Help Debug Circuit

[AnalogKid]

Good news!So being that the 3906 & the 327 have an emitter voltage of 5

Not exactly correct. The breakdown voltage is the minimum value specified *not* to cause reverse breakdown. For the 2N4401 it is 6.0 V. this means that at 6.0 V, the junction will not conduct (except for some very small leakage current), behaving like any normal reverse-biased diode. But at anything from 6.000001 V to 99 V, the diode will break down. I made up the 99 v, but you get the idea. I don't think I've ever seen the max breakdown voltage specified in a datasheet.

My guess is that the actual breakdown voltages for the two transistors are different enough to affect the circuit.

I think you've rediscovered why using a reverse-biased transistor as a zener diode is a very uncommon practice. It is an interesting characteristic, and can be used to make the world's most simple oscillator, but it is so unpredictable that another thing I've never seen (except for Dick Cappels oscillator) is a design paper / app note / whatever proposing its use. In the classic two-transistor multivibrator circuit, it can have a real effect on the output frequency when the reverse biased transistors clip the peaks of the timing capacitor voltages.

For a one-off home project, with a lot of adjustment range, by someone with enough design experience to handle the consequences, it's fine. But you can see (37 posts later) what happens when the idea gets outside of that safety zone. I'm working on a control system for my garage doors, based on using the parts I have at home and it being used ***only*** by me. Nothing in the design is stupid, but I would not take some of the approaches in a commercial design.

ak

 

[ThomsCircuit]

But you can see (37 posts later) what happens when the idea gets outside of that safety zone

I am so grateful for all the time you dedicated to my project. I have learned a bit more about this circuit and even more about transistors. Yes the start delay range has changed (longer 3-9 as apposed to 1-5 seconds) but as you said these are for my shop and i can handle the changes. My initial concern was that i had wired something wrong or mis interpreted the schematic. But now i know that i got it right and understand just what was wrong. The design. And to top it off its working too. Thank you again ak.

 

[AnalogKid]

No problem. There are many ways to do R-C timers, and your staged delays application is not a new idea. Puttering around, it could be done with 2 transistors (without all that reverse-breakdown stuff), a few transistors in a transistor array chip, or any of several CMOS logic chips pressed into timer duty. At its heart, it is two R-C delay circuits in series. But in fact, it could be done with just one transistor and two diodes handling both the turn-on and turn-off delays. Put two designers in a room, and you'll get 5 answers.

I'm sure it's burried in the threads somewhere, but - in a perfect world, what are your desired time ranges for the turn-on and turn-off delays?

ak

 

[rjenkinsgb]

the start delay range has changed (longer 3-9 as apposed to 1-5 seconds

You can compensate for that by halving the value of C1, that would make it roughly about 1.5 - 4.5 seconds.

 

[AnalogKid]

Unfortunately, a common response on forums is to tell the thread starter that everything about their circuit is wrong, you're stupid to do it that way, it should be done this (my favorite) way, use my favorite part, etc. I am not a fan of these responses, and try to stick to the original question. But, now that that is over . . . . . . .

There are other ways to do what you want to do. I won't rant about better or worse, just other. Frankly, I'm surprised someone hasn't jumped in to tell us that the "right" way is to use a microcontroller. And there is something to that. Essentially, you have one input pin, one output pin, and some counting/timing stuff in the middle; a trivial task for a uC. A 6-pin or 8-pin PIC / Atmel / Freescale / whatever costs less than $1, if you don't count the added voltage regulator, software development system, programming device or cable, and the firmware development time (write, debug, debug, debug ...). But I digress.

As I said in another post, there are a lot of ways to do this task. Here is one. This is as minimal as it gets, just four components. This version assumes that the off-delay always is longer than the on-delay. As shown, both the delayed-on and delayed-off times are fixed. Both can be made adjustable by replacing R1 and R2 with combinations of one fixed and one variable resistor, as in your present circuit. This circuit differs in that the two adjustments are not independent. The on-delay is controlled by R1, but the off-delay is controlled by R1+R2. The two indicator LEDs can be added back into this circuit; they are left out for clarity.

Note: the component values shown are very rough estimates; I just winged the arithmetic in my head. If you want to play with this circuit, I can do a more rigorous job on the numbers.

ak

 

[AnalogKid]

Watch this space for alternate circuit #2. This one in fact does use one of my favorite parts, in a way that could be called one of my personal design quirks. We all have a little Ken in us.

ak

 

[ThomsCircuit]

in a perfect world, what are your desired time ranges for the turn-on and turn-off delays?

This is so cool. I dont get out much and this is like a breath of fresh air. Thank you.
Turn on delay 1-5 seconds
Turn off delay 5-10 seconds
12 volt power supply
The trigger must be a latch. (close the circuit) because it is driven by this circuit. (A Peak Detector) A circuit we made a few weeks back. its a huge 10 page thread but that's my style. What I like about this forum is the dedication of the members to help anyone regardless of experience. They stuck it out with me until the goal was reached. The detector simply latches a switch when i pull the trigger or turn on the tool connected (plugged into) to it. It replaces the Hawkeye 800 which is powering two others in my wood shop. And while the Hawkeye 800 needed no external power source the one we developed did so it is plugged into the Delay Circuit as shown here.

Updated Delay Control (now supplies power to Peak Detector)

Peak Detector

 

[ThomsCircuit]

As I said in another post, there are a lot of ways to do this task. Here is one.

I know in a previous post I said I need extreme guidance but after 2 years of being in this group I cannot say "I've learned nothing" The best way to understand something is to start with as few parts as possible.

 

[ThomsCircuit]

Watch this space for alternate circuit #2

Totally on board. No pun intended.

 

[ThomsCircuit]

Peak Detector

My leds are mislabeled in post #47 (open / closed) should be reversed. Oh wait. I forgot i corrected them. NVM

 

[AnalogKid]

Here is another alternate circuit. This uses one of my favorite parts, the ULN2003/2004. This is a transistor array chip, with seven darlington transistors. It is intended as a driver for small relays, lights, LED digits, etc. In fact is has a zillion uses. I've often used it as an industrial-strength logic device. It behaves as a sept-inverter, a hex inverter with one extra gate. Each transistor is rated for 50 V, can sink 1/2 amp, and has inductive kick spike suppression built-in. It is a great part for harsh electrical environments, such as automotive. You would not believe what the power system in an ambulance looks like. If I were doing this for myself, this is the circuit I would use. But part of that is because this is an old friend. And I have about 50 of them.

In this circuit I use four of the sections as open-collector inverters. You can see the internal schematic for each stage in the insert below. In very round numbers, each input has a transition level (the voltage at which the output changes state) of around 5 V, and appears sorta-kinda as a 20 K resistor to GND. Because this relatively low input impedance is connected to the timing capacitors, the cap sizes are increased to get the same kinds of time delays as the original circuit. Again, the parts values shown are coarse estimates.

With no input, the U1A input is pulled high by the very low current through the LED, probably not enough to produce visible light. This holds U1A's output low, keeping the capacitor voltage and the U1B input near 0 V. U1B's output is high (open), so R2 holds the voltage across C2 a 0 V (+12 V on both ends). This drives U1C's output low, which forces U1D's output open, and the SSR is off.

When the sensor contacts close, U1A,s output goes open and R1 starts to charge up C1. When the voltage on C1 crosses 5 V, U1B's output goes low. This yanks the input to U1C low, and charges up C2 very rapidly. With the U1C input low, its outut is high, the output of U1D is low, and the SSR comes on.

When the sensor contacts open, its input is pulled high enough to drive its output low. This discharges C1 rapidly and drives the U1B output open. This allows C2 to start discharging into R2, and the input to U1C begins to increase. When it (finally) crosses 5 V, the output goes low, making the U1D output go high (open circuit), turning off the SSR.

That sounds like a lot, but it is very similar to how the original circuit works - the action of one timer affects the action of another timer - the end of one timer initiates the beginning of another timer.

ak

 

[ThomsCircuit]

ULN2003/4

COOL! Ive got a few TBD62003. A quick search shows that they are similar. I could breadboard with the 62003 then order ULNs
I see the dif between the UL...03 and the 04 is the resistor value. 03 is safe with 5V and 04 is good for 15V.
The schematic is difficult for me to understand but ill place it in DesignSpark and see what happens when i let it make a pcb for me. That will give me an idea of how the other components are connected to the ULN.

the action of one timer affects the action of another timer - the end of one timer initiates the beginning of another timer.

Interesting...

 

[eTech]

Hello,

If you have a couple of 555's laying around, here's an optional DelayOn/DelayOff sequential timer circuit.
The first timer starts when the contact closes, expires after 5 seconds, then triggers the second timer that energizes the output relay for 10 seconds.
Both timers are adjustable. The BJT prevents the first timer from triggering again until the second timer expires. The output can drive an SSR if desired.

 

[ThomsCircuit]

My interpretation of AK's schematic.

 

[ThomsCircuit]

The first timer starts when the contact closes, expires after 5 seconds, then triggers the second timer that energizes the output relay for 10 seconds.

just so im understanding your description. 5 second (adjustable) delay energizing relay after contact closes then 10 second (adjustable) delay de-energizing relay after contact opens.

 

[eTech]

just so im understanding your description. 5 second (adjustable) delay energizing relay after contact closes then 10 second (adjustable) delay de-energizing relay after contact opens.

Yes.

 

[AnalogKid]

2003s will work in the circuit, but R1 and R2 get smaller, and C1 and C2 get larger.

ak

 

[AnalogKid]

My interpretation of AK's schematic.

Except I don't have a standard relay. I show the last stage driving the solid-state relay directly.

ak

 

[ThomsCircuit]

Except I don't have a standard relay

yes. i saw that. i need to learn to just let go.

 

[ThomsCircuit]

If I were doing this for myself, this is the circuit I would use.

Thank you for providing a simple but powerful example. Im going to get some ULN2004's The TBD's are not available at Tayda. But the others are. And only .34 each!
Ive read your explanation of how the project works but i dont know how or where to add the variable resistors so i can adjust the on/off. But based on the info you provided i think R1 for the ON delay and R2 would be for the OFF delay. Now from what ive seen in other schematics you need both a fixed resistor and a variable pot in line to provide the minimum and maximum range. I have tried to understand the formula to determine what combination is needed I have to confess my algebra is very poor. My kids however are extremely smart. I mean it too. My oldest is a scientist and my youngest is the youngest person in her Algebra II Honors class. Nevertheless could you tell me what combination i would need to add the delay function ranges?