4017 Counter skips under load

[[email protected]]

Ah, a picture tells a thousand words.

The board should work just fine *if* you power your load from an
entirely separate and *floating* power supply relative to the plugpack
or power supply which powers the board. You've said the board works
just fine with an LED load, so obviously something is amiss with the
power supply wiring for the load.

Dave.- Hide quoted text -

- Show quoted text -

I tried that earlier, but couldn't get the firing boards (running
4017) to recognize the clock pulses from the parallel port interface
unless the grounds were connected. Am I doing something wrong there?

 

[David L. Jones]

I tried that earlier, but couldn't get the firing boards (running
4017) to recognize the clock pulses from the parallel port interface
unless the grounds were connected. Am I doing something wrong there?

I don't get the whole picture here, so let's see if I have this
correct...
The 4017's are located on your own board at the end of a long ethernet
cable connected connected to this relay interface board?

Please explain in detail what the relay interface board does, what
your 4017 board does, how they are connected together, how they are
powered, and what lengths of cable are used. Unless you do that we'll
all be going around in circles.

Dave.

 

[rebel]

That is a parallel port interface kit that I bought some time ago and
use for any applications where computer control is needed. It is
completely different from the boards that use the 4017. I will post
pics in the morning when I have enough good like to get a detailed
picture.

Now I *am* confused. When you posted that pic, I expect we all took it to be
the board you were using.

As for the parallel board, the 2803 went out, so I replaced
it with an NTE 2018. I'm told that they are an identical match, both
8 channel darlington arrays. Am I wrong in assuming that if the on-
board relays are rated for 12 volts, as is the power supply, that the
output from the 2018 is also 12 volts?

That board uses 12V for relay coil supply: http://kitsrus.com/pdf/k74v2.pdf

The 2803 type drivers are open collector (and typically specified for 30 volts
or so).

So can you explain again where this Kits'r'us board fits into your scheme?

 

[Fred Bloggs]

Sorry, I mistyped. The clock pulse coming into the 4017 from the PC
parallel port is 25 ms in length. The counter then activates the next
relay in sequence and the output stays active for 10-20 seconds before
moving on to the next firing cue. I'll post a printout of the circuit
design when I get home in a couple hours. I'm using a pretty good
sized power supply, and I've tried other power sources as well with
the same results. Should I be doing some sort of decoupling where the
power enters the board as well? I noticed that the parallel port
interface card that I'm using has some sort of diode/cap combination
where the power enters the board. Would this help even things out
further?

The printer port puts out TTL compatible voltage levels and will not
reliably clock a 4017 powered from 12V. The 4017 will have problems
driving the typical relay coil directly. You should forget about
completing this project on time and ask your teacher for an extension.
You have a lot more homework and circuit modifications to perform before
you get close to finishing it.

 

[[email protected]]

I don't get the whole picture here, so let's see if I have this
correct...
The 4017's are located on your own board at the end of a long ethernet
cable connected connected to this relay interface board?

Please explain in detail what the relay interface board does, what
your 4017 board does, how they are connected together, how they are
powered, and what lengths of cable are used. Unless you do that we'll
all be going around in circles.

Dave.- Hide quoted text -

- Show quoted text -

Hopefully these pics will show the full set up.

http://a548.ac-images.myspacecdn.com/images01/105/l_9f67b69cc9c5093d7535e07ff020383b.jpg

This is the inside of the firing box. The barrier strip at the top
brings the +12V and ground for the firing control boards (at bottom)
as well as a separate ground for the igniters that is controlled by a
safety switch. I added the cap to the barrier strip last night in an
attempt to help stiffen the power supply to be boards, but have not
yet tested it.

http://a340.ac-images.myspacecdn.com/images01/79/l_d358560b9c651a8da68c084e88c6d0eb.jpg

The second pic is the top side of one firing control board. There are
5 of these in total. The IC at the top is the 4017. The black
rectangular boxes at the bottom of the board are the SPST reed relays
that activate the igniters.

http://a502.ac-images.myspacecdn.com/images01/82/l_86a86c8ec7313a72a4931cf2d91b4e75.jpg

The underside of the firing board. The wide trace on the right side of
the board is the +12V and the left side is ground. Are these too
small?

http://a861.ac-images.myspacecdn.com/images01/82/l_5aafdf9e0837402f366a766f5936906c.jpg

This is the Kit74 parallel port interface as well as the power
supply. I only have them sitting this close in order to fit them into
one picture. I have wired leads onto the Kit74 that goes to an RJ45
jack. That ethernet cable then goes to an identical jack where each
line is connected to the clock pin of one firing control board.

So in theory the operation goes like this:
The system is powered up and all counters and parallel port are
reset. The ignitor ground is then connected via a remote switch.

From there the software on my laptop plays the soundtrack and at set

intervals activates a single channel output on the parallel port.
That signal is then sent from the Kit74 interface via an ethernet
cable to the firing boards in the display. The 4017 in the firing
controller advances one count, activating the next relay in line. The
wire connected to the relay receives +12v which passes through the
nichrome igniter and returns to ground through the safety switch. The
wire heats, igniting the fireworks. The igniter typically burns up in
less than 1 second. The process repeats until all 9 cues have been
used.

 

[[email protected]]

The printer port puts out TTL compatible voltage levels and will not
reliably clock a 4017 powered from 12V. The 4017 will have problems
driving the typical relay coil directly. You should forget about
completing this project on time and ask your teacher for an extension.
You have a lot more homework and circuit modifications to perform before
you get close to finishing it.

The thing that confuses me is that in tests that I ran last week, the
system worked fine with only a test load. The clock pulses generated
by the Kit74 advanced the 4017 one step each time with zero errors. I
used the exact equipment I am now, including the ethernet cable to
deliver the clock pulses. It wasn't until I installed it into the
firing setup and did a test with an actual igniter that these problems
began to occur. After I increased the wire gauge for the power supply
to the firing boards, it had enough power to fire the igniters, but
began skipping steps and was far more susceptable to any sort of
interference. Even checking the power supply lines with a DVOM was
enough to move the counter several steps.

 

[Jim Thompson]

Hopefully these pics will show the full set up.

http://a548.ac-images.myspacecdn.com/images01/105/l_9f67b69cc9c5093d7535e07ff020383b.jpg

This is the inside of the firing box. The barrier strip at the top
brings the +12V and ground for the firing control boards (at bottom)
as well as a separate ground for the igniters that is controlled by a
safety switch. I added the cap to the barrier strip last night in an
attempt to help stiffen the power supply to be boards, but have not
yet tested it.

http://a340.ac-images.myspacecdn.com/images01/79/l_d358560b9c651a8da68c084e88c6d0eb.jpg

The second pic is the top side of one firing control board. There are
5 of these in total. The IC at the top is the 4017. The black
rectangular boxes at the bottom of the board are the SPST reed relays
that activate the igniters.

http://a502.ac-images.myspacecdn.com/images01/82/l_86a86c8ec7313a72a4931cf2d91b4e75.jpg

The underside of the firing board. The wide trace on the right side of
the board is the +12V and the left side is ground. Are these too
small?

http://a861.ac-images.myspacecdn.com/images01/82/l_5aafdf9e0837402f366a766f5936906c.jpg

This is the Kit74 parallel port interface as well as the power
supply. I only have them sitting this close in order to fit them into
one picture. I have wired leads onto the Kit74 that goes to an RJ45
jack. That ethernet cable then goes to an identical jack where each
line is connected to the clock pin of one firing control board.

So in theory the operation goes like this:
The system is powered up and all counters and parallel port are
reset. The ignitor ground is then connected via a remote switch.
intervals activates a single channel output on the parallel port.
That signal is then sent from the Kit74 interface via an ethernet
cable to the firing boards in the display. The 4017 in the firing
controller advances one count, activating the next relay in line. The
wire connected to the relay receives +12v which passes through the
nichrome igniter and returns to ground through the safety switch. The
wire heats, igniting the fireworks. The igniter typically burns up in
less than 1 second. The process repeats until all 9 cues have been
used.

Schematic showing how the 4017 is driving relay would help.

...Jim Thompson

 

[Fred Bloggs]

The thing that confuses me is that in tests that I ran last week, the
system worked fine with only a test load. The clock pulses generated
by the Kit74 advanced the 4017 one step each time with zero errors. I
used the exact equipment I am now, including the ethernet cable to
deliver the clock pulses. It wasn't until I installed it into the
firing setup and did a test with an actual igniter that these problems
began to occur. After I increased the wire gauge for the power supply
to the firing boards, it had enough power to fire the igniters, but
began skipping steps and was far more susceptable to any sort of
interference. Even checking the power supply lines with a DVOM was
enough to move the counter several steps.

Where are you taking the pulses off the kit74 board? That board has
eight relay outputs, and possibly eight open collector outputs. It would
not be a good thing to use a relay closure to generate the clock pulse
for your 4017. We are starting to suspect that your 4017 interface to
the e-match drive relays is inadequate. The 4017 has a high output
impedance and cannot maintain 12V at its output except for maybe a few
hundred microamperes loading. Your relays are almost certainly several
10's of milliamps loading, this will load the 4017 output to something
less than 8 volts and the relay will not activate.

 

[[email protected]]

Where are you taking the pulses off the kit74 board? That board has
eight relay outputs, and possibly eight open collector outputs. It would
not be a good thing to use a relay closure to generate the clock pulse
for your 4017. We are starting to suspect that your 4017 interface to
the e-match drive relays is inadequate. The 4017 has a high output
impedance and cannot maintain 12V at its output except for maybe a few
hundred microamperes loading. Your relays are almost certainly several
10's of milliamps loading, this will load the 4017 output to something
less than 8 volts and the relay will not activate.- Hide quoted text -

- Show quoted text -

I've soldered leads onto the resistors on the kit74. So (as best I
can tell), I'm getting the signal out of the 2018 directly. I
initially was going to use the relays, but could not find a suitable
debounce solution so I opted instead to go right to the amplified
output of the 2018. Up until I put a full load on the relays,
everything worked fine. Which to me is the confusing part (which I'm
sure is 99% my inexperience combined with ignorance). If the setup
will work fine when the only load on the relay is an LED, then why do
you suspect that the 4017 is unable to power the relays? My
understanding (again, probably wrong) is that the 2018 is basically 8
transistors grouped together on a single IC. They take the 5 volt
input from the parallel port and reproduce the same signal with a 12
volt supply. At that point a pulldown resistor keeps the clock pin on
the 4017 low until a 12 volt signal is received via the 2018. When
the 12 volt signal is removed, then the 4017 sees that as one complete
clock pulse and advances to the next output. In practice, this model
has seemed to work. The firing control board will move one step per
pulse from the 2018 which is verified both by the LED indicator on the
board as well as an LED attached to the lead that runs from the relay
to the igniter.

At this point in the day, it's looking like I'm going to have to go to
plan B. But this process has definently been a learning experience,
and I'd like to continue that, even if it means not finding success
until next 4th of July.

Thanks everyone for all of your help on this. I certainly wish I'd
found this group sooner, probably would have saved myself a lot of
trial & error and subsequent headaches

 

[Fred Bloggs]

I've soldered leads onto the resistors on the kit74. So (as best I
can tell), I'm getting the signal out of the 2018 directly. I
initially was going to use the relays, but could not find a suitable
debounce solution so I opted instead to go right to the amplified
output of the 2018. Up until I put a full load on the relays,
everything worked fine. Which to me is the confusing part (which I'm
sure is 99% my inexperience combined with ignorance). If the setup
will work fine when the only load on the relay is an LED, then why do
you suspect that the 4017 is unable to power the relays? My
understanding (again, probably wrong) is that the 2018 is basically 8
transistors grouped together on a single IC. They take the 5 volt
input from the parallel port and reproduce the same signal with a 12
volt supply. At that point a pulldown resistor keeps the clock pin on
the 4017 low until a 12 volt signal is received via the 2018. When
the 12 volt signal is removed, then the 4017 sees that as one complete
clock pulse and advances to the next output. In practice, this model
has seemed to work.

Okay- that part is okay, except the 4017 advances on the low-to-high
transition of its clock input. When you address the wired output on the
kit74, that output pulls low, and when you de-address it, the output
returns to a quiescent high and triggers the 4017.
Generally, logic circuits are not used to drive relay coils. We usually
use some kind of buffer like so:
View in a fixed-width font such as Courier.

...
..
..
..
.. +12V
.. |
.. .-----+--o-.
.. | | |
.. | - |
.. 1n4001 - |/| <
.. ^ |/| |
.. | |/| '--o---> to ignitor
.. | -
.. | |
.. '-----+
.. |
.. |
.. |/ 2n2222a
.. 4017 >---[10k]--| or
.. output |> 2n3904
.. |
.. |
.. ---
.. gnd
..

Something like this will avoid that quirky behavior you were seeing...

The firing control board will move one step per
pulse from the 2018 which is verified both by the LED indicator on the
board as well as an LED attached to the lead that runs from the relay
to the igniter.

At this point in the day, it's looking like I'm going to have to go to
plan B. But this process has definently been a learning experience,
and I'd like to continue that, even if it means not finding success
until next 4th of July.

That would be best.

 

[David L. Jones]

Hopefully these pics will show the full set up.

http://a548.ac-images.myspacecdn.com/images01/105/l_9f67b69cc9c5093d7...

This is the inside of the firing box. The barrier strip at the top
brings the +12V and ground for the firing control boards (at bottom)
as well as a separate ground for the igniters that is controlled by a
safety switch. I added the cap to the barrier strip last night in an
attempt to help stiffen the power supply to be boards, but have not
yet tested it.

http://a340.ac-images.myspacecdn.com/images01/79/l_d358560b9c651a8da6...

The second pic is the top side of one firing control board. There are
5 of these in total. The IC at the top is the 4017. The black
rectangular boxes at the bottom of the board are the SPST reed relays
that activate the igniters.

http://a502.ac-images.myspacecdn.com/images01/82/l_86a86c8ec7313a72a4...

The underside of the firing board. The wide trace on the right side of
the board is the +12V and the left side is ground. Are these too
small?

http://a861.ac-images.myspacecdn.com/images01/82/l_5aafdf9e0837402f36...

This is the Kit74 parallel port interface as well as the power
supply. I only have them sitting this close in order to fit them into
one picture. I have wired leads onto the Kit74 that goes to an RJ45
jack. That ethernet cable then goes to an identical jack where each
line is connected to the clock pin of one firing control board.

So in theory the operation goes like this:
The system is powered up and all counters and parallel port are
reset. The ignitor ground is then connected via a remote switch.>From there the software on my laptop plays the soundtrack and at set

intervals activates a single channel output on the parallel port.
That signal is then sent from the Kit74 interface via an ethernet
cable to the firing boards in the display. The 4017 in the firing
controller advances one count, activating the next relay in line. The
wire connected to the relay receives +12v which passes through the
nichrome igniter and returns to ground through the safety switch. The
wire heats, igniting the fireworks. The igniter typically burns up in
less than 1 second. The process repeats until all 9 cues have been
used.

Problems:
1) You are driving the relays directly with a 4017, that's *bad*. A
4017 will not have enough output current to do that. You are using low
power reed relays, but it's still too much for a 4017.

2) You can't use a relay driver board to provide clock pulses to a
4017, in fact this probably the worst possible way I can think of to
do this. You'll get ridiculous amounts of contact bounces causing
multiple pluses.

3) You are feeding power to the remote boards via the long ethernet
cable that is also the same ground line as the 4017. Power for the
4017 board should be local.

I'm surprised it even worked at all with no load.

At the very least, this might help until you can design the whole
thing properly from scratch - add a single-shot circuit to the 4017
board clock input so you are guaranteed to only generate one clock
pulse from how ever many erroneous input pulses.

Dave.

 

[[email protected]]

[...]

Problems:
1) You are driving the relays directly with a 4017, that's *bad*. A
4017 will not have enough output current to do that. You are using low
power reed relays, but it's still too much for a 4017.

2) You can't use a relay driver board to provide clock pulses to a
4017, in fact this probably the worst possible way I can think of to
do this. You'll get ridiculous amounts of contact bounces causing
multiple pluses.

I've always taken for granted (meaning I never checked) that reed
contacts don't bounce. Not that that would make them an appropriate
device for clock generation.

Anno

 

[ehsjr]

The thing that confuses me is that in tests that I ran last week, the
system worked fine with only a test load. The clock pulses generated
by the Kit74 advanced the 4017 one step each time with zero errors. I
used the exact equipment I am now, including the ethernet cable to
deliver the clock pulses. It wasn't until I installed it into the
firing setup and did a test with an actual igniter that these problems
began to occur.

The voltage to the 4017 is drooping due to the current
drawn by the igniter, and a spike is bing introduced.
You need to isolate the supply to the 4017 from the
voltage droop and spikes caused by the igniters.
That is the _first_ step, and you may not need to
do anything else. Add a diode and 2 caps as shown:


----o----igniter---gnd
Vsupply + ---+------------}}------------o---^
| Igniter relay
\ / D1 contact
----
|
+-------+---Vcc to 4017
| + |
[470uF] [.1uf]
| |
Gnd ---------+-------+


The above shows only a partial schematic for the addition
of caps & diode. You may need to do other things, but if
the igniters and the 4017 share the same supply, you must
do the above as the first step.

Ed


After I increased the wire gauge for the power supply

 

[David L. Jones]

[...]

Problems:
1) You are driving the relays directly with a 4017, that's *bad*. A
4017 will not have enough output current to do that. You are using low
power reed relays, but it's still too much for a 4017.

2) You can't use a relay driver board to provide clock pulses to a
4017, in fact this probably the worst possible way I can think of to
do this. You'll get ridiculous amounts of contact bounces causing
multiple pluses.

I've always taken for granted (meaning I never checked) that reed
contacts don't bounce. Not that that would make them an appropriate
device for clock generation.

Dry reed relays bounce like any other relay. The special mercury
wetted type do not.
The relays on the OP's parrallel port providing the clock pulse are
not reed, let alone mercury wetted.

Dave.

 

[Jasen Betts]

I installed the final power setup (on the tests I just had the board
jumpered onto the power supply) and now it will fire the ignitors.
The problem is that the counter is now behaving irratically. It will
skip several counts when receiving a clock pulse and will jump around
with the slightest change in ground (even connecting a single lead
from the multimeter to any point on ground advances the count). I've
tried several different caps (.01-10 uF) connected directly to the
power leads of the 4017 with no luck. All of the info I can find on
decoupling keeps talking about selecting the capacitor based on
frequency. But I'm only sending a single pulse of about 25 ns every
10-20 seconds. Does this mean I can use a larger cap? I'm seriously
freaking out at this point cause I only have 24 hours to get this
working. Thanks for the help!

use a separate supply to run the ignitors to the one that runs the
logic and relays. Or put a diode before the logic and use a large
capaciror across the logic supply.

Bye.
Jasen