Hi all,

Let me start off by saying thanks for this great resource and community - it has already proven invaluable for me in the short time I have been using it.

I have been playing around with some 555 circuits with the end goal of driving 12 IGBT switched "smart coils." What I am attempting to do is produce a TTL-esque square wave with 3-8mS high (5v) and 10mS low signal. This signal would be passed around 12 outputs so that each output would "wait" until all the other channels had fired.

Initially I had tried an Astable 555 circuit clocking 2 4017 counters. I was unable to figure how to delay the rising edge of the counter output and I also discovered that the 4017's couldn't sink enough current (I talked to the manufacturer and they couldn't give me a spec - I'm guessing 20mA) and were quickly damaged.

I did discover through experimenting that a standard Astable 555 circuit can drive the smart coil without issue so I set upon cascading 2 555's together (I don't have a schematic but it was along the lines of Bill Bowden's 555 delay below but with an Astable (with adjustable duty cycle) circuit driving the trigger of the second 555. The result was the 2nd rising edge immediately preceding the 1st clock's falling edge despite changing values of the RC filter linking the 2 555's.

I am sure there are better ways to do what I am trying to do although there are a few reasons why a simple 555 circuit is appealing. I tried using an Arduino but again it couldn't sink enough current without added transistors. I am very green in regards to designing/modifying circuits and RC filters in general although I have a fair amount of troubleshooting experience and the tools to do so. Can someone point me in a direction? Can I just delay the trigger of the next 555 somehow?

 

What's wrong with adding transistors to drive the needed current?
Not using transistors is an odd limitation.

 

So you have a square wave pulsing at 3-8 msec high, 10 low, and you want to distribute it to 12 transistors, one at a time,

like a car distributor?

 

What's wrong with adding transistors to drive the needed current?
Not using transistors is an odd limitation.

Right?! It was born of using parts I had on hand and trying to figure out what I thought 'has to be simple.' It very well might be the best way to get it done. Thanks for pushing me in a direction!

 

So you have a square wave pulsing at 3-8 msec high, 10 low, and you want to distribute it to 12 transistors, one at a time, like a car distributor?

Exactly like a distributor - if you were to look at all 12 signals on a scope I would want the 10 mS low between each rising edge. In theory this would be adjustable to let my 12VDC supply for the coils recover (not that it seems to need to).

 

So you need to create a 12way switch, clocking it on the high pulse?

 

So you need to create a 12way switch, clocking it on the high pulse?

Yes that sounds like the thing.

 

Could you use a 12-bit shift register with your 555 clock pulse? Use the logic to shift one "1" state all along the register till reset or repeat. Lot of different ways to do it with logic but it depends on what you have to hand I guess.

 

Could you use a 12-bit shift register with your 555 clock pulse? Use the logic to shift one "1" state all along the register till reset or repeat. Lot of different ways to do it with logic but it depends on what you have to hand I guess.

That's what I did originally with the 4017 chips (which I think are shift registers, no?). I ended up having trouble figuring out how to delay the rising edges so they weren't all right after each other (resulting in one of the 12 signals being on at any given time).

 

Yeah its a decade counter so it is a type of register but I thought you want 12 outputs. I'm not sure what you want with outputs sorry. Do you need continuous astable inputs to each transistor which are just staggered initially? or do you want just one transistor on at a time, 1 through 12 and then all off?

 

Try a HEF4894 12bit shift register.

 

Yeah its a decade counter so it is a type of register but I thought you want 12 outputs. I'm not sure what you want with outputs sorry. Do you need continuous astable inputs to each transistor which are just staggered initially? or do you want just one transistor on at a time, 1 through 12 and then all off?

As I was experimenting I was using 18 outputs based off of a LED chase circuit on Bill Bowden's site. The end goal is 12 outputs/steps staggered so there is 10mS between one output turning off and the next turning on.

 

Try a HEF4894 12bit shift register.

Give this a look as dave suggested you may need to cascade two of them though if your duty cycle has to be less than 50% I'm not too sure.

 

Give this a look as dave suggested you may need to cascade two of them though if your duty cycle has to be less than 50% I'm not too sure.

I do need less than 50% duty cycle. So if I were to tie the clock and data inputs together and feed it the correct signal (3.5mS on 10mS off) I should get the same on the shift register output, no? I don't understand the needing 2 to get less than 50% duty cycle. Looks like the output can handle 40mA so it's more robust than the 4017 I was using.

 

Hmmm - feeling a bit foolish at this point but I'm having trouble getting the HEF4894 hooked up correctly. Does anybody have mind lending me a hand? So far I am trying:

555 output to 4894 Clock Pulse and Data (tied together)
Output enable and Strobe tied to Vdd=5 VDC
Qp(n) tied to Vdd through a ~250 ohm resistor

Result is no signal on the Qp(n) outputs

Strangely when I lift the jumper wire tying the clock and data together so it is just connected to the data pin (other end floating) I get a pulse on the Qp(n) outputs. If connect data to Vdd my outputs are low and if I connect data to Vss my outputs are high.

I have a feeling I am not properly understanding how to implement a shift register despite reading about them a number of times...

 

Some progress:

I tied Data to Vdd and switched the OE by tying it to the clock signal. That gets the inverse of my 555 clock signal - yay!

Now I'm just not understanding how to cycle data to each input successively.

 

So I thought I had a bright idea to use a separate clock for the D and CP to the register. My theory was to put a data signal in which would be what I want at Qp(n) inverted and a clock signal that would give me the off I want before the next stage fires. Didn't work... Can someone explain this to me? I am thinking of it like the clock signal is a snapshot of what data has in its buffer on clock's rising edge. Am I completely wrong here?

If I can find an appropriate shift register macro I'll post a snapshot of the schematic.

 

Here's about all I can muster for a schematic. For some reason the 'one-shot' isn't working in LTSpice but does work on a bread board. For some reason lowering the discharge 'shunt' R9 doesn't seem to have much effect on the discharge time of C2 but I don't fully understand the trigger circuit.

I have yet to get the HEF4894 hooked up to the clock and one-shot but I hope to tomorrow. This has been a long long road... But lots learned! Any comments - even vile derogatory digs are much appreciated.

 

Here's about all I can muster for a schematic. For some reason the 'one-shot' isn't working in LTSpice but does work on a bread board. For some reason lowering the discharge 'shunt' R9 doesn't seem to have much effect on the discharge time of C2 but I don't fully understand the trigger circuit.

I have yet to get the HEF4894 hooked up to the clock and one-shot but I hope to tomorrow. This has been a long long road... But lots learned! Any comments - even vile derogatory digs are much appreciated.

Look at U2 closely, especially around the trigger pin, then tell me what is supposed to trigger that circuit.

 

Look at U2 closely, especially around the trigger pin, then tell me what is supposed to trigger that circuit.

I understand that the trigger pin needs to be low for the 555 to output a high signal. I also understand that initially the trigger pin sees 5.5v through R8 until C2 charges drawing the trigger to ground. What I really can't figure out is what R3 does. It doesn't work without it. LTSpice doesn't seem to care if it's there or not when I model just the trigger circuit as seen in this attachment. Thanks for the input GopherT.