Live Steam article showing capacitor in parallel with LEDs in Signal Head

This forum is dedicated to Riding Scale Railroading with propulsion using other than steam (Hydraulics, diesel engines, gas engines, electric motors, hybrid etc.)

Moderators: Harold_V, WJH

Atkinson_Railroad
Posts: 91
Joined: Mon Jun 08, 2015 6:27 pm
Location: Michigan
Contact:

Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby Atkinson_Railroad » Thu Aug 10, 2017 9:34 pm

I’m on a roll. I sense the crowd will forgive me.

The recent article featured in March/April 2017 of Live Steam magazine written by Dan Swanson shows a schematic diagram of a three light-emitting-diode signal head (drawing 3 page 17) with a .01 micro farad capacitor connected in parallel with the individual LEDs.
There is no reference in the article as to the purpose for the capacitors. I was initially thinking they were in place to allow a “glow down” appearance for mimicking an incandescent lamp during “flashing mode”. A .01 value would not seem to be of sufficient value for that purpose. Any dwellers among the crowd here able to share what they know about why a capacitor is in parallel with the LED in the article and its function?

John

John Hasler
Posts: 272
Joined: Tue Dec 06, 2016 4:05 pm
Location: Elmwood, Wisconsin

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby John Hasler » Fri Aug 11, 2017 5:30 pm

Impossible to say without seeing the schematic.

User avatar
BigDumbDinosaur
Posts: 451
Joined: Tue Jun 28, 2011 9:19 pm
Location: Midwestern United States

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby BigDumbDinosaur » Fri Aug 11, 2017 6:18 pm

Atkinson_Railroad wrote:I’m on a roll. I sense the crowd will forgive me.

The recent article featured in March/April 2017 of Live Steam magazine written by Dan Swanson shows a schematic diagram of a three light-emitting-diode signal head (drawing 3 page 17) with a .01 micro farad capacitor connected in parallel with the individual LEDs. There is no reference in the article as to the purpose for the capacitors. I was initially thinking they were in place to allow a “glow down” appearance for mimicking an incandescent lamp during “flashing mode”. A .01 value would not seem to be of sufficient value for that purpose. Any dwellers among the crowd here able to share what they know about why a capacitor is in parallel with the LED in the article and its function?

John

As Mr. Hasler noted, one would be guessing without seeing the schematic. My educated guess is 0.01 µF is too small with a high intensity LED to cause any persistence. It would have some effect with a T1 or T1-3/4 LED, but those aren't really appropriate for riding scale signalling.

Incidentally, placing a capacitor directly across a switched DC load can cause relay contact sticking.
Science makes it known. Engineering makes it work.

Atkinson_Railroad
Posts: 91
Joined: Mon Jun 08, 2015 6:27 pm
Location: Michigan
Contact:

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby Atkinson_Railroad » Fri Aug 11, 2017 9:48 pm

The 10 page article was carefully re-read to make sure the purpose of the capacitor was not overlooked.
After reviewing every page once again, no description is found explaining the function of the capacitor.

Tossing out a wild guess, I wonder if it has something to do with eliminating "noise". I don't know. I'm pleading ignorant on this one.

John
Attachments
DSC79B.JPG
March/April 2017 Live Steam & Outdoor Railroading
"Iconic Railroad Fixture Gets a Redo" By Dan Swanson

User avatar
DianneB
Posts: 443
Joined: Tue Aug 21, 2012 3:05 pm
Location: Manitoba, Canada

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby DianneB » Sat Aug 12, 2017 1:48 pm

The only reason for the capacitors would be to suppress noise if the LEDs were being multiplexed - a fairly common technique for illuminating multiple LEDs at a fraction of the current required if they were all on. The resistor values would indicate that the LEDs are being multiplexed. Without multiplexing, the LED current would be 36mA which is awfully high. Multiplexed at 1/3 duty cycle, the average current would be 12 mA - just about perfect.

hammermill
Posts: 2938
Joined: Sun Jun 27, 2010 10:43 pm
Location: pendleton or

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby hammermill » Sat Aug 12, 2017 11:55 pm

i agree with the caps acting as a filter for ac noise here texas instruments take on the matter

High-brightness LEDs are available today with forward currents more than
100 times greater than their predecessors. These new devices are not just high
brightness, but are high power as well. Single die with dissipations of 5W and
multi-die modules with power in excess of 25W are now available. The
requirements of high efficiency and low dissipation dictate a switching power
supply for this new generation of High-Brightness (HB), High-Power (HP)
LEDs, as a voltage regulator and a current limiting resistor are no longer
appropriate. High-brightness, high-power LEDs require a constant-current
source to take full advantage of their ever-increasing luminous efficiency and
vibrant, pure color. The topology of choice for this new breed of switching
constant current sources is the basic buck converter. The most convincing
argument for using a buck converter is the ease with which this simple
DC-DC converter can be turned into a constant-current source. This article
will explain the selection of, or possible exclusion of, an output capacitor
when designing a buck regulator for constant-current drive of HB LEDs.
No. 116
NEXT ISSUE:
Design Challenges In Step-Down
Regulator Applications
POWER designer
Expert tips, tricks, and techniques for powerful designs
Feature Article................1-7
LED Drivers for
1W/3W LEDs ......................2
60V Low-Side MOSFET
Controller............................4
Switched Capacitor
Flash LED Driver..................6
Power Design Tools..........8
Driving LEDs: To Cap or Not to Cap
— By Chris Richardson, Applications Engineer
VIN
L
+
D +
– CO
RFB2
IO
RFB1
+
VO

VIN
L
+
D +
– CO
IO
RSNS
+
VF

Figure 1a. Traditional Buck Voltage Regulator Figure 1b. Buck Current Regulator
2
New LED Drivers Power 1W/3W LEDs up to 1A
LM3402/04 Regulators Offer Low Current-Sense Voltage to Reduce Power Loss
and Deliver High Efficiency up to 95%
For FREE samples, datasheets, and more information, visit
www.national.com/pf/LM/LM3402.html
www.national.com/pf/LM/LM3404.html
CB L1 VIN
CIN RON
IF
VIN
RON
RSNS
CF
BOOT SW
CS
DIM GND VCC
D1
LM3402/02HV
LM3404/04HV
Wide VIN range:
LM3402/04 = 6V to 42V
LM3402HV/04HV = 6V to 75V
LED drive currents up to:
500 mA for the LM3402/02HV
and 1.0A for the LM3404/04HV
Current sense voltage of 200 mV
reduces the power loss in RSNS
DIM pin for fast PWM dimming
LM3402/04 Features
• Easy-to-use and low external parts count
• Switching frequency up to 1 MHz
• Constant on-time control with voltage feed-forward
for nearly constant switching frequency over the full
line voltage range
• Low 0.2V feedback reference minimizes power losses
• Low shutdown current when RON pin held low
• High voltage capability supports long strings of LEDs
• Choice of 0.5A or 1.0A peak current capability
(LM3402/LM3404)
DIM
ILED
4 µs/DIV
5V/DIV
100 mA/DIV
T
Dim Response
Typical Application Diagram
Ideal for use in automotive, industrial lighting, gaming/vending machines, general illumination, and
architectural lighting applications
Controlled Current
The buck regulator is uniquely suited to be a
constant current driver because the output
inductor is in series with the load. Regardless of
whether a buck regulator is used as a voltage source
or a current source, selection of the inductor forms
the cornerstone of the system design. With an
inductor in series with the output, the average
inductor current is always equal to the average output
current, and the buck converter naturally
maintains control of the AC-current ripple. By
definition, the LED drive is a constant load system;
hence a large amount of output capacitance is not
necessary to maintain VO during load transients.
No Output Cap Yields High Output Impedance
In theory, a perfect current source has infinite
output impedance, allowing the voltage to slew
infinitely fast in order to maintain a constant
current. For switching regulator designers who
have concentrated on voltage regulators, this
concept may take a moment to sink in. Completely
removing the output capacitor from a buck
regulator forces the output impedance to depend
on the inductor. Without any capacitance to
oppose changes in VO, the output current (referred
to as forward current, or IF) slew rate depends
entirely upon the inductance, the input voltage,
and the output voltage. (VO is equal to
the combined forward voltage, VF, of each
series-connected LED)
LED manufacturers generally recommend a ripple
current, ΔIF, of ±5% to ±20% of the DC forward
current. Over the typical switching regulator
frequency range of 50 kHz to 2 MHz the ripple
itself is not visible to the human eye. These limits
come from increasing thermal losses at higher
ripple current (a property of the LED semiconductor
PN junction itself) and a practical limit to the
inductance used. The percentages are similar to the
recommended current ripple ratio in buck voltage
regulators. Inductor selection for a fixed-frequency
current regulator is therefore governed by the same
equations as a voltage regulator:
One difference is that the inductance used for
current regulators without output capacitors tends
to be higher because the drive currents for the
emerging standards of 1W, 3W, and 5W HB LEDs
are 350 mA, 700 mA, and 1A respectively. Modern
buck voltage regulators tend to use
inductors in the range of 0.1 µH to 10 µH with
saturation currents from 5A to 50A. Current
drivers at similar switching frequencies tend to
require inductors ranging from 10 µH to 1000 µH
and saturation currents ranging from 0.5A to 5A.
The main goal of high output impedance is to
create a system capable of responding to PWM
dimming signals, the preferred method of controlling
the light output of LEDs. The dimming signal
might be applied to the enable pin of the regulator,
in which case the output current can slew from zero
to the target and back to zero without the delay of
CO being charged and discharged. For even faster,
higher resolution dimming, a shunt switch, usually
a MOSFET, can be placed in parallel with the LED
array, allowing the continuous flow of current at all
times. Again, with no output capacitor to slow the
slew rate, dimming frequencies into the 10’s of kHz
are possible. This is a critical requirement in applications
such as backlighting of flat-panel displays,
and the creation of white light using an RGB arrayUsing an Output Capacitor Reduces Size and Cost
Some amount of output capacitance can be useful
as an AC current filter. Applications such as retrofitting
of incandescent and halogen lights often
require that the LED and driver be placed in a
small space formerly occupied by a light bulb.
Invariably the inductor is the largest, most expensive
component after the LEDs themselves. For the
sake of efficiency (especially important in cramped
quarters), the designer generally chooses the lowest
switching frequency that allows the solution
(mostly the inductor) to fit. Allowing a large ripple
current in the inductor and filtering the LED
current results in a smaller, less expensive solution.
For example, to drive a single white LED
(VF ≈ 3.5V) at 1A with a ripple current ΔiF of ±5%
from an input of 12V at 500 kHz would require a
50 µH inductor with a current rating of 1.1A. A
typical ferrite core device that fits this application
might be 10 mm square and 4.5 mm in height. In
contrast, if the inductor ripple current is allowed to
increase to ±30% (typical for a low-current voltage
regulator) then the inductance required is less than
10 µH, and an inductor measuring 6.0 mm square
and only 2.8 mm in height size can be used. The
output capacitance required is calculated based on
the dynamic resistance, rD, of the LED, the sense
resistance, RSNS, and the impedance of the
capacitor at the switching frequency, using the
following expressions:
Typical values for output capacitors range from
0.1 µF to 10 µF, a perfect fit for ceramic capacitors.
In many applications, the addition of an output
capacitor reduces both the size and the cost of the
total solution.
Output Capacitor Placement
For buck regulators that use PWM-based control,
such as Voltage Mode (VM) and Current Mode
(CM) the output capacitor should be connected
from the regulator output to system ground,
power.national.coidentical to a normal buck regulator. (Figure 3a)
This way, the control-to-output transfer function
of the system can be analyzed with the same
equations used when designing a voltage regulator.
When using comparator-based control, such as
hysteretic or Constant on-Time (COT), the output
capacitor should be placed in parallel with the LED
array. (Figure 3b) In hysteretic voltage regulator
circuits, this technique is often used to increase the
percentage of in-phase voltage ripple at the
feedback node. For the current regulator, it forces
both the ripple current through CO and the
forward current through the LEDs to sum at the
input to the switching comparator. The voltage
waveform across RSNS is therefore in-phase with the
switching node waveform, and the result is
predictable operation with high noise rejection.
The combination of low output capacitance and
high inductor current ripple actually makes
hysteretic and COT current regulators more
reliable and easier to design than voltage regulators.
Conclusion
The high brightness, high power LED represents the
biggest change in lighting design since the introduction
of fluorescent bulbs. Using LEDs requires a
fundamental change in the complexity of electronics
used for lighting systems. Currently a large portion
of LED lighting design is retrofitting of incandescent,
halogen, and fluorescent installations. Such
systems rarely include sophisticated dimming
control, and place a high value on small size. These
are the applications where an output capacitor is a
welcome addition to the driver circuit.
In the future, the higher cost of LEDs for general
lighting will be balanced by new levels of control
over brightness, tone, and color. Lighting in homes
and businesses will require fast PWM dimming,
requiring current drivers to minimize or eliminate
their output capacitance. These systems will draw
upon experience from today’s fast-dimming
applications which have already shed the output
capacitor to provide the best response time. ■
7
POWER designer
Driving LEDs: To Cap or Not to Cap
power.national

Atkinson_Railroad
Posts: 91
Joined: Mon Jun 08, 2015 6:27 pm
Location: Michigan
Contact:

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby Atkinson_Railroad » Sun Aug 13, 2017 2:25 pm

Appreciate the copy-n-paste article/document “To Cap or Not to Cap” hammermill.

Thank you for sharing Chris Richardson’s writing.

I may have made a mistake posting a question of electronic nature within the scope of a mainly machining focused forum.

As best as can be understood by the Live Steam article, there is no mention of the LED signals being
driven or controlled by a sophisticated system. Multiplexing was mentioned earlier for example.

It doesn’t help that those peeking in on this thread are unable to read and see what I’m talking about.

Reflecting back, a better approach would have been to sign up to the Live Steam forum and directly
ask about the article in their forum. From there, the publisher(s) maybe could at least contact the author
of the original article for clarification.

Another aspect I’m reminded of is how an illustration can sometimes confuse rather than inform.
If the elements depicted in the drawing posted earlier were significant enough to augment the written portion of the article,
perhaps the lack of clarity falls on the editor of the piece for not better explaining what the reader was/is looking at in the drawing.

It's possible someone else may have additional info to share about the circuit.

John

John Hasler
Posts: 272
Joined: Tue Dec 06, 2016 4:05 pm
Location: Elmwood, Wisconsin

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby John Hasler » Sun Aug 13, 2017 2:39 pm

Dianne writes:
> Without multiplexing, the LED current would be 36mA which is awfully high.

More like 25ma taking the forward drops into account. In any case if the 12 volt source is just DC the capacitors won't do anything.

BryceGTX
Posts: 640
Joined: Sat Dec 01, 2007 9:17 pm
Location: Michigan

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby BryceGTX » Wed Aug 16, 2017 9:25 pm

Atkinson_Railroad wrote:. Any dwellers among the crowd here able to share what they know about why a capacitor is in parallel with the LED in the article and its function?

John


No question of why this cap is here. It is used to shunt high frequency RF signals around the diode. If you place an LED between a very long free wire and AC ground, it will turn on if anywhere near any broadcast station or high RF source.

The AC ground is through the 12 battery. The antennas are the three other wires connected on the other side of the diode. Some of you may have experience with crystal radios.

Hook a high impedance head side on this circuit without cap and you may hear a radio station :D

Bryce

Atkinson_Railroad
Posts: 91
Joined: Mon Jun 08, 2015 6:27 pm
Location: Michigan
Contact:

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby Atkinson_Railroad » Thu Aug 17, 2017 3:40 am

Hi Bryce;

Your term “shunt” gave me a better word to use in seeking additional information.

I’m also finding info stating, “…The shunt capacitor across the LED also helps filter the pulsating DC voltage to the bulbs and reduces annoying flicker.”

The above makes sense to me because the signal is likely being fed by a power supply other than a steady battery source.

Thanks for your help on this!

John

Cary Stewart
Posts: 437
Joined: Wed Jun 08, 2011 5:54 pm

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby Cary Stewart » Thu Aug 17, 2017 7:27 pm

Kind of related to this thread. In the late 1970s my company, part of a large conglomerate, built on spec. for the Army a 4ft X 4ft led panel of thousands of red/yellow LEDs which gave three colors. This panel needed a monstrous power supply which we called Boulder Dam. Soooooo. our power supply engineers and others designed a very high speed switching power supply in combination with digital routing that made the power requirements manageable. The idea was to put mylar overlays of battle ground maps over the top of the panel and with programing show in real time the battle action. The cost - oh my. We built one, the Army choked and we did something else to solve the problem. It was demonstrated and it worked as advertised in a tent out doors.
One of the little things that we discovered is that for flicker free vision of LEDs pulses of 70 hz minimum must be used. The pulse width can be very, very short.

Best, Cary Stewart

BryceGTX
Posts: 640
Joined: Sat Dec 01, 2007 9:17 pm
Location: Michigan

Re: Live Steam article showing capacitor in parallel with LEDs in Signal Head

Postby BryceGTX » Thu Aug 17, 2017 8:58 pm

This capacitor has nothing to do with pulsating D.C. If you are worried about pulsations from a switcher, you have a poor switcher. A better way to reduce those pulsations is a cap between 12v and ground at the supply.

No engineer will put a cap here to reduce pulsations. It's just the wrong place to put the cap. A cap to reduce fluctuations is best applied at a higher voltage drop point to increase the effect of the capacitance. The voltage drop across the LED may only be 1 volt or so. It would be a waste of a cap here.

And it is clearly too small to provide slow turn on/off.

It clearly is for RF noise. Keep in mind you will have a very long wire between this signal and the signal house. That long wire is the antenna. Please take my advice, I have quite vivid experience with RF and LEDs.

Also, this circuit is not multiplexed nor PWMed. There is would be no reason given the trivial design.

The term "shunt" as applied to this design refers to a circuit that tends to short circuit The RF energy around the LED rather than let it go through the LED.

An interesting tid-bit of information.. early automotive designers could not figure why their new car stopped every time it came to a certain place on a certain road.. turns out the radio station there caused the new fangled microcomputer controlling the engine to go haywire.

Thanks for listening,
Bryce


Return to “Riding Scale Railroading”

Who is online

Users browsing this forum: No registered users and 2 guests