n-c alternator modifications: discussion and testing

[wcorey]

Did some casual, practical testing somewhat in context with Bevel bob's situation.

Sec I and III (more or less equivalent to stock alt but with the coils rerouted to the 'new series' arrangement), each with independent full wave rectification, outputs combined and connected directly to a 12v battery, 55w (rated) headlight to positive on battery then through the current meter to negative bat terminal, all run at 3450 rpm.

Battery unconnected, 12.4vdc
Battery loaded with 55w light, 3.82a, 12.0vdc
Battery loaded with 55w light, alternator running, 3.95a, 12.66vdc

I left the light on the battery (no alternator running) for about a half hour to let it rundown to 11.5v.
Disconnecting the light let the battery voltage rise and settle to 12.4v.
Running the alternator with the light reconnected, the voltage went down to 12.2 and gradually over about ten minutes of running, up to 12.42v, were I got tired of babysitting it (and shut it down).
The alternator temp seemed to stabilize at about 30 degrees above ambient.

A half hour later the battery had settled to 12.3v.
Ran the alternator again without the light on the battery, almost immediately went up to 13v, and after about half a minute seemed to stabilize at 13.33v.
Battery read 12.74v with alternator stopped and half hour later had settled to 12.45v. I'm assuming it would have eventually taken a full charge...


Bill

[DewCatTea-Bob]

" Did some casual, practical testing somewhat in context with Bevel bob's situation. "

____ Right Bill, "somewhat in context" is thoughtful added wording to let Bob know that his-own results may (likely) differ.
____ I'm fairly pleased that you thought to do this kind of testing before I got-around to suggesting it myself, Bill !
In fact actually, I had meant to mention in my last reply-post (to your last post), what the next testing should be that ought to get done,, and if I had remembered to do so, then I'd now be assuming that I had inspired your latest post, (so it's just as well that I had forgotten again).
__ I've been waiting for someone to ask about some things which really ought seem to be of concern...
Like: "What about the spools of wire which Bill has been using for his 'loads'? - Are they truly providing the same results as could be expected from my bike's system-load? "
And the answer to that is: 'No, not quite!' - As those wire-COILs (he's been using as test-loads) have more than merely just their 'resistance' for the pulsating-DC to have to contend with. _ (But if those coils of wire were stretched-out into non-coiled lengths, THEN they would be just pure 'resistance' only and thus just like a normal system-load.) _ However, the coiled-type test-loads are fine enough for all the comparison-type testing which has thus-far been completed.
And: "What about the absence of a battery from these tests? - Does that make any difference?"
And the answer to that is: 'YES, a battery makes a big difference!' - Cuz batteries have characteristics which change the state of electrical-power in substantial ways. _ For instance, with a battery (& it's pure DC) in the circuit, it would then make it irrelevant that Bill's test-loads are in coiled-form !
____ So while the test-stator is still set-up for what we've determined is in optimum-arrangement, tests should be done with a battery (in good state) connected into/with the load-system.
But to do this testing (with battery added) appropriately, then either an ammeter or two amp-meters need to be properly connected into the test-circuit.



" Sec I and III (more or less equivalent to stock alt but with the coils rerouted to the 'new series' arrangement), each with independent full wave rectification, outputs combined and connected directly to a 12v battery, 55w (rated) headlight to positive on battery then through the current meter to negative bat terminal, all run at 3450 rpm. "

____ I Believe that you had determined that 'se.I' was slightly stronger than 'se.II', therefore since 'se.III' & 'se.I' are more the same, that test-outcome may be slightly greater than 'se.I' with 'se.II' (as Bob & everyone-else has). _ And-also to be more like the stock arrangement, the two stator-sections should've been connected in 'series', instead of parallel (as you evidently must've had them in order to full-wave rectify each section independently).
This type of dual full-wave rectification-circuit is very much similar to the charging-system circuit-plan which I've been recommending to Ducati-owners since the mid.70s .
I'll post a scheme-diagram of this one which you've just tested, down below.


" Battery loaded with 55w light, 3.82a, 12.0vdc "

____ This makes me question the 'state' of the employed battery, cuz a 55w.light ought draw 4.5 amps, (not just "3.82"). _ The 3.82s & 12.0v indicates that the light is only getting/consuming 46-watts.


" Battery loaded with 55w light, alternator running, 3.95a, 12.66vdc "

____ Now with the charging-circuit's assistance, the light is next getting 50.0-watts, (still 5w shy).
This now makes me question the light's true 'wattage' (or that it's really designed to operate at just 12.0-volts), since the circuit-voltage was holding-up near to 12.7-volts.
In order for that light to become the full 55-watts, it would require 13.1-volts (with 4.2-amps) !


" I left the light on the battery (no alternator running) for about a half hour to let it rundown to 11.5v.
Disconnecting the light let the battery voltage rise and settle to 12.4v.
Running the alternator with the light reconnected, the voltage went down to 12.2 and gradually over about ten minutes of running, up to 12.42v, "

____ This leads me to think that the battery is in relatively good-shape. _ What's it's 'amp/hour-rating' ?
__ Even without an ammeter, the results of this testing makes it clear that the breakeven-point is being surpassed (at the test-RPM) !
And this is with just two of the three sections !?


" A half hour later the battery had settled to 12.3v. "

____ If the battery can set ignored for a couple of days and still hold at least 12.0-volts with a modest load (like a brake-light) then connected, I'd then assume that it's in good enough shape to trust for use with further experimental-testing.


" Ran the alternator again without the light on the battery, almost immediately went up to 13v, and after about half a minute seemed to stabilize at 13.33v. "

____ Under this condition (being kept at over 12.8-volts [for over a couple hours]), a voltage-regulator would be called-for, in order to keep the battery-plates from over-heating (and boiling-off the water-level). _ However if a connected system-load could use-up most of the excessive power-juice, then the battery could handle what's left, without any need for a regulator !


" Battery read 12.74v with alternator stopped and half hour later had settled to 12.45v. I'm assuming it would have eventually taken a full charge... "

____ At that rate, it most certainly was in the process of taking-on a full-charge, and would fairly soon even become over-charged !
An ammeter is certainly called-for, in order to check the rate of charging-current.

____ Thanks for the thoughtful real-world type charging-circuit testing Bill.
__ I've now included a schematic-diagram of the test-circuit which I believe you've described. _ Please let me know if it looks wrong in any way.


Content-Cheers,
-Bob

[wcorey]

__ I've now included a schematic-diagram of the test-circuit which I believe you've described. _ Please let me know if it looks wrong in any way.

Yes, that's it (without the light).
Battery is 7.2ah, one of those typical gel types used in emergency lights and such.
I noticed that I've been using the low beam ( high beam is fried) on the (sealed beam) headlight which I bet is 50w rather than 55w. On the DOT H4 bulbs it's usually 50/55w...

Tried something else, combining outputs of the three sections onto the input side of one rectifier. I seemed to recall someone advising against it but couldn't recall the reason, figured it wouldn't hurt to try it for a brief time.
Well, it appears to work much the same as with triple rectifiers, 3450 rpm, 1ohm, 8.05a, 8.51vdc, 68.5w, stayed at about 25 degrees above ambient.

Battery only, 5.37a, 13.32vdc, 71.5w
Battery with light, 5.05a, 13.91vdc, 70.2w
Light only (I'm assuming this is due to a drop in impedance?), 3.25a, 11.56, 37.57w

Two sections into one rec, 3450 rpm

Battery only, 3.56a, 13.66vdc, 48.62w
Battery with light, 3.85a, 13.02vdc, 50.12w
Light only, 3.21a, 10.86vdc, 34.86w

One thing I don't understand at all is that when I remove the alt/light connection from the bat neg terminal and run the light without the battery (when looking at Bob's drawing the light is connected in the same place the volt meter is), the voltage goes down but the light gets brighter, makes me wonder if I somehow have something connected wrong...

The Battery certainly seems to be the Magic Wand, which is funny to me because that would've been my next logical step (back when I first got into this thread and was using only a headlight for a load) if it weren't for all the impedance/load testing requests. I guess it could be that the convoluted path we've taken is what has gotten us to this point and a straighter route may have passed by some important findings. I hope some of the plethora of misc. load testing (and other) data proves useful to someone someday, in any case I know I've surely learned a lot...


Bill

[DewCatTea-Bob]

" Battery is 7.2ah, one of those typical gel types used in emergency lights and such. "

____ I'm not familiar with the characteristics of those type of batteries, so any stumbled-upon abnormality could be hard to pin-down it's exact cause.
Do you have any regular lead-acid type 12v.batteries (for test purposes) ?



" I noticed that I've been using the low beam ( high beam is fried) on the (sealed beam) headlight which I bet is 50w rather than 55w. On the DOT H4 bulbs it's usually 50/55w... "

____ I can't say for sure anymore these-days, but as I recall, the quartz/H4-bulbs were 55w-low/60w-high !
__ I'd say a good test to try would be (after making sure that your battery is fully charged-up), to check the current-draw & voltage-drop with your light directly connected to just the battery alone.
Then do the same test with a lead-acid battery (of which it's state of condition is relatively equal).



" Tried something else, combining outputs of the three sections onto the input side of one rectifier. I seemed to recall someone advising against it but couldn't recall the reason, "

____ Okay, seems you decided to connect-up all three stator-sections (directly) in parallel and THEN connect their combined outputs to just one FW.bridge-block. ...
Providing that all three outputs have their polarities kept properly matched-up, then the only other concern would be that they all three individually happen to produce the EXACT same amount of voltage (which is quite unlikely). _ If there's any imbalance, (which IS likely), then the sections with the least voltage will become as partial loads to the section which happens to produce the highest voltage. _ Thus at least one section will waste a portion of the power produced by one or both of the other two.
Your tests have shown that the sections are indeed slightly miss-balanced, so there's sure to be some amount of waste, however little it might actually be.



" figured it wouldn't hurt to try it for a brief time. "

____ If the miss-balance were to be substantial, excessive heating and more significant unnecessary energy-waste would result.
The extra diodes will pay for themselves.



" Battery only, 5.37a, 13.32vdc, 71.5w "
Battery with light, 5.05a, 13.91vdc, 70.2w "

____ This needs clarification. _ Just what is the circuit ?
Cuz it doesn't make sense that the battery alone draws "5.37" amps,, yet with the light added, both together only draw just "5.05" amps. _ (Even if those figures were interchanged, things still don't add-up,, so there's got to be an associated circuit which escapes me.)
Exactly where is your amp-meter located in the circuit ?



" Light only (I'm assuming this is due to a drop in impedance?), 3.25a, 11.56, 37.57w "

____ I don't get what you're referring-to about due-to impedance !?
We need an explanation of the associated circuit, but if it's what I assume it to be,
then it makes normal sense that the light would not be able to draw as much current, without the battery's added supply of juice (to fill-in the nulls left by the pulsating-DC).



" Two sections into one rec, 3450 rpm
Battery only, 3.56a, 13.66vdc, 48.62w
Battery with light, 3.85a, 13.02vdc, 50.12w "

____ This pair of test-outcome result-figures seems to make better sense (relative to each other), but still need to know exactly where the amp-meter is located in the circuit, (in order to be sure of what's actually going on).



" One thing I don't understand at all is that when I remove the alt/light connection from the bat neg terminal and run the light without the battery,
the voltage goes down but the light gets brighter, makes me wonder if I somehow have something connected wrong... "

____ It does make logical sense that the light would get brighter, because it no longer has to share the 'POWER' provided by the alternator, with the (power draining) load presented by the battery (as it passes the current which tends to charge it up).
This was just an example where 'voltage' is irrelevant compared to 'power' !



" The Battery certainly seems to be the Magic Wand, which is funny to me because that would've been my next logical step (back when I first got into this thread and was using only a headlight for a load) if it weren't for all the impedance/load testing requests. I guess it could be that the convoluted path we've taken is what has gotten us to this point and a straighter route may have passed by some important findings. "

____ All fairly quite true.
A battery has effects which both mask & review facts about electrical-circuits. _ And that's the reasoning why electrical-systems with & without batteries should not be treated as the same !
Hopefully, you'll be able to review still more significant differences between test-results with & without a battery.


Content-Cheers,
-Bob

[wcorey]

Do you have any regular lead-acid type 12v.batteries (for test purposes) ?

I actually don't own a standard lead acid bike battery and the only good AGM battery I have is in the bike that's on the road and is a pain to get in and out, there's another functional one in the 996 but it's a bit weak.

____ This needs clarification. _ Just what is the circuit ?
Cuz it doesn't make sense that the battery alone draws "5.37" amps,, yet with the light added, both together only draw just "5.05" amps. _ (Even if those figures were interchanged, things still don't add-up,, so there's got to be an associated circuit which escapes me.)
Exactly where is your amp-meter located in the circuit ?

More or less what your drawing is, I didn't take pictures so I can't go back and figure it out.

" Light only (I'm assuming this is due to a drop in impedance?), 3.25a, 11.56, 37.57w "

____ I don't get what you're referring to about due to impedance !?

The light only is probably close to a half ohm and outside of the ideal working impedance, whereas with the battery seems to be closer.



Bill

[wcorey]

I've been playing around with 'real world' setups again, unfortunately not very methodically, just throwing some things together and seeing what happens (back to my old ways, lol).

I had a couple 10kuf 34v caps wired in parallel, they seem to behave very similarly to the battery when used in it's place. Made the light brighter...

Still running all three sections with combined outputs, with the caps and no battery, used a typical newer type Shindengen 12v reg/rec instead of the plain rec and added a 5ohm single Dyna coil. Not sure how to simulate how the coil would behave (what current it would draw) while operating without making something to trigger an ignition at the appropriate rpm, I assume it would draw less at lower rpm. I tried it briefly 'at full power' and it pulled the voltage down to from the mid 14's (with the light only) to mid 11's and current up to high 6's, I think it added up to about 70w. I ran the motor at 4300 rpm and got something close to 6.9a, 13.7vdc, 94.5w.

So that's pretty close to a functional system that would operate a bike and it will fully charge a 12v battery at somewhere around 4krpm...

Sorry for the lack of details but it was either write it up like this or not at all...

Bill

[DewCatTea-Bob]

By: wcorey...
" there's another functional one in the 996 but it's a bit weak. "

____ That may be adequate, and it seems you ought to get it out and perform maintenance on it and get it checked-out anyhow, (but I'm not really trying to coax you to do so).
__ I expect that the battery you've been testing with is capable of handing loads as well as a std.lead-acid type, but those amperage-draw figures you've stated strike me as being rather high, in comparison. _ Could be that that type of battery act's more like a load, which is not something you'd really want to have added to your intended load-system !
Unlike a condenser, storage batteries really don't actually 'store' the same power-juice which they've taken-in, rather they use the power-juice which passes through them to perform a chemical reversal of the chemical-action which lets the battery discharge.
During the process, the ratio of the charging-juice that's effectively saved compared to that which is straight-through wasted, varies somewhat,, and I've since been wondering whether that ratio is any less desirable with the type of battery you've been testing with.
If you intend to use that gel-battery on your 450, and it indeed does waste a greater amount of juice than it effectively saves, then we ought to consider circuit-mods which will prevent it from drawing so much current (which could be better used, diverted to your intended load-system).

Just what is the circuit ?
Exactly where is your amp-meter located in the circuit ?

"More or less what your drawing is, "

____ I've been meaning to draw-up another/better related drawing. _ Then you can point-out which is closest to whatever test-circuit you had set up.

Light only (I'm assuming this is due to a drop in impedance?), 3.25a, 11.56, 37.57w

____ I don't get what you're referring to about due to impedance !?

" The light only is probably close to a half ohm and outside of the ideal working impedance, whereas with the battery seems to be closer. "

____ Okay, I now gather that you were meaning that the poor-power result could be due to the light-bulb's resistance being lower than the working-impedance of your tested power-circuit.
Well if the light-bulb's filament retained the same resistance as when it's cold, (which I suppose is close to being same as a 'short-circuit'),, then you'd be right to consider the possibility which you had. _ However, when the light-filament heats-up, it's resistance increases, (thus becoming greater than the impedance mismatch you were guessing might be involved). _ In fact, considering your given amperage & voltage figures, the light-bulb's resistance works-out to be 3.557-ohms.


Hopeful-Cheers,
-Bob

[Bevel bob]

When I tested my system I found that at 3000 revs the voltage at the batt terminals was 14.4 volts with a 35 watt hl on. The lead/acid battery was part charged 12.5v unloaded, 12v ign on, Voltage went from 14.5 to 14.4 when light switched on.I will try it with a higher watt bulb.Rotor seemed to have good magnetism,battery probably 8 years old at least,but has been well maintained.I usually run at higher revs so it seems the regulator is earning its keep.

[DewCatTea-Bob]

When I tested my system I found that at 3000 revs the voltage at the batt terminals was 14.4 volts with a 35 watt hl on. The lead/acid battery was part charged 12.5v unloaded, 12v ign on, Voltage went from 14.5 to 14.4 when light switched on.I will try it with a higher watt bulb.Rotor seemed to have good magnetism,battery probably 8 years old at least,but has been well maintained.I usually run at higher revs so it seems the regulator is earning its keep.

____ This all seems pretty incredible for only 3k-RPM.
Also, it seems your voltage-regulator is not very good, cuz it's job should keep system-voltage below 14-volts ! _ You should measure system-voltage above 6k-RPM, to learn if your regulator-unit is doing anything at all about excessive voltage.
On top of that, your battery seems insufficient... It as well ought to prevent such high voltage. _ So either it has a very low amp-hour rating, or it's original plate-area has become somewhat reduced, (which ought be expected for it's claimed age).
How many amp-hours is your battery rated for ?

[DewCatTea-Bob]

By: wcorey...
" I had a couple 10kuf 34v caps wired in parallel, they seem to behave very similarly to the battery when used in it's place. Made the light brighter... "

____ The light can get brighter then because the caps will store any excess power-juice which the light happens to not consume, then during the nulls (when the pulsating-DC is mostly off), the caps can then pass-on their stored-juice to fill-in the otherwise empty nulls, thus the light consumes more power (and be brighter).


" with the caps and no battery, used a typical newer type Shindengen 12v reg/rec instead of the plain rec and added a 5ohm single Dyna coil. "

____ It would be interesting to check for any effect the regulator has on power output & consumption. _ That could be done by letting the rect.block do the rectification and taking the regulator in & out of the circuit.


" Not sure how to simulate how the coil would behave (what current it would draw) while operating without making something to trigger an ignition at the appropriate rpm, "

____ I'd determine what the maximum current-draw it's capable of, then find another load that's equivalent to about 20% of that maximum figure, and leave it constantly connected to simulate ign.system power-consumption. _ (As there's only about a 1 in 5 chance at any frozen-point in time, that max.power-consumption [by the ign.system] is occurring.)


" I assume it would draw less at lower rpm. "

____ I can see where more sparks per second would seem to equate to more energy-consumption, however regardless of RPM, (in a points-system), the 'off' & 'on' percentage-ratio remains the same,, so average current-draw ought remain unchanging & fairly constant despite RPM.


" I ran the motor at 4300 rpm and got something close to 6.9a, 13.7vdc, 94.5w. "

____ Wondering how you got the "4300 rpm", exactly !?


" So that's pretty close to a functional system that would operate a bike and it will fully charge a 12v battery at somewhere around 4krpm... "

____ So that help's to prove what I've been claiming about 'power' being all that really matters with alternators, and that 'voltage' is relatively an irrelevant aspect. _ Cuz all an alternator can offer is just 'power', regardless of whether it's coils are in series or parallel.
__ (This in turn relates to a reply-post I had posted but has possibly been overlooked because I had responded to a new-post of Bill's before I later responded his very-previous post, [which had been posted on the previous/older page]. _ So if ya think ya may have missed it, here's a link back to it... viewtopic.php?f=3&t=556&start=280#p4544 .)


Onward-Cheers,
-Bob