrote:

But I may have other problems.  Hard to tell from the photo but most of my LEDs are very faintly powered up.

That doesn't sound right.  You may want to check everything is wired up correctly, diodes placed with the right polarity, etc.

 

And the power button doesnt do too much from what I can tell.

The power switch in the upper right is entirely software controlled.

 

the little TBD62783APG chip got very toasty, so I powered off.

"Very toasty" sounds like a cause for concern.

[Andy]

Yep I am aligned same as yourself with the pins.  I installed pigpio and enabled the service to run at all times.  I also installed verilator, but I was unsure about configuration.

Is there some step I am missing to integrate pigpio, verlator and the mkptyfl files?

cheers! Andy

[Andy]

double checked my diodes and soldering.  My soldering will probably not win any awards, but it is good enough I think.

I am running noobs on my Pi. Correctly aligned ribbon cable and the chip does not run hot, but diodes are faint.  Pressing buttons and switches does make some of the LEDs brighten or dim - albeit somewhat arbitrarily....

Clues appreciated!  Andy

[Andy]

Hey Angelo/everyone! 

While we wait in earnest for the PiDP-10... I made a quick video showing my exploits with the Blincolnlights 18.  As you can see - *something* is going on but I do not know where to go from here.  I followed instructions on Hackaday and on https://github.com/aap/blincolnlights.

(apologies for the darkness of the video - my LEDs are a bit faint)

Any suggestions are appreciated!

tks

Andy

  https://www.youtube.com/watch?v=K5RgxHW4-24

[Andy]

As part of my continuing quest for all things blinky - I finally got three of my PCBs mounted and put in my rack.  Feels like Christmas!

[Heinz-Bernd Eggenstein]

Amazing! So, what's "behind" the panel, what is driving the lights?

Cheers

HB

[Andy]

Well it's all a bit of an inglorious hack- with a $4 ebay meanwell for power and three Pi zeroes running Angelo's Whirlwind/TX0/PDP1 software (https://github.com/aap/blincolnlights.)

I added an aliexpress hour meter, powerswitch a voltmeter and a powered USB outlet for good measure.

But after a good while of heavy duty blinking, the Pis become a bit unresponsive over wifi/ssh so I have them periodically restart via crontab.

cheers!

Andy

[terry-...@glaver.org]

The Raspberry Pi always seems to be "4 steps forward, 3 steps back" in both hardware (the early Pi 4's "I'm a pair of headphones, I don't need USB power") and software.

I updated a Pi 4 that had been running a few releases back to Bookworm, and WiFi consistently stopped working a few hours after booting, where this never happened in the older OS. I had to add "/sbin/iwconfig wlan0 power off" (where "power off" confusingly doesn't mean "turn the power off", but instead "don't try to do power management") to /etc/rc.local. Maybe something similar is causing the problem with your Pi's?