Close
0%
0%

Numitron Hexadecimal display module

Another hexadecimal display, this one is sans transistor for the extra vintage touch, and uses only Soviet-era parts

Similar projects worth following
This is another spin-off from the relay-based YGREC project that needs a hexadecimal display. I first thought about using the pretty cool DYPLED module but the sophisticated and cheap module would be out of place. I've chosen Numitron tubes for their incredible vintage look and ease of driving. The hard part is to make a minimal module for decoding a 4-bits bus into the 7-segments display everybody knows and loves...

Specs:
* Supply : 0V, 3.5V, 7V (up to 0.45A, not symmetrical)
* 10 РЭС15 minirelays, one IV9 display tube and 59 D9K (germanium diodes)
* Decodes 4 bits, from 0 to F
* Input impedance : one relay coil per bit

All files are under https://creativecommons.org/licenses/by-sa/4.0/

Digital data display is a recurring and nagging problem when designing DIY computers !

Look at @Artem Kashkanov 's #BrainfuckPC Relay Computer  : it uses VFD tubes but they are driven by semiconductors.

In the case of #YGREC-РЭС15-bis the constraint is to avoid using transistors. This all-Russian neovintage module uses

  • РЭС15 minirelays
  • IV-9 Numitron tubes
  • D9K Soviet germanium diodes

This project is also a continuation of the log 20. Display technology again (yes, another fork !) and I'll move the relevant discussions here.

The logs explain the 3 methods I've invented to keep the parts count low. Some optimisations are pretty neat !

Ideally, like the #DYPLED, I should be able to design and build individual modules for my own use and for eventual #Hackaday TTLers fellows  :-) This would be a step even further than the NUMI-311 (TIL311 module replacement) because of the increased power draw and the increased room. However it will be warmer (both sentimentally and thermally) and you could show a kid the inside "for educational purposes". Who would not want to play with the buttons ?

One design constraint is to keep the input impedance identical and low for all the data bits, to ease integration with other YGREC devices.

Power supplies are derived from experience with the #YGREC-РЭС15-bis so we're using  -3.3V, 0V and 3.3V power rails. Using a single power supply would be too much of a hassle and increase the complexity and the Numitron draws little compared to the relays...

Logs:
1. Prior art
2. References
3. Physical configuration
4. Circuit and schematics
5. First prototype and first setback
6. More diodes !
7. Diode optimisation
8. Going further
9. A less preliminary layout
10. A little idea to revive...
11. Another mistake
12. Second valid prototype
13. A new ROM layout generation script
14. Beyond the prototypes
15. PCB
16. The demonstrator goes portable !
17. Enhancements
18. New layout
19. PCB prototype
20. Cascading the LZB
21. PCB delivered !
22. What would the Russians do ?
23. Patched
24. And it works !
25. More boards
26. 5 modules
.
.
.

RES15_doc000248413.pdf

Russian datasheet of the RES15 relay family. See https://hackaday.io/project/10889/log/37255

Adobe Portable Document Format - 58.02 kB - 04/16/2018 at 03:18

Preview
Download

convo7seg_09.html

Updated script that generates the final ROM configuration (swapping 5 and 6)

HyperText Markup Language (HTML) - 2.74 kB - 10/30/2017 at 17:20

Download

convo7seg_05.html

A script that searches which input permutation saves the most diodes

HyperText Markup Language (HTML) - 3.09 kB - 09/24/2017 at 01:35

Download

  • New layout

    Yann Guidon / YGDES05/12/2018 at 17:07 0 comments

    Many changes and updates, following the first round of prototypes.

    I changed the connector, a 2×8 0.1" pitch, including a key. It provides data, power and the Z output.

    The Numitron pinout is corrected, at the cost of a couple of vias.

    The board is slightly smaller, 72×85mm now.

    Many pads and drills are adjusted for better soldering.

    I must finish the first tests on the prototypes before I send the new boards to fab.

  • 5 modules

    Yann Guidon / YGDES04/27/2018 at 08:05 0 comments

    Soldering was pretty smooth and the 4 other modules are rather easily assembled. They all worked on first power-up !

    So I played with them a bit, to see if the dimensions are good.

    I finally decided to solder tha relays inside, as previously planned, to keep the overall volume low. Reducing the thickness would have increased the total space used by the display...

    The result is not as small as I expected but it's still pretty nice !

    Now I have to implement cascading...

  • More boards

    Yann Guidon / YGDES04/23/2018 at 23:41 0 comments

    I prepare the 4 other boards to make a 16-bits display. The diodes are looking good so far.

    This time I'll solder the relays outwards to save a couple of millimeters in thickness.

  • And it works !

    Yann Guidon / YGDES04/22/2018 at 05:49 2 comments

    Here is the module, much closer to the intended format.

    The module is quite larger than initially desired but the thickness is ok, so I can stack several modules and keep the numbers easy to read (with a decent gap between digits).

    After a while, the relays can get warm...

    I had to increase the voltage because some relays wouldn't switch. Strings can get unbalanced so I will have to "bin" the relays ( #ReTest-RPi  is meant to address this). So the module gets up to 3.50V (×2) and F draws up to 0.45A (the power consumption is around 2W). Digit 0 draws quite little.

    The power connector must definitely be redesigned...


    Now, I think that surface in itself is not SUCH an issue and I can make the module thinner by bending the relays outwards/outside. That's the program for the next 4 modules :-)

  • Patched

    Yann Guidon / YGDES04/22/2018 at 02:12 0 comments

    I patched the traces to let the Numitron sit at a suitable angle, close to what I envisioned in the first logs.

    I'll have to modify the schematic and PCB layout but fortunately it's not a significant change.

    The testing circuit must still be finished, I'm annoyed by the power supply connector. I should use a HE10 header for more than the data.....

  • What would the Russians do ?

    Yann Guidon / YGDES04/14/2018 at 03:13 3 comments

    At two points, a string of 3 coils in series is switched (by another relay) in the middle (that is not high-side or low-side). I expect a significant spike when the contact is released...

    I don't want to waste 2×2 more diodes for the freewheel. They are not free, they use space and they might not stand the repetitive abuse. I prefer a passive approach, using dumb off-the-shelf parts that a Russian engineer might find, decades ago.

    There is not much choice : a capacitor or a resistor...

    A capacitor would be quite expensive, but a 100nF ceramic across the relay's contacts could reduce the dirty HF parasites. However I have no idea how high the voltages could go... And even though the Russians are known for their high-voltage capacitors, it might not be the cheapest approach, particularly if no electrical data is available. And increasing the capacitance will also increase the current spike when the relay closes again...

    So I'm left with resistors. The prototype uses two vintage 1K 1W resistors from Russia. In theory, it would work, even though it adds a bit of quiescent current (6mA×2 ? that's as much as one incandescent segment). The bet is that as the voltage difference increases, the resistor will let more current flow. I should probably measure this on the scope....

    Any advice ? @SHAOS ?

  • PCB delivered !

    Yann Guidon / YGDES04/12/2018 at 12:34 1 comment

    I just received the boards and did a thorough quality check.

    Thank you to the sponsor @Fredrik Högberg !

    It seems that it could be made on bakelite :-)

    Notes for later :

    • PCB trace widths : OK
    • Clearances for mounting holes : OK
    • Clearances for tracks and pads : OK
    • However the pads for the diodes are still too small. I can't even distinguish the round ones from the orthogonal ones. Fortunately there is a silkscreen that shows the diodes' polarity. But for a bakelite/FR2 run, this will be a problem. The 2×5 connector could also be modified, or even changed for a larger pitch.

    Now let's solder one :-)


    More notes :

    • The resistors' pads are too small
    • The diodes spacing is OK and could be tighter
    • I totally messed with the order of the pins of the IV9 when I tried to reverse the order of the segments on the schematic.

    I think I tried to put the Numitron on the other side of the PCB but it didn't work as well as I planned :-(

    After some pin bending, I was able to display 0, 1, 2, 4, 8 but I have to create a PCB to display the rest.

    It's enough to validate the basic system but I must redesign the PCB :-(

  • Cascading the LZB

    Yann Guidon / YGDES04/06/2018 at 07:58 0 comments

    LZB is "Leading Zero Blank" and it's a tiny visual feature that eases reading. When the most significant bits are cleared, the corresponding 7-segments digits are not turned on. I introduced this option in #DYPLED  but I guess it existed long before that.

    The first project that uses these display modules is #YGREC8  which is 8-bits wide, or only 2 digits. In this case, there is no real cascading. The circuit is very simple : the Zero output is tied to +3V when the input code is 0000, so this powers a relay coil (though a diode). The relay is "normally closed" and connects the 0V rail to the module. So when the input code is 0000, the relay opens and disables the display of the digit "0".

    At first I wanted to mount that relay on the module. Then I decided that the circuit should be on the backplane because each module would need a special assembly, depending on their position.

    The good news here is that the overall current is reasonable : the 6 segments draw about 13mA×6=80mA, OR max. 60mA for the coil (which cuts the 80mA when activated). Using a prebias resistor, this can be reduced a bit (at the price of more static consumption). So a spike would draw maybe 60+80=140mA, or 200mA worst case, and the two other relays in the decoding tree will not suffer too much.

    (If you missed one episode, the RES15's contacts are rated at about 100mA)

    This was the description of the wiring for the Most Significant Digit. The Least Significant Digit doesn't use a relay because its digit is always on, and its wiring is more straight-forward:

    So a 8-bit bus requires 2 modules, as well as 2 diodes (one for freewheeling to the 3V rail) and one relay on the backplane.


    For more bits, it gets a bit more complicated. In the case of 16 bits, 4 digits are required. The MSB looks like the first diagram and the LSB is like the 2nd diagram. The two other digits need more sophistication...

  • PCB prototype

    Fredrik Högberg04/05/2018 at 16:37 1 comment

    Just a quick update.

    The PCB prototypes was just finished, supported by # Lets make OSHW more successful! .

    The boards has been sent out to Yann Guidon / YGDES who insisted to do the assembly. Here is your sneak peek at what the board looks like before assembly:

  • New layout

    Yann Guidon / YGDES03/28/2018 at 04:38 0 comments

    I've spent the whole night polishing the new version (Rev.1) of the PCB.

    I iterated a lot and finally, I only output one more signal that is tied to +3.3V when the input code is 0000. Cascading is easier and totally optional, the backplane must do its own circuit depending on the desired behaviour and cascade length.

    The layout finally contains three mounting holes (3mm diameter) and several things have been moved around. You'll notice the crossed transistor symbol ;-)

    Dimensions are 90×72mm. I have removed one via but 2 new ones have been required in the end.

    The layout is VERY conservative. No silkscreen is needed, all the necessary information is available on the copper. For example the pad shape of the diodes are octogonal at the cathode end. The idea is to create the boards with old technology (even on bakelite :-D) so even the lousiest fab should work.

    I'll have to figure out the orientation of the IV9 tube... There was a little issue with the footprint but nothing critical.

    Anyway, the board is looking good®, it allows blanking-cascading and (I just realised) the single-relay input allows these relays to be used as hysteretic data storage...

View all 27 project logs

Enjoy this project?

Share

Discussions

Yann Guidon / YGDES wrote 04/25/2018 at 21:40 point

Brian does seem to do a lot of promotion for this side-project, which is pretty unexpected but always welcome :-)
http://hackaday.com/2018/04/25/video-quick-bit-numitrons-and-infinite-build-volumes/
Time to finish the 4 other boards !

  Are you sure? yes | no

Dr. Cockroach wrote 04/25/2018 at 23:02 point

Oh, cool :-D

  Are you sure? yes | no

Yann Guidon / YGDES wrote 04/25/2018 at 21:35 point

And Benchoffed again :-D

  Are you sure? yes | no

Nixie wrote 04/23/2018 at 01:29 point

You got Hackaday'ed ^^

  Are you sure? yes | no

Artem Kashkanov wrote 04/22/2018 at 21:12 point

Hey! I remembered that I have yet another beautiful indicators - Flip-Flop!

And I have 5 of them https://hackaday.io/project/153330-flip-flop-segment-display

  Are you sure? yes | no

Yann Guidon / YGDES wrote 04/22/2018 at 21:17 point

Now you have to make a decoder with diodes and relays.
But unlike incandescent lamps, you'll have a harder time because you need to turn on AND turn off, so the pulses must be in 2 directions...
More diodes, fewer minimisations :-/

  Are you sure? yes | no

Artem Kashkanov wrote 04/22/2018 at 21:19 point

Yep, I found that I already have Diode-relay schematic  for this indicators :) 

But I have no idea when I'll start this project;)

  Are you sure? yes | no

Artem Kashkanov wrote 01/22/2018 at 07:48 point

Wow! So cool indicator, but I couldn't find enough number of them with acceptable price :(  So have to use VF ones.

Interesting schema of BitToSeg encoder - you used two-polar power supply to reduce number of parts?

  Are you sure? yes | no

Yann Guidon / YGDES wrote 01/22/2018 at 07:58 point

Hi ! eBay prices for Numitrons range in the 3-4 USD per piece. Yes it's a bit expensive, more than VFD, but VFD is more complicated due to high voltages.

I have used a symmetrical power supply (2×3.3V) because it's already available in the system I build. I have also made one version with +6V single supply and that uses one more relay to reverse the polarity.

What would you like to do with them ?

  Are you sure? yes | no

Artem Kashkanov wrote 01/22/2018 at 08:08 point

Same as with IV6 - use them for showing current data/address bus states, instruction retired count, etc.

Yeah, +48V requires some additional adjustment logic for controlling them. And you also need 1.2V for heating...

As for the price - I got dekatrons 5USD each... and IV6 for free, so I have no choice but to use VFD ;)

  Are you sure? yes | no

Just4Fun wrote 10/30/2017 at 17:50 point

Cool!!

  Are you sure? yes | no

Yann Guidon / YGDES wrote 11/01/2017 at 21:44 point

And it's only the beginning :-)

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates