Useful "Delay" Circuit

[DigitalGriffin]

I can successfully power up 8 sound engines at once with 3 amps.  Any more then that and things get "touchy" as surge current overloads a lot of throttles and circuit breakers.  Once the capacitors stablize on the decoders, you can run 20 or 30 engines at a time.

This can be a real problem after a short where you have 20 or more trains.  A good example of this is a diesel house and roundhouse.

A number of people try to solve this problem by adding power switches to groups of stalls, then powering them up one by one.  This is effective and cheap, but isn't the easiest as you have to remember to power down your engine house after every short circuit.

So...without much ado, here is a circuit which will delay power over x seconds using a capacitor, transistor, resistors and a relay.  (I'll explain how to set it up for whatever seconds you wish)  You can chain these in parallel, so that stalls automatically power up in sequence.  IE: stalls 1->5 immediately.  STalls 6->10 after 5 second.  Stalls 11->15 after 10 seconds.  Stalls 16->20 after 15 seconds.  This gives the capacitors and surge current time to stablize.





Using the R/C circuit time formula we have:

.7 = 1 - e(^{-t /R1 * .00022)}

t = time in seconds of the delay we want.  .00022 = 220uF of the cap.

We plug this into the formula to solve for R1 in the above circuit.  

 

^{For example:} 

.7  = 1 - e^{-22727/R1     (5 seconds / .00022) = 22727}

-.3 = -e^-22727/R1

ln .3 = -22727/R1

-22727/(ln .3) = R1

R1 = 18.875kOhms = 5 second power up delay

 

[hon30critter]

Don H:

Interesting solution. Do you happen to have a part # list from somebody like Digi-Key?

Also, because my eyes glaze over at the sight of electronics formulas, would it be possible to provide the resistance values needed for the longer delays? To be honest, I don't have a clue what the formulas you posted mean.

Thank you

Dave

[rrinker]

 Forget that is has anything to do with electronics and just solve the equation. Because logarithms are involved, it's now longer basic algebra like Ohm's Law for calculating current limiting resistors for LEDs, but the concept is the same. That's all Don did in his example for a 5 second delay. Instead of the 5, plug in 10 in his formula if you want a 10 second delay. If you use Windows - in the Calc app, under View, select Scientific and you get access to the ln and e (inv ln) functions. First two buttons on the top left (under the degree/radian/grad selection)

 The confusing part for non-electronic people is that the result is a rather exact resistor value - and resistors don't come in detailed values like that. Basic example - for lighting we typically use 10% tolerance resistors. That means the actual value can be + or - 10% of the stated value. Standard values are thus approximately 20% apart - it wouldn;t make sense to have a 100 ohm and a 110 ohm available, because the 100 ohm could be as much as 110 ohms and the 110 ohm could be as little as 100 ohms and still be in tolerance. Here's a trick for understanding some of this stuff - in this case, it's the same concept as tolerance in machining. If you understand that - you understand this. Just think of the units as somethign you are comfortable with, if the idea of ohms or kilohms scares you.

ANd realize, this is not a high precision circuit. With some very overkill expensive parts, you could make it fairly accurate, but with standard common parts that won't break the bank, you will probably have a second or two variation  between two of these built with the same value parts. Ultra precision is not required for this task, and these circuits are about as simple as it gets. You can do similar (and control multiple relays) with one Aurdino or similar. Such a circuit woudl look a lot simpler, since most of it is the 'black box' of the Aurdino chip. Opened up though, it is a million times more complex than this circuit. It just transfers the effort from circuit design to program design.

                  --Randy

[hon30critter]

Randy!

I'm laughing at me, not you!

I haven't done math like that for 45 years, and I can hardly remember what I had for lunch, let alone what I learned back then. Besides, I wasn't any good at it back then either.

Somehow I missed the boat. My dad was a scientist and my son loves math. Heck, Cole does math for entertainment! Maybe it's like baldness - every second generation. I got on the wrong end of that stick too.

Anyhow, for the sake of us dummies, could somebody maybe just post the resistor values, and maybe even a part # list too? If you give me the pieces I can put it together for sure.

Thanks

Dave

 

[DigitalGriffin]

I'll do you one better.  I have to build a couple more because I have 36 stalls all filled.

Once the parts come in I'll take photos of the assembly on a breadboard and make a video of it working.  I'll also post a part list.

I'll show you then how to calculate various values.

Randy is right this is not a very precise circuit and exact resistor values are rarely available.  But it does get the job done.

[hon30critter]

Thanks Don!

I will eagerly await your posts!

Dave