The Magic Of Muxing!
HTML-код
- Опубликовано: 12 июн 2024
- Back to the Packard Seven Segment Numitron Driver Board, this time to get it to do what it is designed to do - Multiplexing! Enjoy...
Part One: Resurrecting A Vintage Numitron Board - • Resurrecting A Vintage...
Join Team FranLab!!!! Become a patron and help support my RUclips Channel on Patreon: / frantone
#franlab #multiplex #nixie
- Music by Fran Blanche -
Fran's Science Blog - www.frantone.com/designwriting...
FranArt Website - www.contourcorsets.com 
Наука
I loaded PC boards as summer jobs during college. Paid twice minimum wage! A really good high school electronics course meant I could easily read resistor codes, and a Manpower rep at a job fair said, “Want to be employed tomorrow?” 😄 Got me through college! And yes, I was usually the only male. 😂
Its hard these days to simply get young folks to come to work 5 days in a row.
@@resipsaloquitur13 it wouldn’t be if you paid twice minimum wage.
Same here but wire-wrapping! The cables were colour coded for length. Sounds tedious but I loved it.
A great candidate to have a watchdog set up, as a crashed sketch could end up dumping the filament "overvoltage" without the "PWM" style lower average voltage if any outputs get stuck on
At one point I had the idea of building a large matrix display with micro incandescent bulbs, 1.2V/15mA per bulb or so (still sitting in a drawer..) - I stopped when I realized that soldering 800 flimsy wires to my 2.5x2.5m framed wire grid and also programming the graphic functions was a bit tedious. But what helped to increase the response time of the bulbs was to preheat them with a current that didn't make them glow, or to scan them fast with a pulse width that also didn't make them glow just yet. As far as I remember this way it was not necessary to bump up the voltage this much. (So in case of an error they wouldn't break because of the overvoltage.)
I have the identical idea. The “always on” current, will be high, and effectively creat a I.R. source (I would fine-tune that voltage/current in a darkened area, for a barley perceivable glow), effectively, a radiant heater.
This technique will increase the glow-time, after the ON signal is replaced with the Keep-Warm, which is a small inconvenience compared to the projected extended life of the lamps in your display.
I’m looking forward to seeing your build!
Somewhere I saw a spec sheet / data sheet about using bleeder resistors for a trickle-current thru the filaments.
It wastes energy but reduces the thermal shock so it probably extends the lifetime of the filament since it's already at high resistance.
@@SawdustSoftwareSiliconChippy :) I've did this in 2006 or so - unfortunately I'm struggling with my health nowadays and due to the tremor of my left hand (and a huge backlog of ideas/projects) it's a bit unlikely that I'll revive the project. Anyway, I was fascinated to have an almost point-like, tiny light source that shines 360° in all directions. Today you can probably find LEDs for this.
Props to the ladies including you Fran.
It would be very nice for a nightstand clock. Analog voltage dimming for night time, tubes would last forever if not pushed that hard.
Yep. Made one. Pi powered clock radio.
@@liquidsonly you posted it somewhere for us to see? And why a Pi? Itsy quite power hungry.
@@zaprodk My posts to the link get deleted each time.
@@liquidsonly Only solution seems to be to setup your own RUclips channel, put your stuff in there then post that RUclips link here - RUclips does not mind links to itself
On your own RUclips site, as your now a creator, you can put in as many links as you like, just like Fran has on here.
@@paulstubbs7678 Oh! Interesting. I already have a channel. I'll look into that.
What i am missing (or did not see?) is something for emergency shutdown of the filament power when something with the mux goes wrong. Just a random software crash may kill the expensive display quicker than the watchdog timer would kick in....
This demo doesn't seem to have such a feature and that's probably why Fran is turning on the filament supply after the logic supply. But it should be noted, light bulbs can often last for a few minutes at 3x rated voltage before burning out. Knowing this, a few seconds of 7v on a 5v filament isn't going to cause any harm. If Fran was running higher voltages additional protection would be much more important.
The nerdy parts are the best! Always loving you taking deep dives WaaaaaaaaaaaaaayDownInsiiiiiiiiide :)
love these old displays, some of my older MPH police radar units have these displays, I have always liked the way they looked.
On my 100% DCI-P3 displays it looks very much warm like an incandescent and very orange at 5V!
I'm not sure if the 5 ms and 2 ms mentioned were per digit or per the entire display. I suspect "per digit." You have to get above the human eye "flicker fusion threshold." Long ago we might have called this "persistence", but that term also means how long an image may be perceived after it was gone, so the more specific term is now used. It is about 60 to 90 Hz for humans. If "per digit", 5 ms would be 35 ms to scan the entire display which is about 29 Hz and "flicker" would be expected. At 2 ms it is 14 ms for the entire display which is about 71 Hz and probably barely fast enough for at least many humans. But if certain other animals are watching this, such as an owl or a house fly, they will still see a flicker. You have to be a bit careful designing such things as not all humans are identical and if you use yourself as the test subject, somebody else might still see flicker.
I probably would have added a few lines to the arduino to change the timing from a button input or similar.
Fran has a real soft spot for display systems
I could be wrong, but i dont believe solder flows over the component side during wave soldering.
Seems pretty risky, "feels" like high chance of blobs n bridges gettin' stuck under the chunky components. I have no direct experience of PCB manufacturing in the '70s though, I started dabbling with it during mid/late 90's 🙂
Agree. The solder is ONLY on the bottom of the board. If it were to go over the board it would immediately be scrap. The thick solder traces are covered in solder before the components are loaded and it's called HASL.
Yes. They tinned the traces before putting the components. Once saw a board where the wave accidentally went over the board. What a mess.
Agreed. The one time I got a tour of the PC board manufacturing facility, I saw how the solder "kissed" the bottom of the board and wicked up.
The board is suspended trace side down over a tank of liquid solder. A linear actuator at one end of the tank "slapped" the surface of the solder, sending a single wave along the tank and washing the underside of the board.
When you stuffed the board the tails had to be folded over because the solder would push the components out.
The women stuffed the boards but men flowed them, at least where I worked.
How much effort would it be to drive them all individually? Might be a fun challenge
Most camera's have an IR filter. Would a camera without one pick up the true color of the filaments at lower voltages?
Shrunkified 😂 👏 👍
Edit: I’ve never heard a better one word description. Tnxs!
Hi Fran! I want to get into electronics doing hands on projects for beginners. I'm looking for a good starter kit. The AM Tech kit looks good, but the founder has been idle for about a year. I'm not sure if he is in business anymore and he had a successful launch. Can you recommend a good kit with components, circuit projects, data sheets, etc.?
I love these "weird, crazy videos"! Love you too! Thanks for sharing!
NEVER multiplex vintage displays if you have the option to do direct-drive. As a longtime creator of clocks using nixie, numitron, NIMO, etc devices, the last thing you want to do is reduce the operating life of a device that is no longer manufactured.
In the old days, multiplexing was done because the drive electronics was expensive. Today, the electronics portion is dirt-cheap compared to the cost of the tubes, so saving a few dollars on the IC's doesn't make much sense if your tubes cost hundreds of dollars apiece. For example, the gigantic, though butt-ugly 14-segment 7971 nixie tubes are going for 200USD apiece now, and the few NIMO tubes I've seen were over $300, and those were untested. A six-tube clock is a hefty investment, so treat your tubes as best you possibly can.
Why is multiplexing bad ? Simple. When you multiplex, you can see the brightness goes down. To offset this, the current must be increased in order to increase the brightness. With gas-filled tubes, the wearout mechanism (generally sputtering) is exponentially proportional to the current. Incandescent tubes undergo thermal shock, from cold turn-on as well as thermal cycling. Similar wearout problems can be expected from multiplexing fluorescent and electroluminescent displays.
Some displays, notably gas-discharge panels and multicharacter displays, require multiplexing because the number of pins disallows direct-drive, so in those cases you have no choice.
I add circuity to limit display currents to safe values, such as using current regulators on my nixie clocks, and filament preheat/cooldown on the incandescent/NIMO ones. Knock-on-wood, nothing has died yet.
Thank you so very much Fran for sharing this, very relaxing fran cad..i enjoyed your content. . Looking forward to seeing your next video 😊
Would the cost of one diode block be 40x more expensive? It might be possible that producing this block would be only 3x more expensive than same amount of discrete diodes. ICs weren´t that much integrated at that time. Encapsulating a single diode vs. encapsulating a dozen diodes should not be a factor of 40! Still a valid point of cost saving by using ICs.
If the blocks were custom-made it would be more than 3x the difference. If it was off the shelf then I could believe 3x.
I've always had a bit of a facination with electronics. However it is comparable to brain surgery...lol
I am incredibly jealous of your B-7971 display, but I'm glad you're continuing your Numitron video!!
Awesome as always Fran! Another scheme for multiplexing I am using 74hc244 tri-state buffers with the outputs tied together. BCD inputs would be controlled by a CD4017 which is pulsed by a 555 timer at the frequency of your choice. The 4017 fires the OE pins of the tri-state buffers like a distributor cap on a car. A "good" way to run those pesky VFD tubes with all of the digits inside one tube like an old microwave oven. Arduino would be easier of course, but this is another "easy" option. I'm kind of new to this, but I believe that most multiplexing is an integral part of the IC's or micro- controller(s) being used. A 74hc244 has 8 channels with two OE (low output enable) pins on each so that would allow you to only use seven of them to control the segments and possibly a decimal point. If not, be sure to tie one of the inputs low or high. Hope this helps someone out there. Comments welcome.
Very cool!
Wow, thank you!
The flicker was lovely, Happy days.
I have a few hours free if you like me to debug FranCad. Funny that the "boring" part of the video was what I wanted to learn about most. I have some ICs that I wanted to use for multiplexing, shift registers actually IIRC, but I have never gotten them to work. Been a while since I tried, maybe I wasn't using diodes. Funny that that I do digital programming and analog electronics, but I know nothing about analog programming and digital electronics.
If you're multiplexing with overdrive and the processor locks up, I guess you run the risk of burning out a display?
Yes. Also imagine just how many TVs got a horizontal phosphor burn from the vertical deflection malfunctioning. With the specific display Fran has the original host device most likely used hardware multiplexing, minimizing the risk.
1800k
Neat explaining how it works, only need one BCD to 7 segment decoder IC.
@Franlab. Is that a Yamaha drum machine in your theme music?
When I grow up, I’ll own a nixie clock 😎