Passive Vs Active decoupling

Hi,
I am using a 555 Timer to generate a 38Khz IR signal from the same battery as the electronics in my RC Car.

Whatever combination of capacitors I try I cannot eliminate the electrical noise coming from the 555 Timer (its switching about 30ma).

While looking for other options I came across the idea of using a regulator (7805) as an active decoupling component, this is working perfectly.

So the question is - legitimate technique or am I missing a trick with passive components ?

Duane B

rcarduino.blogspot.com

Errr...that is a bit unusual, perhaps we can see your schematic before and after the 7805 was added? The 7805 has a fairly poor high-frequency response, which is really what decoupling is for, so I'd be surprised if it was really acting as an "active decoupling component".

--
The Gadget Shield: accelerometer, RGB LED, IR transmit/receive, speaker, microphone, light sensor, potentiometer, pushbuttons

Hi,
Its a very ordinary 555 Astable thats been working well for six weeks or more as a 38Khz IR modulator. The noise is only a problem when it is installed in an RC Car and there is no transmitter signal - at which point the car electronics, particularly the servo is very susceptible to noise in the power circuit.

I have run the circuit in the car with its own power and it is fine, I have run the circuit in the car with the 7805 and it is fine, however using passive components - capacitors of all types and values and in various combinations across the power lines I cannot eliminate the noise.

While I would not normally have the car turned on without a transmitter signal its no reason not to eliminate the noise that is clearly there.

So for the sake of clarity my current solution is the 7805 its working perfectly, the inspiration came from here -

http://williamson-labs.com/480_dec.htm

And the fact that my Arduino is powered through a regulator and while switching loads causes no interference where sharing power with the rest of the car.

But back to the original question - while it works is it a legitimate approach or is there something I am missing that could be achieved with passive components ?

Duane B

After a few more hours of testing - the 7805 'active decoupler' is working perfectly and I am still not able to get anywhere with capacitors alone.

Duane B.

rcarduino.blogspot.com

What from that page gave you inspiration? I'm not sure I understand how you are using the 7805. I also think some actual example if how you were using capacitors would help. "Every combination" isn't accurate nor descriptive.

To remove noise from power lines you require more than just capacitors.

Caps are great for killing some noise, but they won't get it all.

What you really want is a "PI" LC filter. This consists of two capacitors across the power lines, with an inductor between them, in the shape of the greek letter "pi":

You can read more on them at Wikipedia: Capacitor-input filter - Wikipedia

I am using a 555 Timer to generate a 38Khz IR signal from the same battery as the electronics in my RC Car.

Whatever combination of capacitors I try I cannot eliminate the electrical noise coming from the 555 Timer (its switching about 30ma).

From your description, it's not at all clear what you mean by "electrical noise coming
from the 555 Timer". What is being affected by this noise? What's not working?

In general, for such a situation, it's usually the exact opposite. The motors generate
so much noise that the rest of the electronics is disrupted.

Heh, I bet ya that 7805 is acting like a 1N4001 in that circuit. The motor was giving him that AC noise ;).

The 7805 is a perfect answer to the noise issue as the 555 WILL work up to a MHz or so and the switching noise is hard to deal with. Try the Pi filter though... use a coil of about a millihenry or so. as long as it's DC resistance isn't an issue you can use anything above a millihenry, I would add a couple of .1 uf caps, one right on top of the "Triple Nickle" (NE555) an alternative would be the same Pi configuration with a resistor in the range of 10 to 47 ohms in place of the choke and also use low ESR caps as well. My "Noise Cocktail" are 1nF, 100nF and 100uF caps placed right at the noise source and at the entry point of the power to the board. I Never have noise issues. Because I am aware of just how well "Distributed" capacitors work on a PCB. I designed PCB's and the circuitry on them for 20 years... Successfully.

Doc

Heres a quick suggestion/request - I don't have a scope, but many of you do. It would be a interesting and useful exercise for the community to see the noise generated by a 555 on the power lines of a circuit and see how this can be reduced or eliminated.

As a common application of the 555 is to generate a 38khz IR signal this would be a good example circuit.

What you really want is a "PI" LC filter. This consists of two capacitors across the power lines, with an inductor between them, in the shape of the greek letter "pi":

The 7805 is a perfect answer ..... Try the Pi filter though

Thanks, I will also give the PI network a try, any suggestions on cap values given that I will be trying the resistor approach as it will be some time before I can get hold of an inductor ? I am switching at 38khz.

Thanks

Duane.

Update -

I have tried the PI Network using 10 and 100 Ohms for R and C1,C2 = 1,10,100uf without success.

Prior to this I have tried .01,0.1,1,10,100,470,4700 capacitors across the power rails individually and in combination all with no observable improvement.

I am sure that there is a solution using readily available passive components however without a scope I cannot measure the nature of the noise nor an improvement I may be achieving.

The only successful strategy so far is using the 7805 as a regulator, its isn't fussy about the caps used, it just works. Currently I have two 100uf capacitors across input to ground and output to ground as you would find in any typical application of a 7805.

The noise I want to suppress is coming from the 555, I am not providing an RC Signal and so the ESC is not arming and no power is going to the motor. The noise is observable as very pronounced erratic movement of the steering servo - much more than would be described as jitter or juddering. Again the 7805 completely cures this, I would like a passive solution, but so far have had no success.

While the car would not be operated with no RC Signal, the fact that I can eliminate the noise through the 7805 shows that the 555 is the source and at the extremes of my transmitter range when the ESC has already armed, the spurious signals from the 555 may be enough to cause a runaway.

Again, is there anyone with access to a scope that could setup a 555 Timer based IR Driver and measure the noise then apply a few common approaches with passive components and measure their effectiveness.

I have a solution, its the 7805, but if I had a scope I would still be working on a passive approach.

Any help appreciated

Duane B

rcarduino.blogspot.com

Firstly 30mA isn't that high, I suspect the problem isn't the load, its either cross-talk from the load to sensitive circuitry or that you're using a standard NE555 instead of the far better CMOS version of the 555 (the standard 555 crow-bars the supply heavily, to quote "The Art of Electronics': "The 555 generates a big (~150mA) supply-current glitch during each output transition. Be sure to use a hefty bypass capacitor near the chip").

A 10uF ceramic and 1000uF low-ESR electrolytic might be a good starting point I think, best to avoid the original 555 entirely - the ICM7555 actually boasts "No Crowbarring of Supply During Output Transition" in the datasheet!

The 7805 probably works by letting the supply voltage to the 555 drop a couple of volts on each transition without pulling its input low (its too slow to follow the 555's glitch!)

DuaneB:
I have a solution, its the 7805, but if I had a scope I would still be working on a passive approach.

rcarduino.blogspot.com

Have you tried a 1N4001 diode in place of the 7805?

I have a diode where my power enters the board, why do you think a diode will help ?

Duane B

rcarduino.blogspot.com

DuaneB:
I have a diode where my power enters the board, why do you think a diode will help ?

Duane B

rcarduino.blogspot.com

I am just trying to figure out how that 7805 is attenuating the noise from your 555. My thought was putting a diode on the 555 pin 8 after any junctions.

I assume your 555 circuit is similar to this:

The only thing I could observe on the 7805 (unpowered) is ~1 volt difference on Vout switching directions and 5.7kohm (no difference in direction). Pretty much throws a wrench into my theory. You have quite an enigma wrapped in bacon :wink:

Hi,
Thats the enigma, the 7805 is not unpowered, its being used in a typically fashion i.e. as a 5 volt regulator.

It is acting as a very effective 'buffer' between the rest of the project and the noisy 555.

The 555 project is an RC Car so you can be certain that there are components looking for any scrap of noise they can try and turn into a useable signal.

A 10uF ceramic and 1000uF low-ESR electrolytic might be a good starting point I think, best to avoid the original 555 entirely - the ICM7555 actually boasts "No Crowbarring of Supply During Output Transition" in the datasheet!

I had read lots of articles suggesting that the 555 is notorious for noise, including some that suggest its very difficult to eliminate. As you mention there are a also references to newer versions being better behaved.

I am just finishing off soldering a 7805 version of my transponder but will try even more caps on the breadboard version in the next hour or so.

Thanks

Duane B

rcarduino.blogspot.com

On the wiki page I pointed to about the pi filter it does say:

Capacitor input filters can provide extremely pure dc supplies, but have fallen out of favour because inductors tend to be unavoidably heavy, which has led to the often-preferred choice of voltage regulators instead.

:slight_smile:

the ICM7555 actually boasts "No Crowbarring of Supply During Output Transition" in the datasheet!

Well, that could certainly expain the problem, as the R/C receiver could be cutting
out, or maybe the microcontroller in the receiver resetting itself due to brownout.

Doing a search on "555 timer crowbarring" shows the problem is classical, well
known to those who use the device.

Thats the enigma, the 7805 is not unpowered, its being used in a typically fashion i.e. as a 5 volt regulator.

It is acting as a very effective 'buffer' between the rest of the project and the noisy 555.

Apparently the reservior cap on the 7805 is effectively preventing the crowbar pulse
from passing backwards to the input side.

I had read lots of articles suggesting that the 555 is notorious for noise,

Good reason not to use one.

Hi,
The ESC is happy enough but the steering servo gets very upset. I have tried different combinations of ESC, Receiver and servo the current combination being the worst (best for testing).

Apparently the reservior cap on the 7805 is effectively preventing the crowbar pulse
from passing backwards to the input side.

I can't see it being just the cap, I assume that there is something in the current path back through the 7805 that is blocking the pulse - not sure what this might be or whether its an unintended application and will have an effect the 7805 lifespan.

It is looking like regulators are a recognised solultion in this situation though, in which case I am leaning towards adopting the 7805 for my final solution on the basis that its cheaper than a high capacity low ESR capacitor and is widely available.

Any comments ?

Duane B.

EDIT - or this - Renesas Electronics Corporation