Ripple Current ?

[PCBONEZ]

Re: Ripple Current ?

This is a linear power supply not a switcher.
You should probably have a load resistor in parallel with the cap.
There is no switcher so any ripple you have will be around 100-120 Hz, not in the kHz.
You do not need low ESR and you don't even need an electrolytic.
Most non-electrolytic caps are bi-polar by nature which fixes that problem.
.

[starfury1]

Re: Ripple Current ?

THANKS Davmax for that VERY BIG correction

So yeah "never" with bipolar setup (better really to buy a cap made that way)

Your right! I got the resistor sharing thing mixed up with Capacitors in Series Only!

+ --+[]------+[]----- -




DOH!

[Joe Rampolla]

Re: Ripple Current ?

Hi Folks,

Again, many thanks for all the input and the time to look over my questions!

I don't use or need a bridge in the loco when I am using the bipolar cap as the flywheel with DC to the track. I have several different situations. Obviously when I have a bridge in the loco, reversing polarity doesn't reverse the loco. The reason I am asking about ripple current is because I am assuming that if I reduce ripple to the flywheel cap, I can use a cap rated closer to the maximum track voltage and keep the cap's dimensions smaller as opposed to a higher rated cap in voltage, not capacitance. My concern about the ripple has nothing to do with the motor. So if the track is getting about 14 VDC, I would like to use a cap to be no bigger than 16 volts, and be safe. I am also assuming that the reason I have needed a 3300 mfd 35-50 volt cap is because of the ripple current.

The flywheel cap prevents stalling and gives a little coasting. The technique is successful and I use it many ways in HO, S and O27 (3-rail O gauge).

Let me ask this another way. If I were using 9 VDC from 6-7 D cell batteries to the track, is there any ripple current getting to the loco with a cap in parallel with the motor?


And if my assumptions are correct, if I were to build a highly filtered power supply, I would have little ripple current in the loco's flywheel cap.

Thanks!

Take care, Joe.

[PCBONEZ]

Re: Ripple Current ?

A "DC flywheel" makes no sense.
Do you mean you are using it as a temporary voltage storage device?
That would be more like a battery than an oscillator as 'flywheel' suggests.

Xc will be so low at 100-120Hz that you don't need to worry about ESR.
Xc at 110Hz w/3300uF is ~ 0.44 ohms.


So your concern is the sum of the DC plus the ripple?
-
That value is created by the characteristics of you bridge rectifier and your input AC voltage.

Your 14v is probably the RMS value and peak will be 1.414 times that.
14v x 1.414 = 19.8v peak.

You want at least a 25v cap.

.

[PCBONEZ]

Re: Ripple Current ?

You still haven't said how you are changing the motor's direction.
-
If you charge up the cap in one direction then reverse polarity then you are going to send one huge current spike through that motor.

.

[davmax]

Re: Ripple Current ?

Hi Folks,

Again, many thanks for all the input and the time to look over my questions!

I don't use or need a bridge in the loco when I am using the bipolar cap as the flywheel with DC to the track. I have several different situations. Obviously when I have a bridge in the loco, reversing polarity doesn't reverse the loco. The reason I am asking about ripple current is because I am assuming that if I reduce ripple to the flywheel cap, I can use a cap rated closer to the maximum track voltage and keep the cap's dimensions smaller as opposed to a higher rated cap in voltage, not capacitance. My concern about the ripple has nothing to do with the motor. So if the track is getting about 14 VDC, I would like to use a cap to be no bigger than 16 volts, and be safe. I am also assuming that the reason I have needed a 3300 mfd 35-50 volt cap is because of the ripple current.

The flywheel cap prevents stalling and gives a little coasting. The technique is successful and I use it many ways in HO, S and O27 (3-rail O gauge).

Let me ask this another way. If I were using 9 VDC from 6-7 D cell batteries to the track, is there any ripple current getting to the loco with a cap in parallel with the motor?


And if my assumptions are correct, if I were to build a highly filtered power supply, I would have little ripple current in the loco's flywheel cap.

Thanks!

Take care, Joe.

The circuit you have drawn does not require the rectifier, I assume that this is still in the loco. Ripple comes from rectified AC so no ripple from batteries. Reversing the loco requires reversing the voltage to the motor, this would destroy the cap shown. You can use a bipolar to solve this problem and retain the "flywheel" effect. See data sheet here: http://www.samxon.com/PDF/prodPDF_58.pdf

Looking at the data 2200uF and 3300uF in the 16V to 25V range have a diameter of 16 -18 mm.

One way to solve your problems is to use a variable regulated power supply with an output reversing switch. The variable control will allow you to control the train over a range of speeds, the "flywheel" will be unnecessary. You can buy these power supplies in kit form to save cost. You need a 0 -15Volt supply rated at a current to supply your needs, maybe 1-2Amps (so a 3-5 amp supply will be ideal). A small bench top power supply like this might do the job: http://www.circuithut.com/index.php?.../view/full/273

This one has both a positive and negative supply meaning that forwrad and reverse speeds could be set independently. A switch at the output could chnage direction. I am assumming that each output is overload protected to limit any current at change over.

The above is still patching a solution, the ideal is a train controller operating from a 12 volt supply see here: http://www.siliconchip.com.au/cms/A_101082/article.html

or here: http://www.sprint.net.au/~terbut/mtc/artindex.htm

You should be aware that most such train controllers use Pulse width switching to control speed. This method goes right away from your desire to minimise ripple. The chopped waveforms delivered to the motor have large ripple components because it does not matter with a motor.

[davmax]

Re: Ripple Current ?

Looking over the Internet I have been facinated by what is on offer for model train control eg computer control: http://www.cti-electronics.com/index.htm.

Joe, I guess you know all about this stuff.

[starfury1]

Re: Ripple Current ?

You posted will I was doing this davmax so you covered most of what I said and then some Thanks

The flywheel cap prevents stalling and gives a little coasting. The technique is successful and I use it many ways in HO, S and O27 (3-rail O gauge).

Yeah I think the caps is more like a battery in the above case, smoothing DC fluctuations to the motor.
(Flywheel, energy storage Kenetic, Wikki )
more I suppose from a mechanical aspect them term is being used and I guess the parallel is the smoothing action
(given frequency of ripple not quite sure why but)

Anyway don't know, thats my thought on it anyway.

So back to the original question.

I myself would used a regulated supply in that case
as far as cap voltage and cutting it fine is concerned.

As bonez has restated the AC RMS converted to DC will result in Higher DC Voltage So you may have to use a lower AC RMS to get the DC at the equivalent AC rms voltage.

the other aspect is I am assuming this will be a somewhat constantly changing DC in relation to control of speed
this may incur some sort or overshoot, as the feedback resettles
I dont know either way for sure thought.

To me if you wanted to use 16V
Id probably think don't go better then 14 DC Volts max
(wet figure in the air)

I know Bonez and davmax would have a better feel on this for how close you can cut it

I suppose really your limited by whats actually needed too.

Just on the circuit

Ripple I suppose is really a matter of Supply and Demand

With AC the supply is intermittent over time
(other words the cap gets a current refill in bursts)
The motor is Demand at a constant rate over time
(so the cap is supplying current to the motor all the time)

The difference between the 2 levels (voltage wise) will determine the amount of ripple your going to get from an unregulated supply circuit

Adding extra capacitance will reduce the difference between the 2 levels
there by reducing ripple (quite simply you got a bigger bucket to draw from)

One more statement
if the Demand is zero then the Ripple is zero
(ideally speaking)

So using a "constant current supply source" such as a battery will basically negate the ripple you would get from AC but other factors then may have an effect on DC voltage.

So basically your Supply is constant as is your Demand

in the case you got in the circuit above you would lose about 2V across the bridge. So say 7VDC on cap
its not needed really (apart from steering if thats your issue)

with that also since its 9V, a 16V Cap would be perfectly fine

Now if you can run the thing off 9 Volts DC
then a psu built at that DC voltage would be you answer.
(or possibly 12V)

I was assuming that the speed is controlled by the AC or DC voltage being variable

but I think when I had a look at throttles there seems to be more then one way its done.

Like I said no real clue with trains and how they work

Anyway hope that helps

If you think I got something wrong here guys please post.

Cheers

[stretch0069]

Re: Ripple Current ?

daah-eee


mikee like choo-choo......


<wanted to put the "dork" smilie in here but seems TC needs to do some work.......... <wink> see below.... >




i'm a tad inebriated, but are you looking for an el-cheapo way to do DCC????


ps-damn it TC.....we need some good smilies or i'm just gonna have to post pics of me in a plaid skirt, pig-tails, white shirt tied at the bottom, and sucking on a lolly-pop!!!!!!!!!!!!!!!!!!!!!!!

[starfury1]

Re: Ripple Current ?

humm let me see if I can put this in a nut shell

cap 3300uf 16 Volt
reason for this cap is physical size to fit in Loco

reason for cap in first place is smoothing supply for motor

problem
supply for cap so it dont go Psssstttt and do impersonations of volcano.

or i'm just gonna have to post pics of me in a plaid skirt, pig-tails, white shirt tied at the bottom, and sucking on a lolly-pop!!!!!!!!!!!!!!!!!!!!!!!

Not sure the forums ready for that stretch0069....... TC!!!!!
now if you were a good looking woman..... Hummmm then post away by all means


Cheers

[Joe Rampolla]

Re: Ripple Current ?

Hi Folks,

Many thanks for all the replies! Very helpful!

My one question has been addressed; a battery delivers no ripple!

Let me explain the "electronic flywheel" as I use it. When dirty tack prevents electrical contact, the loco stops, so the flywheel cap discharges and gives the motor a chance to coast past the bad section of track and onto a clean section. It works so well that I no longer have to clean track as often and I can even have intentional dead spots and the loco continues forward. I use it several ways, with or without a rectifier in the loco, on AC and DC, sometimes with a bipolar cap. When there is a rectifier or diode in the loco, I have no directional control of the item, usually. If there is no rectifier or diode in the loco and it is placed on the DC track in the wrong direction with incorrect polarity, the regular cap will pop! With toy train stuff, often there is a rectifier in the loco or other motorized item so it can be run on AC layouts. Some have a reversing board that allows me to reverse polarity to the motor. I don't always have a rectifier in the loco, but will be running that stuff on DC. There are many combinations and they all work and have been in use for years, not my idea or invention. And the flywheel cap was employed by some model railroaders before the DCC (digital command control) and after. Not all scales and equipment need the electronic flywheel. Some already have circuits that do approximately the same while being remotely controlled.

It is mostly short, inexpensive rail items with few points of electrical contact with the track that will benefit from the electronic flywheel. It is so simple but out-of-practice now with all the new developments with electronics in the hobby.

Some caps are smaller than others, so I have already tried to find the smaller sized caps. Just in theory I wanted no ripple getting to the rails.

I am afraid that my simple question has gotten more confusing. Basically I just want to build a simple circuit that produces little if any ripple just by adding filter caps from the supply, if it can be that simple. I have seen all sorts of hobby power supplies and throttle circuits. I have them in hobby books and they are on the Web. I don't need any help with the circuit in the loco. That circuit has proven itself. I need to build a simple highly filtered supply powered by an AC transformer.

So to summarize, the flywheel circuit isn't in question. I use it in several forms and it works very well. Sometimes so well I have to reduce the capacitance so equipment does eventually stop within a few inches. What I want to do find a simple way of filtering the ripple from rectified current. I have very large rectifiers and I am very conservative about wire size and size of components.

I was just wondering if those 3 caps in parallel would reduce most of the ripple. I was figuring that the first cap takes most of the ripple.

Is there I way I can measure the amount of ripple? Yes, my concern is in eliminating the ripple, if possible, by just good filtering. I want to understand the theory of filtering.

There is so much out there with all the new electronics, but that is not my interest or my need. I take the low-end, small stuff and try to get reliable dependable performance with simple things like just the capacitor.

I so much appreciate all the input!!!!!! I don't want to take up everyone's time with this.

Thanks!!!

Take care, Joe.

[Joe Rampolla]

Re: Ripple Current ?

Hi Folks,

I believe I have found the circuit I need, a capacitor input filter circuit:
capacitor input filter



So anyone have an idea of the rating of the choke I need if it is critical? I have used a choke before when I made another unusual item, a high frequency AC generator (old model railroad lighting technique, not in use anymore by many), but that is another story.

Again, everyone, thanks!!

Take care, Joe.

[Joe Rampolla]

Re: Ripple Current ?

Hi PCBONEZ,

Sorry I didn't know that a PI filter is a capacitor-input filter. You hit the nail on the head a few days ago. Thanks!

Take care, Joe.

[davmax]

Re: Ripple Current ?

Answering some of your questions.

Basically NO the first cap out of three does not filter more ripple. All are connected in parallel the ripple passing through each cap is determined by the ESR of that cap so it will vary. If two caps are seperated by a resistor or a choke then the first cap passes most ripple.

A PI filter with a choke is not that simple. It needs to be designed to match the expected max load. Chokes in such filters a placed there to resist the rate of change of current (ripple) (acts like a resistor to AC but is a low resistance path for DC) Unfortunately with the wrong choke the DC current can saturate the iron core and the AC resistance goes to near zero. Therefore a choke must be selected that will pass the max DC current without saturating. It is possible to find chokes with a specified DC current. You need to know what max current you will have in the circuit. The caps are not that critical ie higher capacitance = lower ripple, but more contact damage potentential.

Re theory. Unless you really want to get complex your Wiki link and other explanations in this thread should be enough for you. Hopefully.

[davmax]

Re: Ripple Current ?

I should add that the right type of choke is very hard to find. I have looked through catalogs and not found one suitable. For 120hz ripple one millihenry = 0.75 ohms reactance this is not enough. > 100 millihenries is required to obtain > 75 ohms. Such an inductor is quite large requiring many wire turns, these turns will cause saturation so a core air gap is created to remove this problem. The reason you will not find them available is that nobody uses chokes a 120Hz anymore. It is much simpler and smaller to build a simple regulated power supply and get a better result.

The best solution is a plug pack switch mode adapter. Locally a 12Volt 1.25Amp will cost A$29.95. 12V 4 amp A$59.95.

[davmax]

Re: Ripple Current ?

Example of a build your self switcher: http://www.national.com/pf/LM/LM2673.html

This of course requires a step down transformer and rectifier. There is a design wizard on this link. Note that the caps involved are small and the output cap must be low ESR.

[starfury1]

Re: Ripple Current ?

Let me explain the "electronic flywheel" as I use it. When dirty tack prevents electrical contact, the loco stops, so the flywheel cap discharges and gives the motor a chance to coast past the bad section of track and onto a clean section. It works so well that I no longer have to clean track as often and I can even have intentional dead spots and the loco continues forward.

Arh ah that makes sense
basically taking up the slack when there's no juice
like a flywheel supplying energy

Yes Davamx is correct on ESR etc
I was just really trying to keep it conceptually simple in regards to your question
(which does normally result in some bending of the true facts
(also normally ESR with linear supplies is bit neither here nor there but switchers thats another story)

Error Alert
Now I should correct something davmax brought to my attention

when I said battery as "a constant current source." that is NOT correct I should have stated it in context of what I was talking about...quite simply
The battery would be able to supply current anytime unlike the AC situation
(or what I was thinking when I wrote it Fack to Bront as Constant "Source OF Current")

Why? Quite simply the term as used in electronics does "define a mode of operation" and as such does have a "defined meaning"
quote
Constant Current source;
A constant current source maintains the current constant even if the load varies, so the voltage across the load is proportional to the resistance.

unquote

(its a bit of an ohms law stand on your head thing, you tend to think of current as being defined by voltage and resistance but there are circuits that will deliver a constant current (somewhat) regardless of resistance ..used in audio amps for one)

Just so were clear here
Thanks Davmax, quite right to mention that and correct me on that miss-use of terms (battery is a "constant Voltage source")

Joe, I understood this to be
you were trying to do this By having Higher DC voltage on the rails but were looking for a way of dropping this for the CAP in the loco
(due to size constraint)
(assuming here you wanted other loco's whatever working on a higher DC?)

Well I just twigged to one thing
if you use a 16 Volt cap, you cant have any greater across the cap its that simple... therefore as far as that engines motor is concerned the rails may as well be below that
12 Volts say

If you can run the whole thing off a lower DC voltage (less then 16V) then that is it
just build a DC supply that doesn't go higher (well close to)
and since you want to "go close to" it must be a regulated one, linear will work

An unregulated one may just vary too much,
you could as pcbonez said stabilize it a bit by using a resistor across it.
(a little constant loading should keep it a little tighter voltage wise from load to no load...since there is alway a little load there will never be a no load condition)

Yes more cap will reduce ripple, but it wont regulate voltage for you super tight, there will still be fluctuations determined by Mains voltage

Does it have to be variable? DC for throttle control?

Just in theory I wanted no ripple getting to the rails.

Regulated PSU is best in this case

with unregulated psu ripple is really more so going to be determined by your load
(what the others have said still counts thought)

if the motor only pulls a small amount of current few Thousand uf should be ample and fine

Hope this isn't too confusing to understand

short answer for me is build less then 16 Volt regulated supply
(maybe an old "REGULATED" CB 13.8Volt 2~4 Amp PSU)

if the DC has to be variable Then there are DC variable supplies you can get or circuits you can build. (as I posted above)

(On the PI filter don't think its really an option unless build as part of a psu, they do tend to be used in switcher psu's not linear...could be bulky I think for linear PSU but Dont know for sure)

Don't worry about it getting a bit larger then life here, it can sometimes as people explore different aspect of things.

Trust me if they weren't interested in your post they wouldn't have replied so I guess they are enjoying trying to help nut out your query too

cause the thing that making it a bit harder is none of us that have replied
(apart from possibly Mikee like choo choo our resident beer swilling Mod err sorry "SUPER" Mod post above ) are train buffs

just so long as we come up with the right answer for you.

(hope I got this right )

Cheers

[PCBONEZ]

Re: Ripple Current ?

Okay,
I didn't understand that the capacitor was going into the train.
I thought it was on the power supply side of the tracks.

This suggestion is for DC motor versions.
[Someone double check me, been a long time for this stuff.]

You might consider looking into an RC filter vs a PI filter.
http://en.wikipedia.org/wiki/Low-pass_filter

As with anything there are give and takes.
Pros:
It -may- be smaller.
No inductive kick.
It should reduce arcing at the rails.
It will slow the inrush current to the cap when you first apply power.
[The cap will charge slower and so less 'shock' to power supply, and less arcing at track.]
Some wire-would resistors have inductive properties that may be helpful.
[That is very specific to the part you choose.]
Cons:
Depending on how much current you need the size may still be a problem.
They (the resistors) get hot and need air flow (at least vents) and clearance.
Throttle response from a dead stop may be a bit sluggish. (Just Cap affect this too.)
Top speed will be reduced as resistor's value goes up.
(Due to the voltage drop across the resistor. You can compensate by raising the track voltage but make sure the components can handle that.)

You want a small resistance to minimize the wasted power and leave more power for the motor.
You want a large wattage to handle the current.

The optional diode prevents the cap from sending power back to the track.

Essentially a RC low pass filer's job is to remove/reduce all frequencies above a cut-off frequency.
Here is a calculator for the cut-off.
http://www.muzique.com/schem/filter.htm

a 0.50 ohm resistor with 6600uF worth of caps will give a 48.3 Hz Cut-off.
a 0.33 ohm resistor with 9900uF worth of caps will give a 48.7 Hz Cut-off.

VERY IMPORTANT.
How many amps are passing through this thing?
-
The watts of the resistor MUST be adequate or you may cause a fire.
Need to know the max amps to figure that out.

.........

For your other question,
you need an O'scope to actually measure ripple.

.

[starfury1]

Re: Ripple Current ?

Actually may thought was voltage divider
I did think of dropping resistor
(which actually is what you say bonez a low pass filter, that I didn't twig too obvious thought)

I sort of ruled these out cause of possible heat, being probably inside a plastic box

I guess depending on current its doable (thats what you would need to know)

but thinking it though further ...
if the train is in stopped mode, motor doesn't pull current therefore No Vdrop across resistor, capacitor basically *open circuit so therefore Voltage across capacitor becomes the track Voltage DC.
(*except for leakage current)
If that is higher then 16V.... bye bye cap (I think)

think thats when I thought voltage divider
voltage across cap is somewhat fixed at a set ratio
but you also under no track voltage have the resistor (across cap)
acting as a load like the motor.

the tricky part is getting the top resistor to drop the voltage enough under all cases (so voltage across cap never exceeds less then 16V)
but provide enough current for the motor.

the other side is getting the second resistor
(the one that has the cap and motor in parallel on it)
to not bleed to much current of the cap

So I think you could get that to work but heat of top resistor my blow it out of the water
the other thing (and the bit I think would be hard to get a proper balance with)
is each resistor needs to be in opposite directions ...one you want to pull some reasonable size current through (top one)
bottom one you don't want much at all going through it.

If it was a fixed current it probably wouldn't be that hard,
but having a max & 0 point (and all points between)
its would be tricky...maybe the thing Kirchoff's voltage and current laws are made of
(anyone good with maths out there?)

you would need to know what is the max DC voltage on the track is and what the max current of the motor to start with

But if The tracks can be powered by less then 16 Volt (Varying DC) there is no issue.

This is all based on if you are varying the DC track voltage for speed control ? and that it must be greater then (<) 16 Volts
let me go and reread his posts

Anyway hope that makes sense...maybe I am totally barking up the wrong tree here

Cheers

[Joe Rampolla]

Re: Ripple Current ?

Hi Folks,

Thanks, everyone, for all the great information and attention to my questions. This has been very helpful to me, but a good bit of it is beyond my knowledge, as you probably have gathered by now.

I am going to look over all the new links and see what I can do.

The model train hobby is so technical now that it might be the right combination of fun and technical stuff for all of you here on this forum. I like the hobby for stress relief! I suggest that many of you stop into a hobby shop and look at the model train stuff available now. I think you will all be surprised at how far the hobby has come!

But I use the ideas that were abandoned back in the 1970s. For me, the capacitor is magic, solving many of my problems. Unfortunately I had to figure out a lot of stuff for myself and did some basic testing. The first time I used the cap flywheel idea, I did all the figuring on my own. I soldered the cap in parallel with the motor, applied current to rails by a remotely controlled outlet, standing about 20 feet away in case it would pop, wearing goggles. I watched the little HO loco run so smoothly that I couldn't believe my eyes! I turned off power, pulled the shell off of the loco, and felt the capacitor to see if was warm. It was very cool. I was using a cap rated at 2 - 3 times the max track voltage. So that is how I apply these concepts. I take the simple theory and apply it to something I can see and feel, a good combination of thinking in the abstract and applying it to something real. But there are limits to how far someone can go without the formal training.

Thanks, again, everyone!!!!!

Take care, Joe.