There was a time when internal protection was poorly documented, but it is well specified for a very large number of modern components. Specifications are usually either 5 mA or 10 mA though I have seen some at 2 mA. I have seen a many ap notes where manufactures make it very clear that relying on protection diodes is perfectly acceptable it you limit the current. On at least a couple of occasions I have discussed it with ap engineers from major semi manufacturers.

The switch is only saturated when it is saturated. It passes through a linear region, regardless of what you do with it, and creates vulnerability. You can dress it up with external components, but it is still silly ahead of something with NO linear region at its input. Clamp the voltage if you are paranoid, but get rid of the worse-than-useless transistor.

 

Looking back at the original circuit, it seems that part of the problem may be from the input to the schmidt trigger IC exceeding it's supply voltage.

How would that happen? The input has a pull up to VDD, and gets pulled to ground by Q1. I don't see what would drive it outside of that range.

 

Somehow the 60hz line frequency changed.

The 4040 is a ripple counter, which means that none of the outputs change state at *exactly* the same time. This means that there are small "metastable" glitches on your decoded reset signal, and these glitches change with the value of Vdd. Add a 0.1 uF capacitor (1 ms time constant) from the reset pin / 10K resistor to GND. This should filter the microsecond-long glitches but let the real decoded signal go through.

ak

 

I don't think the generators at the power plant are changing in speed.

I have had circuits like yours oscillate during the slow transition time of the transistor. The best I can think to try is putting a small capacitor from the collector of the transistor to ground. I would try something like 100 pf.

Let me know if that helps since I don't have time to try it myself.

You're beautiful, the capacitor from collector to ground worked. Not 100 pf but .002 uf. I am now able to raise VCC to 15V.

 

I don't think the generators at the power plant are changing in speed.

I have had circuits like yours oscillate during the slow transition time of the transistor. The best I can think to try is putting a small capacitor from the collector of the transistor to ground. I would try something like 100 pf.

Let me know if that helps since I don't have time to try it myself.

You're beautiful Richard. A capacitor from collector to ground worked for a good one Hertz output. Not a 100 PF but I used .002 microfarads. I am now able to apply 15 volts of VCC with no problem.

If it wasn't mentioned clearly I'd like you all to know that my objective is to create the one Hertz output from a 60 hertz line frequency. This has been accomplished. Thank you.

 

Is it safe to use a resistor divider to output approximately 5v AC from the 120v AC line voltage without a step down transformer to clock my CMOS counter?

Without a transformer probably violates the terms of this forum.

Use the rough DC from a secondary fed rectifier and use the B/E junction for zero crossing detection - its the most accurate way.

you feed the reservoir cap via another single rectifier so the cap doesn't smooth the pulses you want to sample.

A bigger resistor can do higher voltage - but don't tell the moderators I said so.

 

Without a transformer probably violates the terms of this forum.

.

You get isolation in the Fairchild Ap note H11L1 is 7500pk isolation voltage rated.
Max.

 

You get isolation in the Fairchild Ap note H11L1 is 7500pk isolation voltage rated.
Max.

Max is right, have a look.

 

How would that happen? The input has a pull up to VDD, and gets pulled to ground by Q1. I don't see what would drive it outside of that range.

Looking at the original circuit there is nothing to prevent the input from being driven higher than whatever the supplied Vdd is. So even with input protection diodes, as long as the Vdd source impedance is greater than zero, an input could be driven to a higher voltage, with totally unspecified results. That was my point.When Vin>Vdd, performance is not defined. I am not saying that was the causeof the problem, but rather that it could possibly be the cause. Few of us know what is happening in undefined areas. At least I don't know.

 

Looking at the original circuit there is nothing to prevent the input from being driven higher than whatever the supplied Vdd is. So even with input protection diodes, as long as the Vdd source impedance is greater than zero, an input could be driven to a higher voltage, with totally unspecified results. That was my point.When Vin>Vdd, performance is not defined. I am not saying that was the causeof the problem, but rather that it could possibly be the cause. Few of us know what is happening in undefined areas. At least I don't know.

Are we talking about the circuit in post 5, or is there another one that I'm missing? Assuming it's post 5, how would over voltage get to the input? Are you worried about voltage going backwards through the base-collector junction?

Hmph. Wait a minute. I was writing that last paragraph thinking the BC path was silly... but maybe I just forgot how transistors behave. It's slowly coming back to me. BC junction is sort of like a diode, so if base is higher voltage than collector, current will flow. Am I getting that right now? If so, I understand your earlier comments now.

Sorry for my confusion.

 

Maybe add a zener diode and a fast-blow fuse to get reverse polarity protection and overvoltage protection. With big transformers and high inductances, you can get some big voltage spikes. And you may wire things up wrong or have other issues. In normal operation, you only have the resistance of the fuse. If there is overvoltage, the zener clamps it and probably blows the fuse to let you know. But either way, the voltage does not exceed the zener voltage. If there is reverse polarity, the zener clamps it to it's fV when forward biased. It also blows the fuse. Just a suggestion.

 

Are we talking about the circuit in post 5, or is there another one that I'm missing? Assuming it's post 5, how would over voltage get to the input? Are you worried about voltage going backwards through the base-collector junction?

Hmph. Wait a minute. I was writing that last paragraph thinking the BC path was silly... but maybe I just forgot how transistors behave. It's slowly coming back to me. BC junction is sort of like a diode, so if base is higher voltage than collector, current will flow. Am I getting that right now? If so, I understand your earlier comments now.

Sorry for my confusion.

I am talking about a circuit that is probably the same as in post #5, but looks different on this computer. A transverse mode spike will transform quite well, and that is how it would get to the logic gate. Transistors, also, are not as nice as we would like when operated outside of where they are intended to operate. The problems simply that when you venture far outside the specified operating conditions one is in unspecified territory. AND, when the base voltage becomes higher than the collector voltage the results are not always what might be predicted.

 

I've learned that the Schmitt triggers are definitely necessary. The two circuits work well to create 38.768 khz oscillators. Schmitt triggers are necessary for the one using 2 transistors because it generated more than a pulse per second. I went from 6 volts to 15 volt and things are okay. I think that 5v should be OK also. The collector of Q1 joins R1 and R3 and the base of Q2. C4 is 10pf and C5 is 47pf.

 

Indeed, schmitt triggers are very useful and many times really vital.But they are not very adjustable. Some other logic devices also have schmitt trigger inputs that may not work quite as well as we would hope. And it seems that different brands of the same device have different characteristics. I once had to change an IC brand dual 4538 of a whole production run because of that. Fortunately no product had shipped yet.

 

Seems strange that there is 34+ posts for something that has been done countless times before and examples available in this thread and by Google.
Both isolated and none-isolated.
Why try and re-invent the wheel?
Max.

 

Seems strange that there is 34+ posts for something that has been done countless times before and examples available in this thread and by Google.
Both isolated and none-isolated.
Why try and re-invent the wheel?
Max.

The mathematics is false. The physical constants must become variable but those in control of fate want to limit technological possibilities by keeping ratios the same. The count of 1 is acting like 0 or 2 so definitely we must reinvent to verify the doubt of ourselves. We have to understand because this is the nature of human beings-to question ourselves.

 

The mathematics is false. The physical constants must become variable but those in control of fate want to limit technological possibilities by keeping ratios the same. The count of 1 is acting like 0 or 2 so definitely we must reinvent to verify the doubt of ourselves. We have to understand because this is the nature of human beings-to question ourselves.

In addition to the problem being a very interesting one, Max, creating is FUN!! At least I find it fun. While learning from the past is always well advised, this thread started out with a very novel complaint, that as the supply voltage increased suddenly operation changed. While that may be the result of things obvious to some, it was not obvious to a whole lot of folks. That made it interesting, at least for me.
Now it is time to consider those no longer with us, the original intent of Memorial Day.