#1254

Episode 1254 chip of the day
CD40106 Hex Schmidt Trigger Inverter
Be a Patron: / imsaiguy

[@PhoenixRevealed]

Be aware that if you order a cheap CD40106 from China you may get a repainted 74HC14A instead. If you plan on using it at 5V or less this is fine, but unlike the 74Cxx and CD4000 series chips which have an absolute maximum VCC rating of 18V, the 74HC ones are only good up to 8V. I designed and built an automotive control circuit using a "CD40106" and couldn't figure out why it worked perfectly on the bench (using a 5V supply) but kept failing completely after just a few hours of use in the car at 12V (or 13.8V). It took me a while and a half-dozen fried "40106's" to figure out that my 40106 wasn't actually a 40106. The vendor had simply rebadged 74HC14 chips as CD40106/74C14. Caveat Emptor.

[@Spark-Hole]

I also buy from China, found a few fakes, but it works fine no problem. I still doubt your case because it too easy to spot. If it is more sophisticate chip may be.

[@daviddickey9832]

A lot of this going on

[@joseppuig925]

The 74HCU04 which is the un-buffered version of the 'HC04, is praised as capable of low phase noise when used as crystal oscillator. Also, these un-buffered cmos inverters are quite versatile as they behave quite linearly when biased as analog inverter amplifiers.

[@ericksonengineering7011]

Great video, thanks. Another 4000 CMOS trick is tto parallel gates to provide more drive current / lower source impedance. You can wire 2..6 of the inverters in the package in parallel. Good for power circuits, FET driving, speakers, more precision for analog circuits...

[@AnalogueGround]

Always look forward to ‘chip of the day’ and it’s always good to have a refresher on the 74 and 40 series ICs. Much appreciated.

[@jj74qformerlyjailbreak3]

Me too.

[@davidjh7]

I know I'm being pedantic here, but the original CD4000 series was made by RCA, not Motorola. The number letter designation was C, representing RCA, and the D was for digital. Similar designation to RCA's CA series of analog chips, C for RCA, A for analog. The CD4000 series was probably my favorite logic series, even though it was slow. The CA series analog chips, now pretty much all obsolete, were also some of my favorite analog chips. That series had some chips that had better specs than replacements you can get today, like in some of their electrometer op-amps, and some of their transistor arrays.

[@marko.692]

Few years ago I built oscillator using the CD40106 and 100kHz Xtal. I added 10kohm resistor in series with the crystal, and 1Mohm resistor parallel with inverted input and output... the circuit started to oscillate.

[@AlessandroAllegretti]

My recent use of CD40106 is for a one-button switch, configuring a couple of inverters as a flip-flop. This chip also suites well in anti-bounce push button circuits. My 2 cents 👋😉

[@nigeljohnson9820]

I find the the 74HC132 quad Schmitt nAND a more useful component, as it provides a control signal on each gate. It really is a very versatile part, as two gates can be used to produce all the versions of mono-stable or linked together to provide a modulated oscillator output, ideal for sounder alarms. Not keen on using a Schmitt in Pierce crystal oscillators, as they have a habit of not starting or going into RC oscillation, simply because they can. The old come 8085 processor had an internal Schmitt oscillator, due to a production fault, when used with a crystal, internal input leakage current tended to take the oscillator input bias beyond the point of oscillation, beyond the hysteresis voltage window, result the oscillator would stop for no apparent reason. I have used the the HC132 as the basis for a rather simply but functional LED stroboscope, with one gate providing the oscillator function, and two providing a pulse width limiting monostable. Feedback, using a low current N channel mosfet, reduces the pulse width, maintaining an apparent constant light output. The output Less, are four, 6watt, white light, parts, driven as two * two series connected in a network with separate current limit resistors, using a power mosfet. The average supply current is very low, but the peak led current is close to their maximum rating. Using a BPX65 photodiode, it was possible to confirm the light output pulse shape to be a good narrow square wave, with a minimum width of 0.6us, and a repeat rate of up to 20KHz.

[@julianrosas4992]

Hi good morning, great vid as always, I made a circuit back in the 80's with a CD4093B wish also is an oscillator chip but lost the schematic, it was a Roger Beep circuit with two IC's double tone, do you have anything like? Will be nice to see to remember. Have a great day. Peace

[@arthurharrison1345]

One part they never made in the 4000 series was a dual D flip-flop with Schmitt clock inputs. There's a version called the SN74HCS74 that will operate up to 6 volts, but nothing available for higher voltages. Such a part would have found millions of applications.

[@davidjh7]

Just as a suggestion, one simple, but very interesting CD4000 chip you might want to feature, is the CD4007 complementary pair plus inverter. It's been described as a DIY "kit" for interesting cmos circuits. I think Bob Pease back in the day described a cool little op-amp circuit using he 4007. Just a thought.

[@dalemettee1147]

The military version of IC's was a ceramic case and the consumer version was in a plastic case. Of course the military version was way more expensive, no surprise there.

[@brianharper9798]

With a slightly different configuration, as most synth builders will probably know. You can generate a very nice ramp (sawtooth) waveform. And with the addition of a few more components it can be made to track at 1v per octave very nicely....

[@bayareapianist]

Because of their input voltage hysteresis oscillators made by triggers create inherently non-symmetrical square waves. If you want to 50/50 symmetrical square waves you should use two gates. Schmidt triggers are ideal for low/very low frequencies because you can take advantage of their hysteresist. For above 1 MHz you should use either Gates or op amps. Since you have six Schmidt triggers in a chip you can measure lithium ion batteries voltage indicators up to six LEDs.

[@mikeh020011]

Hi This took me back to the late 70's.When I left school in the UK, my first job was in ship to shore radio mf hf vhf and am air band.I used to repair the radios and some of them used TTL logic. I still have my first editions of Don Lancaster TTL and Cmos cook books. We started to use 54 series logic as they were mil spec , more expensive but cut the fault rate down. The manufacture of the tx's and rx's started to replace most of the 74,s with 54's

[@johnwest7993]

I'm going to guess that the 100 kHz xtal had more mass and thus needed more drive than the CMOS gates could provide. You might try paralleling gates to drive the xtal, and also playing with that 51k resistor value.

[@aerofart]

I’m willing to bet that the 100kHz crystal would work if you cut down those extremely long leads. Not only do they act as inductors, but also as antennae.

[@AnalogDude_]

did you try with the capacitors? maybe the casing needs to be grounded as there is a thing on top of the case, sometimes you see oscillator cases soldered to the PCB with a wire. Darn these are huge!

[@jp040759]

I love these 4000 series CMOS devices. I still go back to them for essentially "no current draw" circuits when I need something that will run on batteries forever.

[@richardl6751]

I successfully built a 100kHz oscillator with a crystal (that looked very much like that one) using a 7404 with three gates wired in series. Pins 2 and 3 connected together, pins 4 and 5 connected together and the crystal connected from pins 1 and 6. It seemed more stable using three gates instead of one.

[@joeteejoetee]

The 100 KHz crystal is likely a parallel resonant "Cut" - the Osc will likely work with that Xtal if you place it between ground and the input to the 1st Inverter, with a small feedback cap from the output to the input of that 1st inverter.

[@jj74qformerlyjailbreak3]

That’s the biggest xtal I’ve ever seen. Wow you could house a whole radio station in that thing now a day’s.

[@IMSAIGuy]

I had a bigger once but can't find it. was in a glass vacuum tube

[@jj74qformerlyjailbreak3]

@@IMSAIGuy cool can’t wait to see it. Probably so big your not looking in the right place. 😂

[@PhoenixRevealed]

While you might be right about the cut, it shouldn't matter to get it oscillating. All crystals will oscillate in both modes, just at slightly different frequencies. When you order a crystal you need to specify serial or parallel mode so it will oscillate at the correct frequency in your intended circuit, not because it won't oscillate at all in the other mode.

[@joeteejoetee]

@@IMSAIGuy Me-too! I got to field re-cycle a 1970's (huge) "Office Facsimile Machine" and I found this curious large metal box inside, with few wires going to it, so it was easy to harvest, and I took it home. Taking it apart I found it was filled with that old yellow foam, and centered inside was an aluminum cylinder with a few small wires going inside of it and 2 large wires wrapped around it in a coil. Never seen that before in Radio or TV sets. I crumbled out all of the foam and extracted the cylindrical core (which was a hollow cup ~4x1.5" OD). I unscrewed the cap screws and found a 100Khz Glass Chrystal insulated inside. Later I found out that it was an oven for the Xtal. I still have that Xtal V-tube somewhere in the stacks, and hope that I can find it!

[@arthurharrison1345]

I don't recommend using solderless breadboards for high-impedance circuits like this. The stray reactances will change the parameters appreciably.

[@crabby7668]

Weren't these called relaxation oscillators back in the day?. Very simple and commonplace, but drift with temperature chahges.

[@tarek13ch666]

Hi Great video I guess the 100KHz crystal need to have the GND on the shell

[@youtuuba]

I have restored quite a few vintage digital multimeters, and many of them use A/D chips or chipsets that use negative voltages are part of their biasing for the LCD (if that is the display they have), or sometimes just because the chip(s) need negative voltages for some reason. That often drives the use of peripheral logic ICs having their power pins connected in odd ways, although they always end up still have correctly 'polar' supplies if measured only across their supply pins, e.g. their VCC pin might be connected to a -2V supply while their GND pin is connected to a -7V supply, so they still 'see' 5V power of the correct polarity in their own little context. This is fine as long as any other signals going into and out of those logic ICs does not interface directly with other circuits using positive supplies. And for what it is worth, in this video a common mistake is made regarding the word "simplistic", which correctly means something along the lines of 'something complex that is mistakenly viewed as being simple', OR the mindset that tends to view things in an overly simple way (not giving their complexity enough consideration). "Simplistic" is NOT a synonym for "simple", as they have almost opposite meanings.

[@alnoorratansi9364]

will a 32.767khz watch crystal work with this circuit ?, I have bunch of unmarked crystals and would love to build a test circuit which would cover a wide range and output the clock to a frequency counter ( your earlier project). which I am building now.

[@richardl6751]

It would probably work with 32,768 Hertz crystal and your other crystals also.

[@robballantyne3]

A question on a completely different topic, if I may. Is there a better place than here to ask questions unrelated to the video content? I'm designing an LC filter (well using web based design software). If I measure its insertion loss performance (scope on load resistance) and it seems just OK. A little thought on the subject though seems to indicate that since matched 50ohm systems have 50ohm loads and 50ohm source impedances that the output voltage of even a 'piece of wire filter' between the 50ohm source and load would imply a 1/2 voltage at the load (and 1/4 power into it). This would amount to a minimum of a 6dB insertion loss. All filter specs I've seen show 0dB loss. Is the 0dB a relative kind of thing? What measured IL would you expect to see on a hand constructed LC filter? Perhaps a more important question, my wife asks: when will ImsaiDog will make a reappearance? Thanks, Rob

[@IMSAIGuy]

what you are measuring is S21. The system is calibrated at 50 ohms so a 50ohm signal generator set to 1Vpp will output 2V unloaded. if a 50ohm load it attached you get 1V. if you put a coax cable between the generator and load you will measure a loss due to the cable (S21). the filter would be in the same place as the coax. I have lots of videos measuring filters. I promise ImsaiDog will be in some videos in the future. Just haven't gotten any footage.

[@robballantyne3]

@@IMSAIGuy Thank you. Much appreciated. The mentioned videos helped.

[@shadimurwi7170]

Microchip product maybe is not available around my city and just let you know they use atmel product for some electronic project

[@andymouse]

Microchip have bought Atmel now.

[@ats89117]

You should test out the 100K crystal S21 with a VNA...

[@codelicious6590]

If the crystal matrix were fractured inside the case, could that cause it to simply not work or would that have different effects?

[@IMSAIGuy]

Wouldn’t work

[@__--JY-Moe--__]

gadzooks Holmes! I believe U'v got something!! crystals R magic!! Krystals sold white castles! OMG!! those crystals R massive!! they need 10MW to work!! car 54, I found U! well, parts of U! ha..ha..

[@Sarahbuildsstepsequencers]

I don’t know why so few people use CMOS over TTL.

[@IMSAIGuy]

I did all of my designs in the 1980's in 74HC CMOS. No one I know used TTL at that point. the 74C parts had their place also

[@onecircuit-as]

Hey I just did a vid with the same chip! Snap! Great minds…👍😀

[@nonsuch]

When I first started out with electronics, the CD40106 was one I played with a lot. I found it has a great sounding "square" wave that's closer to a ramp/saw. Pretty much half way between a square and a ramp/saw. Unfortunately, I haven't found a good way to keep it stable enough for analog synth use without using a micro of sorts. It makes a great 6 oscillator drone machine tho. One that you can have them modulate each other if desired. Not bad for a cheap single chip.

[@lmamakos]

I've got an old Fluke 8040A DMM that has a similar postive-ground design. It was surprising when I found out; I was in there replacing the very old 7-segment LCD display by adding a retrofit graphical LCD display and microcontroller that wiretapped the signals to the display. It was really... curious, but I suppose made other stuff work out. It was nice that Fluke had full schematics and theory of operation chapter in their manual.

[@MyProjectBoxChannel]

I've used them for simple pwm LED-dimmer/ DC motor driver circuits. It's the same capacitor relaxation oscillator circuit but with diodes and a potentiometer, to control the pulse width.

[@Manf-ft6zk]

At 2:36 you gave the input a bias and a chance to oscillate, any MOhm resistor will help. When the oscillator works with around 100kHz you have the chance to bring it to 100kHz inserting the quartz.

[@TRONMAGNUM2099]

Thank you for showing the breadboard close up. makes it much easier for me to try and duplicate your work on my own bench.

[@jagmarc]

This is a religion!

[@petersage5157]

Why 51K for the feedback resistor? Why not an E12 value like 47K or 56K?

[@IMSAIGuy]

I'm sure those would be fine too

[@petersage5157]

​@@IMSAIGuy Audio filters, right? I usually see 5K1, 51K, and 510K resistors in Sallen-Key filters in guitar amps (with commonly available E6 Juliet tolerance capacitors, they give corner frequencies that are relevant for guitar); I'm guessing you have these values on hand for similar applications, like your ongoing synthesizer project. Anyway, I was just curious and wondered if there was a specific reason for using that value here.

[@IMSAIGuy]

@@petersage5157 I rechecked the HP schematic. it was 56k. when I breadboarded the circuit I just grabbed a close value. I later found out the only problem with my circuit was I needed to ground the xtal case.

[@GeorgeGraves]

I'e been folowin you for a bit - know HC wasn't volt tallearent (static?) Negatve? Can you please do a run down on how to use these old "cheap chips - but still reelant" Today? I love the gray beards - but they seem to keep their fa chips a ceret - and when not to use them in 5v or 3.3 logic. Noooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo?

[@IMSAIGuy]

cheap microcontrollers have all but killed any digital design. Teaching fan out and karnaugh maps would be like teaching long division when everyone has a calculator in their pocket.

[@frankowalker4662]

I did'nt know they were the same IC. Thank you.

[@BaldurNorddahl]

Much easier to use ready made oscillator packages. Look like a crystal but has four legs. Just feed it power and it outputs the required frequency. The extra cost is insignificant.

[@mikebond6328]

Easier = not as much fun

[@byronwatkins2565]

I suspect the 100 kHz is defunct. You might try 51 k --> 10 k though.

[@IMSAIGuy]

I should put it on the VNA and see what it looks like

[@henceldeanon9233]

Sir... can you extract a triangular and sinusoidal wave with your circuit like a function/signal generator circuit?

[@IMSAIGuy]

No

[@henceldeanon9233]

@@IMSAIGuy Thank you for replying, sir!!! But, can you add additional circuits so that it can give triangular and sinusoidal waves and perform as signal generator, sir? Because, 1MHz is a very good output and the oscillation is balance since it is a shmmitt trigger circuit.

[@IMSAIGuy]

@@henceldeanon9233 sorry, this circuit cannot be modified to do what you need.

[@henceldeanon9233]

@@IMSAIGuy Greatly appreciated, sir! Thank you very much for the information! Hoping for more good circuits in the future!

[@hanelyp1]

You'd need a high impedance tap to read the downstream side of the crystal. I'm not sure what waveform it would deliver.

[@sorcererstan]

I've wondered why some crystals have the case connected (to ground?) like that large 100kHz one. Is that a necessary connection for it to work? Or is it just to ground the shield?

[@IMSAIGuy]

just shielding I believe

[@aerofart]

Possibly RF ground. At UHF and above, those long leads (and the casing) can act like a tuned circuit.

[@mardav1545]

All of the 74xx chips that I have can't stand more than 6 volts.

[@IMSAIGuy]

it is the 74C series that is high voltage 74AUC series is only 2.5V en.wikipedia.org/wiki/7400-series_integrated_circuits

[@PhoenixRevealed]

74HC is 8V absolute max, but 74C and CD4000 are 18V max.

[@t1d100]

Good one. Thank you.

[@christophernetherton9389]

10 pF either side of xtal to gnd?

[@IMSAIGuy]

www.diodes.com/design/support/perspective/crystal-oscillators-simple-low-cost-and-highly-accurate-clock-sources/ www.gadgetronicx.com/square-wave-generator-crystal-oscillator/

[@POLMAZURKA]

PUF?

[@IMSAIGuy]

real engineers don't say pico farads

[@frankbeckman2663]

For the 100Khz part, try a lower resistor value

[@IMSAIGuy]

do you know that, or a guess

[@frankbeckman2663]

@@IMSAIGuy A guess

[@andymouse]

Cheers.