yes once it triggers I want it to stop oscillating until it is triggered again and then I want a two or three day delay again. But I do have a question. If we want it to count over a 3 day period we will want a very low frequency right? and why do we want 10800 as the binary number?

 

Because it is a counter, and when the count reaches that binary number exactly 24 hours has passed. At that point the counter resets and starts counting again (or in this case, not). The reset pulse is the 24 hour indicator.

The 4017 will be a day counter, you select the two or three output and you have the days down. Replace it with a rotary switch and you can have any number of days you want up to 9.

 

I assume a fixed location as there is a 5V & a 24V supply, operation triggered by a "sensor" closing a switch [?momentary?]. If so, operation is a one shot, with every thing reset at last bell. With 60 Hz: 4093 for shaping & memory, 4040 /512, 4040 /3395, 4017 or 4022 /3 [days], 4017 or 4022 / 2 or 3 for 48 or 72 hr sw selected, 555 or decode existing counters for one hr,[ /512 & /422 for one hr.] + misc invertors & NAND getes.

There is a charm of having a independent oscillator that will keep on chugging if the line power dies. This assumes a battery backup, which isn't hard.

 

So then how do I connect up 3 4060b's to count out a 72 hour period? do I run the output from one to the restart on the next? and how does a 4017 count the days?

 

Qty 3 4060 using a 32Khz crystal and a 5 input NAND gate will count 24 hours, I'm not going to use the extra counter flip flops. That marks a precise 24 Hour period. The 4017 marks the days.

I haven't worked on the 555 section yet.

I've reposted the 4060 time base timer. I wish they had added one extra flip flop to create 1 Hz, but I don't know if they thought of this application when they release the chip, but it works well.

******************************

OK, here is a 24 hour timer. It is still pretty simple, no frills like reset, and the inverter gives a very narrow negative pulse every 24 hours. The negative pulse is important for several reasons, for now it is just to reset the primary counter.

 

On my "quick " post, my token expired, so here's extra short form: Asuming a one second clock. 24 hours is 86400 sec, dividing by two binary numbers, such as 4096 twice gives a new period of .oo51498 or about 194 Hz. So two 4060s a 4017 & 555 + misc. gates should do. If intrested in this approach let us know, and " we'll" add some pin nos.

 

So is the resistor next to the inverter the 24 hour output? Then how do I trigger it? do I put a switch on the voltage supply and when it has voltage it will start counting like a monostable 555 timer?

 

The diode/resistor is a diode AND gate. Like I said, I'm designing this from the ground up, and I haven't started the 555 section yet. Next comes the day counter.

 

Bill- is the IC connected to the diode AND gate an invertor or non-inverting buffer?

 

The symbol is that of a inverter. The circle on the tip of the triangle defines inverting.

 

Call me confused: If the last 4060 is reset, all outputs are low, input to the " invertor" is low, out is high, 4060 held in reset inhibiting colck, counter cannot count??

 

OK, I think I had a brain fart about the direction of the pulse, you want the reset pin low. It will go high when the number representing 24 hours is reached.

Thanks for the catch.

 

OK, latest schematic.

I don't see a way to get rid of the 3 input OR gate. I'll keep peculating on it.

I see a problem with a 555 1 hour (?) monostable looming ahead. I'll bring it up when I get there.

After I finish I will redraw a schematic showing a RC oscillator instead of U1 and U2.

Quick question, why 5V and 24V? Why not 12V and 24V? While I can do 5V, it would be easier with a 12V power supply on this circuit. Ultimately the transistor I plan on using to drive the solenoid is easier to get in the 12V version (that being a MOSFET).

 

I was just thinking 5 volt because that is what it takes to bias the npn transistor that I was planning out. I do not have alot of design experience yet though im only a tech. So if we could bias a mosfet with 12 volts and have it work the same as a transistor that would be great. I have drawn it up if you guys could check over the transistor/mosfet circuit which ever one is ok. Just one more question what is the difference between a depletion and a enhancement mosfet.

 

You will find MOSFETs much better parts for this sort of thing. They conduct much cleaner (less ohmage), and take almost no current to switch, and run much cooler because of it. You can even buy one from Radio Shack (not the best mind, but it is up to the job). The only down side is they are static sensitive, which is to say while building with them you need to pay attention to ESD. I suspect once you've gotten used to them you won't use regular BJT transistors again. Overall the cost is comparable too.

The problem I was alluding to with the 555 is the duration is pretty long. I've never had the problem people describe myself, but large capacitors tend to leak more than smaller ones, and as a basis of reference we are talking a 1MΩ and 3300µF capacitor. Old capacitors have this problem more than new ones, and if it is exercised I don't believe it will fail.

My suggestion is try the design with the 555, and if it doesn't work I'll come up with another 4060 version. Like I said, I haven't had the problem, but if I post it expect a lot of protests to follow. The other problem is capacitors in this range tend towards ±20%, add the normal ±5% of the resistors and you will need a variable resistor to tweak it into the exact time.

My experience tends towards enhancement type MOSFETs, so I can't really answer that question with authority. The few times I've used them (I'm an old fart, so I used regular (BJT) transistors most of my life) I've been really happy with the results. Just shopping at my local electronics outlet they have some incredible specs (50 amps or more anyone?), more than the case (usually TO220) will allow.

Most MOSFETs require 10V to turn on reliably, and require a resistor on the gate to prevent ringing. They have a substantial gate capacitance, combine that with the lead wire and you get a LC circuit than will make them switch badly, but it is all fixed with a 100Ω to 1KΩ resistor on the gate. There is a sub family called logic level MOSFETs, that can turn on completely from 3.8V to 5V. Because of the gate capacitance they surge briefly when switching, but pull almost no current once the capacitance is charged (something less than nanoamps).

I'll bet back on the circuit when I get a chance. I may see if I can simplify the logic gates to nothing but diodes, but I'll need to run some experiments on my end to see if I can get by with it.

I will also be showing some simplified RC oscillator types to allow you to play with it on a protoboard.

 

Todays efforts.

 

Here is one version using 4060 oscillator tuned to 194.18+ Hz. Dividing twice by 4096 gives an output at 24 hrs then divided by selectable switch for 2 or 3 days. Number of connections are reduced with only two added ICs, 4011, & 4071. Every thing starts at reset; operation starts by momentarily closing SW, setting S-R latch U1&2, releasing resets from U13,U6 & U7[ clock enable]. Do not know solenoid drain- just use Bill's 555 ckt. A review would not hurt.

 

bill, is that inverter in your circuit a 555 timer made into a voltage inverter. it is the inverter with the 555 in the middle of it.

 

At this point I'm slinging gates around as needed. When it jells a bit I'll define them a bit better. I've about decided to get rid of the 555 on the left, and the middle inverter is just that, a simple inverter. You have to understand these are my working sketches, I don't work on paper anymore.

Wait till you see my diode gate jell. I will definately have to experiment with them to see if I can use them as is or add some buffers to clean up the logic. I think I can get them to work as is.

The problem with diode gates is they are analog in some ways. I generally try not to stack them two deep as I have done here.

The 555 invertor on the left is a bastard flip flop. If you hold the voltage on a schmitt trigger mid way it will stay at the last state it was in.

I'm thinking of using a simple toggle switch instead. If you like the idea of push buttons better we can keep it. There will be some simplification as I do this.

The big difference between Bernard's designs and mine is the crystal oscillator. If a RC oscillator is used the frequency can be picked to simplify the basic circuit, as his has done. I've got to tell you this crystal oscillator is harder to build, but it is also much more stable over time (and 72 hours is a lot of time).

 

so what is the difference between an n channel and a p channel mosfet I know with regular transistors pnp and npn is the direction of current flow.