I think it's a great ic to make a theremin musical instrument. If the capacitor from one oscillator is connected to a antenna, yo can change the freq with waving your hand. Then the combined wave forms produce a frequency in the audio range.
"Dating myself here! Maybe. But at least you won't give someone a funny look and a "Huh?" if they mention a negative 5.2 volt power supply. It has been years since I've looked, but there were still some ECL parts being made at the time. PECL was in use for a time. I don't know if it still is. There was a rather ancient ECL triple fully differential line receiver that made quite a good amplifier and found use in things like frequency counters. I used to have and ECL board (from a DEC CPU of some sort, iirc) that I used as a show-and-tell piece because it had little loops of tracks on the PCB for time delay matching and even a area with a sort of zigzag trace from which a jumper could be soldered from the appropriate point to an adjacent straight track for tweak-on-test matching of delay. All that stuff largely disappeared for a long time but is now very common on consumer computer boards. There are versions of the old RCA CD4046 complete PLL in newer CMOS families that will run at nearly double the frequency of this LS part.
@@d614gakadoug9 I believe you're talking about the 10116. It was three fully differential sections, typically used as the front end of a frequency counter. The first two sections were used as amplifiers, while the third was used as a Schmitt trigger to square up the waveform.
@@hardlygamaliel455 Yup, that's the one. I couldn't recall the part number. DigiKey has quite a large stock in an SM package. I remember when DigiKey's catalog was about 8 pages.
@@d614gakadoug9 I wasn't aware of Digikey quite that early, but my first catalog was certainly less than 50 pages. Then it grew to phone book (remember those) size before finally going online. I also remember ordering from Jameco (back when it was known as James Electronics) and it advertised in the back of Popular Electronics - along with PolyPaks.
Glad I stumbled across this video cause I bought a tube of these for experimentation with RF circuits and totally forgot about them, thought they may work with a dac output to control frequency.
Would be quite handy with the I2C controlled variable potentiometer from a few weeks ago, in that you could have an I2C controlled external oscillator that requires no actual programming and barely takes any programming on the side of your MCU. You could then cascade them and have a lot of external oscillators being controlled by a single MCU and using only an I2C output. Far more than you could have with digital outputs and far more accurate.
It is an interesting device for sure. I am a little bit concerned about the output waveform. Compared to a TLC555 especially the high output voltage is not very stable. I am curious if it is the same for lower frequencies around 200 to 500 kHz. Maybe this is caused by an underdimensioned capacitor between Vcc and GND. The old TTL devices from the 1970ies have no shoot-through protected output stages (this is the same for the NE555, which has up to 200 mA shoot-through current, that's why I don't like it). Maybe a low ESR ceramic capacitor of 1 to 10 µF could solve the problem. For digital applications this is not an issue, but if you need well defined output voltages this should not happen. At higher frequencies such devices could therefore consume of lot of current and generate considerable amounts of heat.
I remember skimming across a board for numbers years ago and seeing a 324 but quickly realizing it was a 74LS and not an LM. I've never come across an LS625, but I have seen the LS624 a couple times in old gear and Mouser has fresh LS624 chips from TI but yikes, they are indeed pricey.
I wonder if the capacitor "outputs" put out half-of-a-ramp-wave. Not something you're keen to try? .-) There are also versions with two voltage inputs, one for frequency and one for "range", so should have a huge ... range.
Thanks for the great video and sharing another cool part. I also spied that little green PCB at the end of your solderless breadboard. I have searched high and low for one of those but am striking out. Do you have a source?
There are versions of the 4046 in other CMOS families that are much faster. One of the conveniences of the 4046 is the on-chip phase comparators. Iirc, one of the faster versions has three different phase comparators on-chip, but it has been a least a decade since I've looked at these things.
@@d614gakadoug9I used 4046 over 35 years ago to built a frequency synthesizer generator. I used a bunch of down counters and bcd switches. It worked amazingly well up to 2MHz 😅
Normal TTL voltages will be required. The reason for the separate pins is to ease issues with delivering clean supply voltages to the oscillator section. Noise on the oscillator supply will lead to jitter in the output frequency which can be a significant problem in lots of applications. DO NOT expect a circuit like this to perform well on a solderless breadboard with sloppy power connections and some decoupling caps with long leads crudely jammed in.
Yes, but a crystal oscillator will be stable. The stability and linearity of this thing will be poor, but that's not a problem when it is used inside a phase-locked loop.
@@IMSAIGuy There are voltage controlled crystal oscillators. Usually the frequency can't be changed by more than a few tens of PPM. VCXOs are very useful for making something like a very high precision timebase using the VCXO in a PLL locked to a clock signal from a GPS receiver.
I mentioned 50/50 because when (last century) computer overclocking was all the fad, people "just bought a new higher frequency oscillator" and cpu could not be overclocked by a 60/40 freq. That's when the RF oscillator entered the scene, and it was found that an adjustable 50/50 duty cycle clock could indeed benchmark better than the fixed freq quantums that 4 pin clocks were easily available. One of my claims to fame is designing and selling an Amiga accelerator-board --> FOR accelerator boards called SocketRocket. It interposed the MC686881/2 FPU chips clock with a TTL VCO, and a simple pot to adjust the overclocking above 50Mhz. Naturally I included a FPU chip select LED so you'd know when FPU was engaged, an used the power LED as reference temperature sensor so the clock-freq was regulated down w/temp. I have some extra blank PCB's and can send you pics of the finished "SocketRocket". P.S. MC68882RC50's will run at 62+ MHZ on a 50/50 clock. #Raytracing in the 80's
Capacitors are generally horrible when it comes to temperature stability. The "best" that are readily available these days at a reasonable price and suitable for a circuit like this are "COG" (C zero G) ceramic. These used to be called "NPO" said as en pee oh but in reality it is N P zero, meaning negative-positive zero - the temperature coefficient of capacitance is nominally zero. You used to be able to get caps like N750 - a negative tempco of capacitance of 750 ppm/°C. That's just an example. Various tempcos were available. There were "P: types too, hence the NP0 designation for those with neither negative nor positive tempco. By selecting the appropriate tempco you could do things like compensating for the opposite tempco of an inductor used in a resonant circuit as in an RF receiver. Polypropylene capacitor are reasonably good with regard to tempco of capacitance. Polyester ("mylar") are pretty awful. An oscillator like this would typically be used inside a "phase locked loop." In a PLL the stability of the capacitor isn't a huge concern because of the closed-loop nature of the system. Like the poor temperature performance of this part itself, the error gets "cancelled" to a very large extent. We now have ceramic caps with abominably high negative VOLTAGE coefficient of capacitance. A cap actually operated at its rated voltage may have 20% of the capacitance it would at a much lower voltage. Selecting the right cap for any particular application can be quite a challenge at times.
