DC pulsing

[artv]

Hi all ,  I was wondering if this is how the dc out is supposed to behave?
As the mill starts to turn the lights start to flicker, as the mill speed increases the flicker increases and so does brightness. Is this proper behaviour??
Also found that by putting two big caps paralelled, stops the flicker but also takes away some of the brightness.
Anyway just wondered if this how it's supposed to work.   thanks .....artv

[madlabs]

Shooting in the dark here but...

Sounds like you have a light connected to the mill and nothing else. If so, yeah, what you are seeing sounds about right. If you have any other configuration, that probably isn't right. Can ya tell us exactly what you have going on?

Jonathan

[ghurd]

LEDs?  And a single phase PMA?
Usually my experience with light bulbs or 3 phase do not show flicker.  Incandescent bulbs do not react fast enough, and 3-ph is pretty smooth.

AC from the PMA comes out of the bridge as sort of pulses.
http://en.wikipedia.org/wiki/File:Gratz.rectifier.en.svg

The cap smoothes out the bumps.
The bright points of the flicker are sort of used to fill in the low points making the brightness lower with less flickering.
http://en.wikipedia.org/wiki/File:RC_Filter.png

[artv]

Hi thanks for the replies , It is three phase ,4 diodes per phase, (2 series diodes paralelled with 2 series diodes).The ouputs of the phase connected in the middle of the series diodes ,and a 10v -1000uf cap hooked with the positive on one side of the phase connection  (middle of the series diodes), and the negative of the cap hooked to the negative dc output of the diodes.
Each phase is hooked this way , then the two outputs from each phase are paralelled. Leaving two lines coming down the tower, dc out.
They are leds ,and when I hook my meter to the two leads coming from the tower ,without any thing else connected the voltage rises and falls with the speed of the blades. When I hook up the battery ,the voltage seems  to get locked down to near battery voltage, then when I hook up the leds ( 11 of them) the voltage drops again. It is only a small mill  and am using a 6 volt battery (from a cordless framing nailer "pasload").
The battery does make the flickering stop.
My question now is : with the battery and leds connected ,if the readings don't go above the rated battery voltage, it's not charging???..........still a ton of stuff to learn..........thanks for all the help.....artv   

[ghurd]

It is three phase ,4 diodes per phase, (2 series diodes paralelled with 2 series diodes).The ouputs of the phase connected in the middle of the series diodes ,and a 10v -1000uf cap hooked with the positive on one side of the phase connection  (middle of the series diodes), and the negative of the cap hooked to the negative dc output of the diodes.Each phase is hooked this way , then the two outputs from each phase are paralelled. Leaving two lines coming down the tower, dc out.

You lost me on that one.
It does not sound good though.
Especially if it is still wired star/wye.

They are leds ,and when I hook my meter to the two leads coming from the tower ,without any thing else connected the voltage rises and falls with the speed of the blades. When I hook up the battery ,the voltage seems  to get locked down to near battery voltage, then when I hook up the leds ( 11 of them) the voltage drops again. It is only a small mill  and am using a 6 volt battery (from a cordless framing nailer "pasload").
The battery does make the flickering stop.
My question now is : with the battery and leds connected ,if the readings don't go above the rated battery voltage, it's not charging???..........still a ton of stuff to learn..........thanks for all the help.....artv   

I will join Jonathan on his limb.
Uh... Yes?

The turbine voltage is clamped to the battery voltage.
The turbine voltage is the same as the battery voltage, plus wire loss and diode drop.  Meaning they are about the same while charging.  You will not see the mill at 20V and the battery at 13.0V while they are connected together (unless you have done something very very wrong).

Because the battery and turbine are at the same volts, batteries (in a decent system) do not change voltage significantly fast to noticabley 'continiously flicker the lights', the LEDs see the same voltage which is relatively constant.



If you want to know if it is charging the battery, 2 options:

1)  Connect an ammeter between the battery and EVERYTHING else.
Read the manual for the meter, or a similar meter, about properly connecting it to read amps.
If amps are going into the battery it is charging.
If amps are leaving the battery it is draining.
And the amps can change as you watch during a good strong long burst of wind.

2)  Connect a volt meter to the battery.
If the voltage is increasing, the battery is charging.
If the volts are decreasing, the battery is draining.
A typical cheapie meter set to 20V is enough to tell.

"am using a 6 volt battery (from a cordless framing nailer "pasload")"
Is it MORE than 10 years old?
Guessing it is more than 20, and nicd.
Guessing it is shot.
Bad batteries will just confuse the whole thing.   Get a new battery.  Sounds like 4 AA nicd / nimh in a holder will do what you want.  Can get a 4 pack of ~2500mah nimh AAs at Big Lots for $6, and a 4 AA holder at RatShak for $3?

Do NOT throw nicds in the trash.  The CD in niCD is very evil stuff.  Take it to Office Max / Radio Shack / Sears and throw it in the recycleable battery container.

Glad to hear you have something flying.
G-

[artv]

Hi ....it's flying but it aint perty ...lol...I knew that was going to be confusing. I don't have it wired star or delta, ( similar to Jerry)The start and end of each phase (6 wires) 2 for each phase, each phase goes to it's own set of diodes . The start wire of phase 1 goes to the middle of two diodes in series, the end wire of phase 1 goes to the middle of two diodes in series, these two sets of diodes are paralelled.
 Now take a cap hook the positive  to the end of the phase, not the start, and hook the neg lead to the neg. of the diodes(-dc out)..................I hope thats more clear......

Method 2 works and it is charging,.....I don't think I have an ammeter......as for throwing stuff away ...I keep everything....never know when it might come in handy............artv

[ghurd]

Send junk batteries for recycling because you will never have a use for them.

Get an ammeter or VOM.
Even the $3 HF multimeter is fine for most stuff, and usually matches my $350 meter within a few %.

( similar to Jerry)The start and end of each phase (6 wires) 2 for each phase, each phase goes to it's own set of diodes . The start wire of phase 1 goes to the middle of two diodes in series, the end wire of phase 1 goes to the middle of two diodes in series, these two sets of diodes are paralelled.

 Now take a cap hook the positive  to the end of the phase, not the start, and hook the neg lead to the neg. of the diodes(-dc out)..................I hope thats more clear......

I am out on that limb, looking back at Jonathan.

Not sure what you mean.
Guessing from the discription the cap is a polarized electrolytic.  They are (basically) a short circuit when connected backwards.  The PMA makes AC.  The electrolytic cap is connected backwards half the time, so it is a short half the time.

Get the cap out of there.  It is causing more issues than anyone but Einstein can explain.
G-

[Madscientist267]

fire... FIRE!!! 

You got a schematic for this?

That would make things easier for us to see. Or even at least a picture.

It almost sounds like its IRP with full wave bridges on each. If they're wired right.

With the cap across the output of the bridge. If I understand correctly.

What isn't clear at all is how the rectified phases are interconnected.

If the above is all correct, the DC+ and DC- of each bridge should all be connected together.

If it's 3 phase, there should then be little pulsating, with the exception of varying wind (and therefore prop) speed.

And since you're going into LEDs, a high enough wind gust, and they will pop.

Hence the need for a battery. At the very least.

Am I close?

Upload a picture of your work so we can tell. Then we gotta work on the terminology. 

Steve

[JW]

Isnt "Artv" a senior member from the old forum? Doesnt seem right only 200 or so posts. I remember from the merge, if you remembered your original password, the account was automatically carried over. However there is no way to recover the original password, from the old forum.

JW

[ghurd]

With the cap across the output of the bridge. If I understand correctly.

Not the output?
In a 3-ph that wouldn't do much good.

"a 10v -1000uf cap hooked with the positive on one side of the phase connection  (middle of the series diodes), and the negative of the cap hooked to the negative dc output of the diodes"

I don't get it either.  Sounds like a short... half the time.  Most of the time?

Art,
Not sure anybody quite understands your verbal explainations. 
Sketch it out in MS Paint (comes on Windows PCs from the factory), save it as a gif or jpg, post it.
Or draw it on paper, take a low res pic of the drawing, post it.

[Madscientist267]

Agreed, the cap is pointless for 3ph, but just going on what he's given, if it's there, that's where it goes...

Missed the part about 'middle of the series diodes' with the cap. One leg on the AC side? Definitely doesn't make sense to have it there.

Definitely need a visual on this one.

Steve

[JW]

Method 2 works and it is charging,.....

Send junk batteries for recycling because you will never have a use for them.

For some reason I have the urge to mention ETDA... but yes, if there "puffed-up" ther plates are too degrated.

JW 

[wooferhound]

I think artv is describing this . . .

[Madscientist267]

Possibly.

Would make the most sense.

The caps would need to be a half way decent size to provide useful output (or at least efficient output).

Figuring on a couple hundred Hz at each phase, then effective DC filtering of the caps combined being 'sticker' divided by 3 (assuming they are equal)...

Somewhere in the 2200-4700 uF range? Just a quick crunch in my head, didn't really do any math...

Phew, now I understand where Amanda was coming from with the whole schematic thing!

I'm sorry guys! LOL

Good call woof!

Steve

[joestue]

one amp makes 1 volt per millisecond per 1000 uF

if you can see the lights pulsing, caps will need to measured in kilograms, not micro farads.

[artv]

O.K. ....this is my fourth attempt ,I think  it will work......if the diagram shows up........do this with each phase then  paralell the dc outputs....to end up with two leads + and -

JW ,I joined this site last year was a guest six months prior.....

Woofer ...I tried it wired that way but found this design gave a little better output

Joe ...do you mean physically weigh the cap if so they are 2.18grams each.........I just wondering if I should have used bigger caps.....
thanks for all the replies....artv

[Madscientist267]

LOL @ joestue

"Don't feed the humans"  

artv -

Ok, that's what I was originally thinking based on what you had said. The cap is in the wrong place. It isn't doing you any good there.

The entire thing should look more like what woof posted. I can't see how you're gaining anything (in terms of power) by having a single cap across a single diode in a bridge. There are occasions where caps get put across diodes in a bridge, but they're not used for what you're using them for. So forget that for now...

If you're not seeing an improvement by having the caps across the DC side of the bridges (as in woof's diagram), then something else is wrong:

#1 - The cap(s) are bad

#b - The cap(s) don't have a high enough capacity

#3 - One (or more) of the diodes in the bridge(s) are bad, either open or shorted

#D - I'm sure there are other possibilities, they just don't come to mind at the moment

#4 v. 2.0 - Only the capacitor's value (in uF) is important in terms of what it can do for you in the circuit... Joe was kidding - the physical weight of the cap doesn't mean anything. He just meant that they would need to be very large (in terms of capacity), and therefore would weigh a lot as a result.

Also, you've shown us one phase, but now we need the rest of the picture. How are the phases connected to each other?

Steve

[Flux]

If you are getting better results with that circuit it is because your cut in speed is too high for the parallel Jerry ( IRP) connection.

You are adding the phase voltages in series as a sort of voltage multiplier and you are finding the limitation of all these voltage multiplier circuits. The capacitance required to maintain a few amps during the interval when each bridge is not conducting is enormous. This arrangement works fairly well for small machines with limited output but the capacitance values get out of hand for higher power.

If you are hopelessly fast at cut in then this method is your only option and at the small power level you are using it should work fairly well.

If you got some sort of result with the normal Jerry connection, then you would certainly do better with star and a 3 phase bridge as it will give you 1.7 times the basic delta ( Jerry) voltage.

For higher powers the less effective voltage match from star may be better than the series Jerry connection unless you can afford monster capacitors.Jerry was into car audio and the odd Farad knocking around was likely common.

Flux

[fabricator]

Why not just go to radio shack and buy a five dollar bridge rectifier? Why make it one thousand times harder than it has to be? It's like the old computer programmer term GIGO, garbage in = garbage out.
Without a decent battery you are never going to learn anything meaningful.

[JW]

Without a decent battery you are never going to learn anything meaningful

I would like a complete reason for this statement.

Ghurd touched on this, I skimmed thru the comments and I can't be sure that the battery is "expected to be sulfated".

Im just trying to feel my way around in the dark, with this...

I think that "all" the comments are somewhat on target.

JW

[fabricator]

Basically I meant if the mill is clamping to the voltage of a sulfated dead battery he's never gonna know what the thing is capable of.

[Madscientist267]

From what I can tell, I don't even think it is a lead acid battery...

One thing is for sure, this is a cluster**** of speculation. 

Need Input! Number Johnny 5 Alive!

 

Steve

[artv]

The battery is a 6 volt NiCad..........when I charge it with the wall charger , it comes out fully charged and reads 6.8 volts. I can hook it up to the leds and it runs down to 3.28 volts over a couple of hours. Then I hook it to the mill as long as the wind is blowing it does charge  it back up.
That circut diagram I posted earlier does give better output than star ,delta ,and Jerry, but maybe only because I'm working with such small values.
It is one circut per phase ,then paralell the outputs from each circut, to end up with +dc out and -dc out. I don't know if it works with bigger machines .When I build one I will test it and see if it does...........artv

[Madscientist267]

Okie now we're getting somewhere...

LEDs drain NiCds in a couple of hours?

How many (and of what type) LEDs are we talking about here?

It still sounds like the battery is crapping out (which you mentioned it is old, and old NiCds have a tendency to have issues), but that is kinda beside the point.

There are bigger issues at hand. Like why you get a better response out of having one diode in a bridge bypassed with a cap...

Also, final wiring (meaning what you do with the rectified DC from each phase) will play a significant role in how much you can get out of this.

Woof's diagram gives the highest possible voltage (at least OTV), because the caps all charge independently and the voltages are then summed together. The math gets complex under a load however, because since they are in series, the whole 'weakest link' thing applies. They all have to have equal output individually, and the caps need to be identical, and substantial to keep them from discharging to a useless level when under a load.

With the outputs from the bridges in parallel, you will have the highest possible (EDIT - continuous) current, at the lowest voltage. The caps don't need to be as big (or there at all for that matter if you're charging a battery) since they don't need to support each other to complete the circuit. The drawback to wiring this way is that your cut-in for the mill will be very high, and so it will take a substantial amount of wind to see anything at all from it. But when it produces, it will give up the goods. The trick is getting it spinning fast enough to do so.

Delta and Wye are enigmas to someone not familiar with how the phases interrelate due to the connections and phase angles. Unfortunately, I can't provide a whole lot of input in this area simply because I only partially understand them myself. While I get it enough to make use of them if I needed to for myself, I couldn't really explain it entirely to someone else. That's ok, there are plenty here that have their head wrapped all the way around it, and can help with that.

But I know enough to say that since you still haven't told us exactly how you have the phases connected to each other, we have little to go on.

Steve

[ghurd]

I would like a complete reason for this statement.

Ghurd touched on this, I skimmed thru the comments and I can't be sure that the battery is "expected to be sulfated".

Artv said "and am using a 6 volt battery (from a cordless framing nailer "pasload")."
To me, that sounded like a Very old nicd pack,
which he later confirmed in #22 with "The battery is a 6 volt NiCad".

When is the last time you saw a new tool with a 6V nicd pack?  7.2V?  9.6V?  12.0V?  14.4?
nicds get funky with age.
A 5 cell nicd pack less than that old can act like a cell has a diode over a cell while charging or discharging.  It can act like it has a diode in series with the cells. A cell can actually reverse polarity.  A pack can act like it is charging correctly, but discharge like it is 1% of the rating, or vice versa.
(no, I did not log it with anything fancy)

nicds can get get exrta funky with age to the point of not really knowing what is going on.
G-

[Rover]

G, Agreeing with everything said about NiCads, but one little off topic point

"When is the last time you saw a new tool with a 6V nicd pack?  7.2V?  9.6V?  12.0V?  14.4?"

He said this came out of a Pasload framing nailer, its a gas powered nailer. The charge is used to ignite the gas. It may not be an old battery, and its not used to set the nail.

However, using a NiCad for testing .. hmmm probably not the best idea.

[ghurd]

Rover, I figured they had a Piezo sparker in there.

One other potential problem with nicds in this situation.
The "small mill".

When charging with a low current compared to battery capacity, they get up to about 1.3V per cell (6.5V in this case), and sort of sit there.
They sit there while charging, and they sit there when charging stops.

Usually when doing simple testing with a lead acid battery, a volt meter on the battery terminals is easiest.
A couple hundred ma into a 85AH 12V is enough to watch the voltage change by 0.01V.
100ma into a 3AH nicd is not going to change the voltage very fast.

If making a sketch in Paint, make sure to 'save as' a gif or jpg.
bmp files are huge in comparision.
Woof's gif sketch is 5K.  The bmp sketch is 108,050K.  21 thousand times bigger!
If you made a new sketch and it would not upload, that is probably why.
G-

[Bruce S]

artv;
Agreeing with G- and others on the NiCd battery pack.
As it is a 6Vdc pack and it's at 6.8Vdc when fully charged , this is right for the 5-cell battery type.
The part the bothers me is where it drop to 3.xx volts after being on a LED light after a few hours.
Once removed from the light it should bounce right back up to say about 6Vdc. The fact that it is dropping that low also tell me one of the cells is toast or about to be.
going below 0.9Vdc on a NiCd battery is murder on them and can cause a cell reversal.

However using the battery as a test bank for your mill is perfectly okay as it will respond nicely.
This is one of the reason I use multiple sizes of NiCd battery banks, their ability to take rough usage and bounce back.

A word of caution: NiCds will vent and burst if you try to charge them too high whether it's high voltage or high current. Check the cells external temps, if they start to feel more and warm, STOP, IF you see venting on the tops STOP. The vented stuff isn't cadmium so not too worry about that, but not good for the batts either.

Cheers;
Bruce S
 

[artv]

G-..thanks for the tips on posting pics,I had to keep shrinking it to get within the allowed size, next time will do saving as( jpeg or gif).
Bruce S ..not to be concerned about over charge the best I've seen out of this thing is ~20volts dc, but only for a few seconds, in a big gust of wind.
I'm 30 feet in the air, the trees all around are 40-60 feet tall ~70 feet away,and when they get their leaves all I'll get is turbulance.
Steve ...I've already explained how the phases are hooked up twice now , maybe I'm not being clear enough.
Make the circut diagram like I posted ,hook one phase to this circut, now make two more circuts for the other two phases,take the output from the three circuts and paralell them to end up with two outputs........thats as clear as I can explain it.
Jerry gave better out than either star or delta......this was a little better than Jerry ,but maybe only because the #'s are so small...........thanks.........artv

[Madscientist267]

paralell them to end up with two outputs

Don't you mean one output?

Positive and Negative?

We can't help you if we don't know what you're talking about. That's why the schematics are important.

Verbal explanations aren't working in this case.

Does it look like this?



I didn't draw the caps in...

Steve

[artv]

Yes Steve that is it ,but like you said you didn't draw in the caps. I tried that arrangement ,then added the cap and the output increased by .15volts dc. I was testing at 100rpm...with 11 leds connected..........your drawing..3.4vdc......add the cap ..3.55vdc......small increase ,but still an increase just the same.....the drawing Woofer showed..........3.45vdc ........the day I was trying this I tried probably 20 different arrangements..........the pic I posted was the best output I could get...........artv