We had a full reactor core display that used miniature-387 bulbs. But it was 2908 bulbs total!! The only time they were all lit at the same time was lamp test, but that was a whopping 3.2 kWatts!! We rigged up a timer so you couldn't leave it in lamp test for more than 10 minutes. Several fans behind the panel for cooling them. And for our drivers, we had a resistor tied in parallel with the drivers to pass a 'keep warm' current through the bulbs. Selecting a resistor such that the bulb had about 1/3 voltage on it when 'off', reduced the in-rush current when turning on, and bulbs actually lasted a lot longer. (when we upgraded to LED equivalents, we had to remove the resistors since LEDs they passed enough to turn on the LEDs lol )
It turns out if you're willing to pay MILSPEC/Aerospace prices you can get some pretty nicely engineered things. Makes you wonder what you could get now.
Not with those sub miniature D connectors, those are basically a single use non repairable item. The connectors also come with all wires the same length, and you cannot replace a pin or socket in them, as they are potted during manufacture. I have had to replace quite a few, unlacing the old one from a wiring harness, soldering in the new one then sewing up the wiring harness again with lacing twine, then applying the conformal coating again to the PCB and the twine knots so they can be protected.
If you want to use the IR camera on pure metal it is recommended to place a peace of black pvc tape on it. Otherwise you will measure the wrong temperature. You could change the emmisivity but then you have to change it back when measure something else.
Yes, a really neat display and your solution is excellent. The issue with incandescent displays and semiconductor drive is the near zero cold filament startup resistance as you indicated. This drove designers nuts and there are many many discrete transistor driver boards running around in the 70's for this and similar reasons. Once VFD and LEDs arrived, driver design because much simpler, could be run faster, and multiplexing displays became feasible.
@@repeat_defender Cool! I love kit building of all types. Electronic, plastic or wooden models, even K'nex lol. I hope you enjoy building it and using it!
A heatsink for just 1.36W total of tungsten lamps? They don't mind running at a high temperature. And why no fins? Maybe the metal block is meant to absorb vibrations to protect the filaments. I'd guess it was mounted on springs or in rubber.
@@flyguille Idk if CNC was available in 1970. They would have flycut a 40 or so inch bar and cut into short pieces and mill several at a time clamped together. Adding fins would jack up the price, and it's not necessary for glass.
Despite being American, I gawked at how crazy it getting to 90° seemed for a good few minutes, until I realized it was Fahrenheit rather than Celcius. Weird how my brain automagically split thinking in °F for weather and cooking and °C for science and EE! Or maybe I was just too distracted by such a -cool- hot display to think about it much, haha. Thanks for sharing yet another wonderful video, Fran!
Heathens and their °F ;) Still, I think Fran should have hooked up a thermometer to the display driver so it would display its own ever rising temperature. That would have been some kind of... electrical waterboarding.
A piece of equipment I used in my military days used incandescent 7 segment displays. The difference with this one is the front of the display was removable. You then were able to reach in with a pair of needle nose pliers to replace the grain of wheat lamps. This way you didn't have take the equipment out of service to change the lamps.
Interesting old calculator! I recall back in the min 1970's I was serving with the 2/3 Armored Cav down at Fort Bliss Texas. We had a communications building in the Motor Pool area, just a sort of small block house with enough room for our supply of commo gear, and a couple of CONEX containers for the bit stuff, like miles of wire and such. Our section was commanded by a Captain and a Second Lieutenant. These officers, of course had to have some college, unlike we in the NCO corps and enlisted soldiers who did the dirty work. Our Captain had his bachelor's degree in some mechanical engineering with a minor in electrical something or another. During his college career he needed one of those fancy calculators. Of course back then a calculator that could do square root was fancy and his, why his could do trig and crap like that. We were impressed, he would let we peons hold it and press a few buttons so we could OOOH and AAAH a bit, then it was back in his leather pouch and locked in his desk that sat in the southwest corner of our shop. Cool to see one of those back doing it's duty once more!
These old displays are super cool, but it's also remarkable how dominant LED technology has become. If something emits light it's almost certainly some kind of LED, or the sun.
Very nice looking display. When building a high-end piece of hardware required real engineering, both electrical and mechanical. Considering the dozens of steps required to make this, I'm guessing they were never produced in large quantities, and they must've been pretty expensive! Even machining the block of aluminium to the required precision would've been challenging, since they probably didn't have the kind of CNC machining systems used for this sort of thing today. Each step would require a different machine with a different jig, and an operator doing a single task, e.g. drilling the mounting holes, or tapping the threads.
That is just cool... definitely military or aerospace. On incandescents lasting a long time...I have a clock radio I got at a yard sale in 1976 with a incandescent dial lamp. Still works. Other than moves and power failures, it's been on since then.
maybe you could run a constant current source through one filiment to get an idea of how much the cold filament current changes to hot filament current. start with 1/10th ma and measure Vdrop then working current. the circuit uses a linear 3 term reg and a resistor.
The BCD decoder is designed with the CMOS gates sized to limit the base current in the NPN transistor. The 20mA is only available when you have a red LED with a 1V2 forward voltage drop when lit, as this them will get 20 mA drive, reasonably consistently across each particular die, and reasonably close per wafer lot. The most common way to drive those displays, along with the Numitrons, was to use CD4049/50 COS to TTL buffer chips, as they are rated to deliver 50mA minimum source or sink current, into a TTL logic 1 or logic 0, nominal voltage of 3.4V or 0.8V respectively, with a 5V supply. Just needs to have a single extra gate per digit, or two if you were also driving the decimal points actively as well.
They must've cost a fortune to make! Would've been so awesome as a frequency display on a high-class early 70's hi-fi reciever. Re-engineering your driver circuit on the fly to include a buffer amp is frankly above and beyond the call of duty! You exhibit a degree of determination and patience I can only dream of!
It's gorgeous. I was going to complain, I can't find one. But seeing how much trouble you went through to get yours working, It's fine. I don't need one. LOL. (I appreciate the work you put into these demonstrations, I would'nt have the patience.) Great video as always. Thank you Fran.
FYI If you check Chicago Miniature Lamps CML-IT, a lots of 5V filament lamp have a 40,000 hours average lifetime. For example, a T-1 #680 is rated at 60,000 hours.
+4 volts was the operating voltage for RTL, Resistor Transistor Logic. A logic family made by Motorola in the form of standard DIP chips. A forerunner to TTL. Then there was also DTL, Diode Transistor Logic.
I think you forgot about the warm up current. Typical incandescent lamps have about 1/10th the resistance cold, as they do warm. There's a huge current draw when first powering up, which stabilizes after it is emitting light. It would be interesting to watch the current with a scope on a single digit to see how quickly these light (they seem quite "snappy"!) and what the "inrush" is really like.
@@jukingeo I watch so many of these videos and like to build some of the things that are presented. Many say you can replace the BC547 with the 3904 and I was doing that, but the pinout is different and I am a bit dyslectic. To solve the problem was quite easy Ebay has sellers that sell both so I now have boxes of 547's 2222's 2904's and all their p channel mates so I no longer have to do the conversion of pinouts and screw up my projects. The cost of these is really negligible, perhaps a few cents to a dime each so in today's money it's simply the price of a cup of coffee (which I LIVE ON!) for 100 of each.
I can hear Dave Jones comments about all those transistors flapping around in the breeze hahahaha Your collection of exotic displays is quite something, I had no idea there where so many weird implementations
Micro Ds are extremely rugged, but they are unforgiving if you misalign them when plugging them in. Within the company I work at, the instructions are only to mate them by engaging the jackscrews and turn 1/4 turn max at each end alternately.
Is this a similar display that was used in BttF for the MPH display in the Delorean and on Doc's remote control? Those looked like they were incandescent 7-segments to me.
It appears that the inability of the CD4511B to sink 20mA can be explained by Fig. 1 on page 2 of the TI datasheet. It shows the output characteristics of the FET used for sinking current. There are 3 curves on that plot. The only one that reaches 20mA is the top one which uses a 15V supply. Even 10V isn't enough to sink 20mA. With 5V, the most you can sink is 4mA, but at that point you have 5V of Vds so the external supply for the lights will need to be at least 9V for 4V lights. With a 15V supply you can turn on the FET hard enough to sink 20mA, but Vds will be about 6.5V. For 4V lights, that means you need at least 10.5V to run the lights, or a resistor (15-10.5)V/20ma = 225ohms between the 15V supply and the light to limit current / voltage to each segment. Looking at the static electrical characteristics, this part was really designed more for SOURCING current rather than SINKING current. At 5V, it can typically source 20mA with an output voltage of 3.75V, minimum 3.4V. Not quite enough for the 4V lights here. However, at higher supply voltages, you can get higher voltages out for the same 20mA. So it looks like a supply of about 6.5V for this chip would allow you to SOURCE 20mA well for 4V lights.
Right. I believe Fran was sourcing not sinking but I'm not sure. The On Semi data sheet shows the part will drop about 1.5V when sourcing so I am assuming she was using Vdd = 4V not 5. So yes 5.5-6V supply SHOULD work, but we don't know what specific device she has.
@@michael931 She is sourcing current, the decoder is direct driving, and does not have active low outputs. The sink transistor there is to prevent display ghosting when driving multiplexed displays, actively pulling leakage current, from either other segments not selected or from the chip itself, to ground so the digit drives are at logic 0 levels when not enabled for a digit. however the display type it is really designed for was the original Monsanto red LED arrays, where the forward voltage drop is 1V2, with the multiplexed common cathode arrays, using saturated NPN drivers, so the CE voltage drop was around 0.2V extra added to the display.
@@michael931 From her explanation at 8:00 to 8:40, it seems clear that she is trying to sink current through the pull-down FETs. The explanation about current *sensing* is incorrect. Current sensing can not be done by looking at the gate voltage, and the output pins aren't fed back in anywhere. The logic simply drives those gates high or low. However, it IS being current limited by the FETs, basically because their Rds_on is so high that it leads to a high Vds voltage when the FET is on, as shown by the curves in Figure 1 of the CD4511B datasheet (that part number is visible at 7:58).
@@SeanBZA Good point about the active-high outputs - I didn't check that before looking at the current limits. Sourcing current is easier than sinking with this part, but she would need to power it with more than 5V to get 4V & 20mA out of it.
@@michael931 You are right, she would have to be sourcing current to drive the lights - her focus on the pull-down FET made me think she was sinking current through them to turn the lights on. But this part is designed to drive the outputs high (and source current) to turn them on. I should have checked that before focusing on the FET limits!
Kind of a funny side story...when I worked at a mobile product company, one of .y colleagues related a story about looking at a product from a Swedish company, in which they accomplished LED dimming by actually "stealing" current from the LEDs, just 'burning it off', so we always joked about "Swedish dimming" .... imagine needing a heatsink for the dimming!
Hey Fran, a simpler interface is to string 3 X tpic6c595 chips (same as 74hc595 but has open drain outputs capable of 100mA pin sink) in series hanging off the spi port or can even be bit banged out. Much less soldering 😜
There are a LOT of videos available on how to build a small oscilloscopes on TH-cam. I have built a couple that work fairly well. There are also small kits for around 40 bucks that you can build, or if you lack the skill to build they can be had in prebuilt form where you simply plug the screen to the circuit board, and if you ordered it, build the little plastic case to protect it. I have built two of those, they both do the same job but I like the larger one simply because it has a bit larger screen. I did build on that plugs into the sound card on your computer, you can download freeware that does the display for you. I feared frying my sound card so I also purchased a small USB sound card to protect my computer but after a time I found it not necessary. The build has two 7805 voltage regulators to keep the output to your PC under 5 volts no matter what you plug into so it does a fairly good job of protecting things. I don't think I would ever plug the probes into a 110 or 240 outlet but for work on digital circuits it works just fine, you can also do a print screen on the output and actually print out a sine wave if you want to, comes in handy when trouble shooting and a saved screen shot can be sent to a more knowledgeable Elmer if you have one so he can look over your problems. KE0JBL
Indeed, the incandescent lamps have a strong increase of resistance at higher temperatures. Which is actually a saving grace for their longevity. I was for a long time using a rule of thumb that the resistance went up by a factor of ten from room temperature to the illumination temperatures.. The actual change is close to 90-fold. Think of this as self regulating feature against operating voltage variation. Not perfect, but helps the lamp operation. Compare that to LEDs that require constant current supply rather than constant (or loosely regulated) voltage.
Wow incandescent 7 segment, never heard of! Much look beautiful IRL. I'd have been so tempted to just solder another 4511 on top of the original one to bump up the current...
My goodness , a whole transistor per segment to supply power. That's insane compared to standard LED 7-segment displays that only sink 5-20mA per segment (standard outputs can handle that current). Love your old skool tech videos!
I think that this kind of display was used by Concorde's cockpit to display DME distance and various other things. Maybe CIVA INS systems used them too for their CDUs.
@@FrozenHaxor When you have a reel of NOS transistors, and reels of resistors, why order the ULN array when you already have the internal parts already to hand for essentially zero cost and wait time. Plus you will need 4 chips to do this, with half of one going wasted.
I wonder if the heat from the display operating kind of keeps the unlit filaments warm to some degree, which would then lessen the thermal shock of them being powered up when it is their turn?
It's like all addictions - it starts when somebody offers you a single decimal place. Once you are hooked, you move up to the stronger stuff. Beware of the risks of the 14 and 16 segment alpha-numeric displays. Once you've tried those there is no going back!
Made me think of my doorbell button, which I guess I had assumed was lit by a neon or something, since it was only installed a little over 20 years ago, but when the light went out I took it apart and found that it had a little circuit board inside with a tiny incandescent lamp soldered on it. It was about 18 volts AC, so I replaced it with a red LED, with an appropriate resistor and a series diode just for good measure, so now at night it looks like HAL 9000.
Cool! I am getting a bit well my kids say "over the hill!" now, lost my wife to cancer last summer, a horror I wish on no one, and I am having a hard time getting out of my electric lift chair. So for Christmas my children kicked in and bought me a ring doorbell for Christmas. My grand-son-in-law, a licensed electrician did the install for me, and now I have the ultimate doorbell. It gives me a ring any time there is anyone walking up my sidewalk, and I can press a button on my cell phone to talk to them, even if I am 500 miles away! I love it! For a tech nut like myself it is great fun. I later picked up an Echo Dot so I can use that to communicate over the doorbell with those who come up the walk if I am in my chair, I can invite them in or send them packing depending on who they are, I gave my permission for local law enforcement to tap into it should the need ever arise, and it may well happen since the meth heads moved in across the street and we are treated to their fist fights arguments and reckless driving. The traffic is getting bad now with cars coming and going all night long, but my Ring does save all the movement data on the cloud for 30 days should it be necessary to see what happened say at 3:00 AM on the 23rd of November or when ever. At any rate the Ring pulls it's power from our Doorbell power pack which is 18 volts and hooks up to our wireless network. Everyone should have one, and it comes with a guarantee that if the doorbell outside parts are stolen, Ring will send you a new one, no questions asked, of course they can look and see who stole it, which is also handy as hell. As an old retire cop, I love it!
The FET in that display chip looks like it is wired about the same as the FET that comprises a "current-limiting diode", something I just recently learned of. So, maybe there is no missing voltage feedback or whatever, in the schematic: the FET device, as wired, may intrinsically put out a fixed max current. Or, I may be wrong!
Not sure why? But I find watching people sodder, very therapeutic! So,... Next time Fran,... please video your work... "WOULD VERY MUCH,... LOVE TO SEE IT!" GOOD JOB! 😉👍 Blessings James 🌈
Fran (or anyone else) do you know why the segment ends on Led 7 seg displays and this one too always seem to be mitred off at an angle but the segment ends on VFD 7 segs always seem to curve over a little way to meet the neighbouring segment.
Love these historic display videos. I know you are using what is on hand, but might be a good idea to get some ULN2003's for future projects. They will save a lot of solder time if nothing else :o) You could turn this project into a thermometer that measures its own temperature :o)
I see the soldering iron nicked the insulation of the green wire. I am glad to know I'm not the only one whom occasionally makes the same mistake. Imagine if a display based on this technology was manufactured large enough to display Time and Temperature as the displays from the top of bank buildings from years gone by.
When an incandescent lamp is cold, it has a low resistance and draws more than 20mA. If the lamp is glowing, the current goes back to 20mA. The incandescent lamp cannot get hot enough from a constant current source.
"Incandescent bulbs can be made to last forever." -- Well, they have done a much better job overall than the CF lamps and now LED lamps that light my home. I'm sure the LEDs could be made much better too. But I have an incandescent from the 1970s that's still going and I've gone through at least 4 LED lamps in as many years that are supposed to last for many thousands of hours. I pretty much just buy the LEDs so I can put a much higher output lamp in a fixture and have it run cool. I don't buy them for longevity at this point as the higher wattage ones have failed quite a bit. They build the power supplies like crap in them. Though the LED nightlight that lights up the water dispenser on my fridge has lasted for over 10 years now.
Seems odd to me that an incandescent display should need a heat sink. I mean, the thing actually _works_ at high temps. But I'm betting that the segments are actually _Lucite_ rather than glass. Lucite is much easier to work with, and would indeed need protection from heat.
They funny enough are only rated for 500 cycles of insert and disconnect, while USB is rated for 1000, though the 500 cycles is when the original contact resistance of any contact has increased by more than 5% IIRC, of the original tens of milliohm contact resistance. Typically they will survive 10 000 cycles before the contact wear is bad enough to cause noise, or the actual wires themselves fatigue and break. Those connectors themselves are probably over $100 each, new, now, if you have to buy them from the manufacturer, and not as surplus.
Julian Weinert: An Atmel Mega (www.seeedstudio.com/blog/2019/11/13/atmega2560-features-comparisons-and-arduino-mega-review/) is one of the microcontrollers used in a number of the varieties of Arduino (www.arduino.cc/) boards on the market. The little board Fran holds up at 11:09 is this one: (www.amazon.com/Gowoops-ATMEGA2560-16AU-headers-Compatible-Mega2560/dp/B07Y9P4JL2/ref=sr_1_15?dchild=1&keywords=atmel+mega+board&qid=1618765123&sr=8-15) though people are reporting problems using it. Fran - would you care to discuss how you were able to get it to work? Cheers!
@@stevejohnson1685 I have a cheap chinese ATmega2560 board, works perfectly. Use it a lot to drive led strips. Now working on some python/micropython code to drive a number of SPI mini displays.
High quality micro sockets all the way✊ Usb 🤣 Currently collecting miniature aviation sockets..the soldering is a nightmare but they feel good. Again..usb 🤣🤣🤣
15 minutes in and I finally realized the display was just some number with 111 added to it every cycle.
That number is 069. _Nice_
P
Oh that's it! Thanks
We had a full reactor core display that used miniature-387 bulbs. But it was 2908 bulbs total!! The only time they were all lit at the same time was lamp test, but that was a whopping 3.2 kWatts!! We rigged up a timer so you couldn't leave it in lamp test for more than 10 minutes. Several fans behind the panel for cooling them.
And for our drivers, we had a resistor tied in parallel with the drivers to pass a 'keep warm' current through the bulbs. Selecting a resistor such that the bulb had about 1/3 voltage on it when 'off', reduced the in-rush current when turning on, and bulbs actually lasted a lot longer. (when we upgraded to LED equivalents, we had to remove the resistors since LEDs they passed enough to turn on the LEDs lol )
What a work of art! The angles of the segment boundaries are so sharp, and the gaps between the segments are so small.
It turns out if you're willing to pay MILSPEC/Aerospace prices you can get some pretty nicely engineered things. Makes you wonder what you could get now.
I have something similar in an aircraft instrument, it can be taken apart and the bulbs can be replaced, by just undoing the screws at the back
Not with those sub miniature D connectors, those are basically a single use non repairable item. The connectors also come with all wires the same length, and you cannot replace a pin or socket in them, as they are potted during manufacture. I have had to replace quite a few, unlacing the old one from a wiring harness, soldering in the new one then sewing up the wiring harness again with lacing twine, then applying the conformal coating again to the PCB and the twine knots so they can be protected.
If you want to use the IR camera on pure metal it is recommended to place a peace of black pvc tape on it. Otherwise you will measure the wrong temperature.
You could change the emmisivity but then you have to change it back when measure something else.
Aimed at the black epoxy and glass, so there is a pretty good black body source there.
You, Techmoan and Technology Connections all do a digital / analogue display video this week! The universe is in Synergy 😁
Too bad only Fran and Mat know what they're talking about lol
Fran if it ever stops working please do a post mortem on it to find out how it is made
The "vise of knowledge" touch... :D
@@catalinbadalan4463 Love the Clive reference :)
Yes, a really neat display and your solution is excellent. The issue with incandescent displays and semiconductor drive is the near zero cold filament startup resistance as you indicated. This drove designers nuts and there are many many discrete transistor driver boards running around in the 70's for this and similar reasons. Once VFD and LEDs arrived, driver design because much simpler, could be run faster, and multiplexing displays became feasible.
hey fran! what a great morning surprise! i bought my first electronics kit the other day because of you, i’m excited to build it. cheers!
What did you get?
@@kenmore01 i got a little atari punk synthesizer kit, and i can see myself easily getting addicted to building tiny noise makers.
@@repeat_defender Cool! I love kit building of all types. Electronic, plastic or wooden models, even K'nex lol. I hope you enjoy building it and using it!
A heatsink for just 1.36W total of tungsten lamps? They don't mind running at a high temperature. And why no fins? Maybe the metal block is meant to absorb vibrations to protect the filaments. I'd guess it was mounted on springs or in rubber.
Omg even more salt hiii😂❤️
Ayeee .... I love my subscribed-to channels interacting with eachother 👏👏👏
I've seen those mini db9's and displays in low orbit gear.
maybe then they have only a CNC machining and it is easier to do solid blocks, less time of process.
@@flyguille Idk if CNC was available in 1970.
They would have flycut a 40 or so inch bar and cut into short pieces and mill several at a time clamped together.
Adding fins would jack up the price, and it's not necessary for glass.
Despite being American, I gawked at how crazy it getting to 90° seemed for a good few minutes, until I realized it was Fahrenheit rather than Celcius. Weird how my brain automagically split thinking in °F for weather and cooking and °C for science and EE! Or maybe I was just too distracted by such a -cool- hot display to think about it much, haha. Thanks for sharing yet another wonderful video, Fran!
Heathens and their °F ;) Still, I think Fran should have hooked up a thermometer to the display driver so it would display its own ever rising temperature. That would have been some kind of... electrical waterboarding.
°C s for science...??? For shame... K-shame.. :)
@@catalinbadalan4463 Hey mang, I'm 'Murican, gotta be proud to be above using furlong firkins per fortnight at least :P
We could use °R instead!!! Would you like that rest of the world???
Rankine is just Kelvin using Fahrenheit degrees.
I used to love doing projects like this in my high school industrial electronics class.
A piece of equipment I used in my military days used incandescent 7 segment displays. The difference with this one is the front of the display was removable. You then were able to reach in with a pair of needle nose pliers to replace the grain of wheat lamps. This way you didn't have take the equipment out of service to change the lamps.
It was a calibrated device so if I took the back cover off it had to be sent in for recalibration.
You had me at Tung-Sol. Thanks for sharing this with us. That was when serious lamps were made.
Friend's baby: "Meh" / Micro D connector: "That's adorable!"
I have that same helping hand, I got glued mine to one of those small flooring samples from home depot to make it more stable.
And what a beautiful soldering job it is!!
Interesting old calculator! I recall back in the min 1970's I was serving with the 2/3 Armored Cav down at Fort Bliss Texas. We had a communications building in the Motor Pool area, just a sort of small block house with enough room for our supply of commo gear, and a couple of CONEX containers for the bit stuff, like miles of wire and such. Our section was commanded by a Captain and a Second Lieutenant. These officers, of course had to have some college, unlike we in the NCO corps and enlisted soldiers who did the dirty work. Our Captain had his bachelor's degree in some mechanical engineering with a minor in electrical something or another. During his college career he needed one of those fancy calculators. Of course back then a calculator that could do square root was fancy and his, why his could do trig and crap like that. We were impressed, he would let we peons hold it and press a few buttons so we could OOOH and AAAH a bit, then it was back in his leather pouch and locked in his desk that sat in the southwest corner of our shop. Cool to see one of those back doing it's duty once more!
These old displays are super cool, but it's also remarkable how dominant LED technology has become. If something emits light it's almost certainly some kind of LED, or the sun.
With the current LED trend , I'd say even the Sun's incandescent days are numbered! :)
90 doesn't seem too hot for several incandescent globes. It looks very well constructed.
Thanks for showing this beautiful display off.
One learns to appreciate the invention of the LED, almost as important as the transistor.
Very nice looking display. When building a high-end piece of hardware required real engineering, both electrical and mechanical. Considering the dozens of steps required to make this, I'm guessing they were never produced in large quantities, and they must've been pretty expensive! Even machining the block of aluminium to the required precision would've been challenging, since they probably didn't have the kind of CNC machining systems used for this sort of thing today. Each step would require a different machine with a different jig, and an operator doing a single task, e.g. drilling the mounting holes, or tapping the threads.
That display is a thing of beauty.
I needed this one today. Always glad to see Fran.
Hi Fran. What a great "little" build! And a very cool - well, warm given the heatsink temperatures - display. You continue to amaze.
You're wonderful Fran! Thanks for the video
9:28 I am in awe of Fran's exceptionally clean and beautiful soldering. 🌸
That is just cool... definitely military or aerospace.
On incandescents lasting a long time...I have a clock radio I got at a yard sale in 1976 with a incandescent dial lamp. Still works. Other than moves and power failures, it's been on since then.
maybe you could run a constant current source through one filiment to get an idea of how much the cold filament current changes to hot filament current. start with 1/10th ma and measure Vdrop then working current. the circuit uses a linear 3 term reg and a resistor.
The BCD decoder is designed with the CMOS gates sized to limit the base current in the NPN transistor. The 20mA is only available when you have a red LED with a 1V2 forward voltage drop when lit, as this them will get 20 mA drive, reasonably consistently across each particular die, and reasonably close per wafer lot.
The most common way to drive those displays, along with the Numitrons, was to use CD4049/50 COS to TTL buffer chips, as they are rated to deliver 50mA minimum source or sink current, into a TTL logic 1 or logic 0, nominal voltage of 3.4V or 0.8V respectively, with a 5V supply. Just needs to have a single extra gate per digit, or two if you were also driving the decimal points actively as well.
I would love to see one of these cut open, wish there was a way to do that without destroying it
I guess a CT scan would do it... One can dream!
They must've cost a fortune to make! Would've been so awesome as a frequency display on a high-class early 70's hi-fi reciever.
Re-engineering your driver circuit on the fly to include a buffer amp is frankly above and beyond the call of duty! You exhibit a degree of determination and patience I can only dream of!
It's gorgeous. I was going to complain, I can't find one. But seeing how much trouble you went through to get yours working, It's fine. I don't need one. LOL. (I appreciate the work you put into these demonstrations, I would'nt have the patience.)
Great video as always. Thank you Fran.
FYI
If you check Chicago Miniature Lamps CML-IT, a lots of 5V filament lamp have a 40,000 hours average lifetime.
For example, a T-1 #680 is rated at 60,000 hours.
+4 volts was the operating voltage for RTL, Resistor Transistor Logic. A logic family made by Motorola in the form of standard DIP chips. A forerunner to TTL. Then there was also DTL, Diode Transistor Logic.
And IIL (Integrated Injection Logic). Also written as I2L or I²L.
That was meant as an expansion of your last sentence ("Then there is also..."). IIL does not use 4V. It works on current, not voltage.
I think you forgot about the warm up current. Typical incandescent lamps have about 1/10th the resistance cold, as they do warm. There's a huge current draw when first powering up, which stabilizes after it is emitting light. It would be interesting to watch the current with a scope on a single digit to see how quickly these light (they seem quite "snappy"!) and what the "inrush" is really like.
the 2N3904 is a solid choice of NPN transistor :D
I am very impressed with her reel of them.
when i was a lad the BC108 / 9 and the BFY51 were my faves :)
Here in Europe the bc547 is pretty common, I believe very similar to the 2n3904
2N2222 is popular as well. For power transistors, the TO-3 2N3055 was widely used in power supplies and power amplifiers.
@@jukingeo I watch so many of these videos and like to build some of the things that are presented. Many say you can replace the BC547 with the 3904 and I was doing that, but the pinout is different and I am a bit dyslectic. To solve the problem was quite easy Ebay has sellers that sell both so I now have boxes of 547's 2222's 2904's and all their p channel mates so I no longer have to do the conversion of pinouts and screw up my projects. The cost of these is really negligible, perhaps a few cents to a dime each so in today's money it's simply the price of a cup of coffee (which I LIVE ON!) for 100 of each.
I can hear Dave Jones comments about all those transistors flapping around in the breeze hahahaha
Your collection of exotic displays is quite something, I had no idea there where so many weird implementations
The late 60s - early 70s were probably a quite wild era... :)
@@evensgrey I guess there's also MicroLED and possibly DLP? I'm not sure how old DLP is.
(And yes - amazing indeed!)
Edit: E-ink as well!
Micro Ds are extremely rugged, but they are unforgiving if you misalign them when plugging them in. Within the company I work at, the instructions are only to mate them by engaging the jackscrews and turn 1/4 turn max at each end alternately.
Is this a similar display that was used in BttF for the MPH display in the Delorean and on Doc's remote control? Those looked like they were incandescent 7-segments to me.
Those BTTF displays were Stanley - th-cam.com/video/LvJN4Maea9I/w-d-xo.html
It appears that the inability of the CD4511B to sink 20mA can be explained by Fig. 1 on page 2 of the TI datasheet. It shows the output characteristics of the FET used for sinking current. There are 3 curves on that plot. The only one that reaches 20mA is the top one which uses a 15V supply. Even 10V isn't enough to sink 20mA. With 5V, the most you can sink is 4mA, but at that point you have 5V of Vds so the external supply for the lights will need to be at least 9V for 4V lights.
With a 15V supply you can turn on the FET hard enough to sink 20mA, but Vds will be about 6.5V. For 4V lights, that means you need at least 10.5V to run the lights, or a resistor (15-10.5)V/20ma = 225ohms between the 15V supply and the light to limit current / voltage to each segment.
Looking at the static electrical characteristics, this part was really designed more for SOURCING current rather than SINKING current.
At 5V, it can typically source 20mA with an output voltage of 3.75V, minimum 3.4V. Not quite enough for the 4V lights here. However, at higher supply voltages, you can get higher voltages out for the same 20mA. So it looks like a supply of about 6.5V for this chip would allow you to SOURCE 20mA well for 4V lights.
Right. I believe Fran was sourcing not sinking but I'm not sure. The On Semi data sheet shows the part will drop about 1.5V when sourcing so I am assuming she was using Vdd = 4V not 5. So yes 5.5-6V supply SHOULD work, but we don't know what specific device she has.
@@michael931 She is sourcing current, the decoder is direct driving, and does not have active low outputs. The sink transistor there is to prevent display ghosting when driving multiplexed displays, actively pulling leakage current, from either other segments not selected or from the chip itself, to ground so the digit drives are at logic 0 levels when not enabled for a digit. however the display type it is really designed for was the original Monsanto red LED arrays, where the forward voltage drop is 1V2, with the multiplexed common cathode arrays, using saturated NPN drivers, so the CE voltage drop was around 0.2V extra added to the display.
@@michael931 From her explanation at 8:00 to 8:40, it seems clear that she is trying to sink current through the pull-down FETs. The explanation about current *sensing* is incorrect. Current sensing can not be done by looking at the gate voltage, and the output pins aren't fed back in anywhere. The logic simply drives those gates high or low.
However, it IS being current limited by the FETs, basically because their Rds_on is so high that it leads to a high Vds voltage when the FET is on, as shown by the curves in Figure 1 of the CD4511B datasheet (that part number is visible at 7:58).
@@SeanBZA Good point about the active-high outputs - I didn't check that before looking at the current limits. Sourcing current is easier than sinking with this part, but she would need to power it with more than 5V to get 4V & 20mA out of it.
@@michael931 You are right, she would have to be sourcing current to drive the lights - her focus on the pull-down FET made me think she was sinking current through them to turn the lights on. But this part is designed to drive the outputs high (and source current) to turn them on. I should have checked that before focusing on the FET limits!
Kind of a funny side story...when I worked at a mobile product company, one of .y colleagues related a story about looking at a product from a Swedish company, in which they accomplished LED dimming by actually "stealing" current from the LEDs, just 'burning it off', so we always joked about "Swedish dimming" .... imagine needing a heatsink for the dimming!
Quite an impressive display technology. Great job at making the driver.
Looking forward to incandescent bulb video.
The arcade I went to until recently had big incandescent 7 segment displays where you could see the bulbs
Hey Fran, a simpler interface is to string 3 X tpic6c595 chips (same as 74hc595 but has open drain outputs capable of 100mA pin sink) in series hanging off the spi port or can even be bit banged out. Much less soldering 😜
If only an oscilloscope was more affordable. I want to know what that initial voltage spike is in each of those bulb segments...
There are a LOT of videos available on how to build a small oscilloscopes on TH-cam. I have built a couple that work fairly well. There are also small kits for around 40 bucks that you can build, or if you lack the skill to build they can be had in prebuilt form where you simply plug the screen to the circuit board, and if you ordered it, build the little plastic case to protect it. I have built two of those, they both do the same job but I like the larger one simply because it has a bit larger screen. I did build on that plugs into the sound card on your computer, you can download freeware that does the display for you. I feared frying my sound card so I also purchased a small USB sound card to protect my computer but after a time I found it not necessary. The build has two 7805 voltage regulators to keep the output to your PC under 5 volts no matter what you plug into so it does a fairly good job of protecting things. I don't think I would ever plug the probes into a 110 or 240 outlet but for work on digital circuits it works just fine, you can also do a print screen on the output and actually print out a sine wave if you want to, comes in handy when trouble shooting and a saved screen shot can be sent to a more knowledgeable Elmer if you have one so he can look over your problems. KE0JBL
Indeed, the incandescent lamps have a strong increase of resistance at higher temperatures. Which is actually a saving grace for their longevity. I was for a long time using a rule of thumb that the resistance went up by a factor of ten from room temperature to the illumination temperatures.. The actual change is close to 90-fold. Think of this as self regulating feature against operating voltage variation. Not perfect, but helps the lamp operation. Compare that to LEDs that require constant current supply rather than constant (or loosely regulated) voltage.
@ 10:20 .. Now I don't feel so bad. 2 day build? 20 hours into my Nixie tachometer and it's still not flying straight. I thought I was getting old.
Awesome, indeed, both the display and the Goldbergian driver.
Those mini 2560 boards are great. Especially for MIDI merge boxes.
Wow incandescent 7 segment, never heard of! Much look beautiful IRL. I'd have been so tempted to just solder another 4511 on top of the original one to bump up the current...
My goodness , a whole transistor per segment to supply power. That's insane compared to standard LED 7-segment displays that only sink 5-20mA per segment (standard outputs can handle that current). Love your old skool tech videos!
Apparently the chips on the middle layer weren't giving out 20mA on all segments, closer to 12mA take a look at 7:00 to almost 10:00
I think this is a good time to say: "don't take it apart, turn it on!"
Great choice of car! I wonder why those don't get te resto mod treatment, sush a candidaye for it
I really enjoy the pile of circuit boards, this is exactly how I would end up doing it.
I think that this kind of display was used by Concorde's cockpit to display DME distance and various other things. Maybe CIVA INS systems used them too for their CDUs.
You must surely know about the ULN200x ICs right?
Why do it with discrete transistors?
...and resistor arrays!
@@stevejohnson1685 with the uln ICs you wouldn't even need them!
Yeah that's a weird choice of design.
@@FrozenHaxor When you have a reel of NOS transistors, and reels of resistors, why order the ULN array when you already have the internal parts already to hand for essentially zero cost and wait time. Plus you will need 4 chips to do this, with half of one going wasted.
I wonder if the heat from the display operating kind of keeps the unlit filaments warm to some degree, which would then lessen the thermal shock of them being powered up when it is their turn?
Please post the specs for the Tung-Sol DT 1511 and/or bulletin T430 for us to follow along at home, particularly the current requirements!
I have an addiction for 7 segment displays.
It's like all addictions - it starts when somebody offers you a single decimal place. Once you are hooked, you move up to the stronger stuff.
Beware of the risks of the 14 and 16 segment alpha-numeric displays. Once you've tried those there is no going back!
@@johncoops6897 I have an Alpha Clock Five which uses 5 16-segment displays 🙂
I'm doomed, lol.
Made me think of my doorbell button, which I guess I had assumed was lit by a neon or something, since it was only installed a little over 20 years ago, but when the light went out I took it apart and found that it had a little circuit board inside with a tiny incandescent lamp soldered on it. It was about 18 volts AC, so I replaced it with a red LED, with an appropriate resistor and a series diode just for good measure, so now at night it looks like HAL 9000.
Cool! I am getting a bit well my kids say "over the hill!" now, lost my wife to cancer last summer, a horror I wish on no one, and I am having a hard time getting out of my electric lift chair. So for Christmas my children kicked in and bought me a ring doorbell for Christmas. My grand-son-in-law, a licensed electrician did the install for me, and now I have the ultimate doorbell. It gives me a ring any time there is anyone walking up my sidewalk, and I can press a button on my cell phone to talk to them, even if I am 500 miles away! I love it! For a tech nut like myself it is great fun. I later picked up an Echo Dot so I can use that to communicate over the doorbell with those who come up the walk if I am in my chair, I can invite them in or send them packing depending on who they are, I gave my permission for local law enforcement to tap into it should the need ever arise, and it may well happen since the meth heads moved in across the street and we are treated to their fist fights arguments and reckless driving. The traffic is getting bad now with cars coming and going all night long, but my Ring does save all the movement data on the cloud for 30 days should it be necessary to see what happened say at 3:00 AM on the 23rd of November or when ever. At any rate the Ring pulls it's power from our Doorbell power pack which is 18 volts and hooks up to our wireless network. Everyone should have one, and it comes with a guarantee that if the doorbell outside parts are stolen, Ring will send you a new one, no questions asked, of course they can look and see who stole it, which is also handy as hell. As an old retire cop, I love it!
That's really awesome!!!
The FET in that display chip looks like it is wired about the same as the FET that comprises a "current-limiting diode", something I just recently learned of. So, maybe there is no missing voltage feedback or whatever, in the schematic: the FET device, as wired, may intrinsically put out a fixed max current. Or, I may be wrong!
7 Segments? I love that band!
Not sure why?
But I find watching people sodder, very therapeutic!
So,... Next time Fran,... please video your work...
"WOULD VERY MUCH,... LOVE TO SEE IT!"
GOOD JOB! 😉👍
Blessings James 🌈
Fran (or anyone else) do you know why the segment ends on Led 7 seg displays and this one too always seem to be mitred off at an angle but the segment ends on VFD 7 segs always seem to curve over a little way to meet the neighbouring segment.
Hahahah yeah, nailed it, 7 seg display driver but not a VFD as I thought it might be.
Love these historic display videos. I know you are using what is on hand, but might be a good idea to get some ULN2003's for future projects. They will save a lot of solder time if nothing else :o) You could turn this project into a thermometer that measures its own temperature :o)
Can you do a tear down on the electric eye from the mail bag
It's look a good place for ULN2003 chip
It's possible the lamps have the evacuation nipple on the back end and a magnifying tip like that found in a battery operated pen-light.
I see the soldering iron nicked the insulation of the green wire. I am glad to know I'm not the only one whom occasionally makes the same mistake.
Imagine if a display based on this technology was manufactured large enough to display Time and Temperature as the displays from the top of bank buildings from years gone by.
Really nice display.
Counted the 21 transistors, but what are you using to drive the decimal point? That would be 22.
You don't need a transistor for something that is permanently on. Just wire it directly to the ground.
@@olmostgudinaf8100 But in her video, it's not permanently on. It is flashing.
I had an extra digital output pin on the Arduino and used that as a sink.
When an incandescent lamp is cold, it has a low resistance and draws more than 20mA. If the lamp is glowing, the current goes back to 20mA. The incandescent lamp cannot get hot enough from a constant current source.
Wow no planned obsolescence in a bulb . Very nice
My new favorite host and channel! Thanks Fran!!!! SUBSCRIBED!!!
FRAN, watch episode 1 of the show "UFO" - it has tellites, and some of your more rare display types. 😊
Tung-Sol really made some heavy-duty equipment. I have a lot of their tubes in my guitar amps, and they seem to last forever:
14:45... why not chuck an oscilloscope onto it see what the turn on current looks like
Cool! Whenever you get the chance to have an X-ray image made of this thing, pleeeeaase take it :-D
Going to be hard with all that aluminium there though, but doable.
don't have a reel of ULN2003s?
Today I learned you can buy transistors by the reel
Waiting patiently for the bulb life video
No closed captioning?
Great video
It would be interesting to see an xray of this device.
track name in the end?
I love incandescent and Nixie tube👍
im getting NASA vibes from this lil guy.
Heat sink draws the heat into the aluminum block
"Incandescent bulbs can be made to last forever." -- Well, they have done a much better job overall than the CF lamps and now LED lamps that light my home. I'm sure the LEDs could be made much better too. But I have an incandescent from the 1970s that's still going and I've gone through at least 4 LED lamps in as many years that are supposed to last for many thousands of hours. I pretty much just buy the LEDs so I can put a much higher output lamp in a fixture and have it run cool. I don't buy them for longevity at this point as the higher wattage ones have failed quite a bit. They build the power supplies like crap in them. Though the LED nightlight that lights up the water dispenser on my fridge has lasted for over 10 years now.
Four-stair cooling!
That looks like a mini version of serial port
I was really thrilled by your manufacturing hypothesis.
Seems odd to me that an incandescent display should need a heat sink. I mean, the thing actually _works_ at high temps.
But I'm betting that the segments are actually _Lucite_ rather than glass. Lucite is much easier to work with, and would indeed need protection from heat.
Pretty sure the mini-D will survive a few more insertions than the average USP plug
They funny enough are only rated for 500 cycles of insert and disconnect, while USB is rated for 1000, though the 500 cycles is when the original contact resistance of any contact has increased by more than 5% IIRC, of the original tens of milliohm contact resistance. Typically they will survive 10 000 cycles before the contact wear is bad enough to cause noise, or the actual wires themselves fatigue and break. Those connectors themselves are probably over $100 each, new, now, if you have to buy them from the manufacturer, and not as surplus.
What is a "Atmel Mega board"?
Would be nice to know what product exactly that is...
Julian Weinert: An Atmel Mega (www.seeedstudio.com/blog/2019/11/13/atmega2560-features-comparisons-and-arduino-mega-review/) is one of the microcontrollers used in a number of the varieties of Arduino (www.arduino.cc/) boards on the market. The little board Fran holds up at 11:09 is this one: (www.amazon.com/Gowoops-ATMEGA2560-16AU-headers-Compatible-Mega2560/dp/B07Y9P4JL2/ref=sr_1_15?dchild=1&keywords=atmel+mega+board&qid=1618765123&sr=8-15) though people are reporting problems using it.
Fran - would you care to discuss how you were able to get it to work?
Cheers!
@@stevejohnson1685 Cheap knockoffs require a CH340G driver installed separately, then they work fine.
@@stevejohnson1685 I have a cheap chinese ATmega2560 board, works perfectly. Use it a lot to drive led strips. Now working on some python/micropython code to drive a number of SPI mini displays.
High quality micro sockets all the way✊ Usb 🤣
Currently collecting miniature aviation sockets..the soldering is a nightmare but they feel good.
Again..usb 🤣🤣🤣
That's awesome
Fascinating display, if it ever does (hope not) fall off the twig please autopsy it !...cheers.