Alternator - questions

[Keeping Up]

Reading this has set up a train of thought in my mind (not easy on a Sunday afternoon).

 

I have a diode splitter and an external (Adverc) alternator controller. Because of the voltage drop in the diodes, the output (B+) voltage from my alternator is nearly a volt higher than the battery voltage.

 

The voltage on D+ tracks that on B+ so my rotor is being fed from a source which is nearly a volt greater than the battery voltage. Therefore when it is being fully driven, I presumably get about 7% more rotor current than I would otherwise have had. This presumably results in an increased charging current.

 

So including a diode splitter increases your charging current. I wonder if a chain of diodes would give me even more?

[Timleech]

Reading this has set up a train of thought in my mind (not easy on a Sunday afternoon).

 

I have a diode splitter and an external (Adverc) alternator controller. Because of the voltage drop in the diodes, the output (B+) voltage from my alternator is nearly a volt higher than the battery voltage.

 

The voltage on D+ tracks that on B+ so my rotor is being fed from a source which is nearly a volt greater than the battery voltage. Therefore when it is being fully driven, I presumably get about 7% more rotor current than I would otherwise have had. This presumably results in an increased charging current.

 

So including a diode splitter increases your charging current. I wonder if a chain of diodes would give me even more?

 

...but surely the charging current gain will be reduced because of the voltage drop in the diode.

I suppose also it'll depend on how close the rotor is to saturation?

 

Tim

[Arnot]

Er..... This bit is wrong, and therefore all the rest of your post (which follows on from this) is also wrong.

 

It's because of this:-

 

"When an alternator is charging a battery that is very flat, the current output will be limited by the regulator providing maximum field current."

 

That bit is correct......

 

"However this current is related to the resistance of the rotor winding and the regulated voltage."

 

That bit is correct.

 

"If you increase the regulated voltage"

 

That bit is incorrect. You can't increase it. Well you can increase where it tries to go. But it won't get there. The alternator has reached its current limit, the voltage doesn't reach the regulation voltage. Therefore you can't increase the voltage across the field, therefore you can't increase the current through it. Therefore you can't increase the alternator output current.

 

At low engine speeds the output current is limited by the RPM. No amount of playing with regulators is going to change that.

 

Gibbo

 

You are absolutely correct sir! It serves me right for trying to over simplify what is quite a complex relationship in a rush.

 

When the alternator first kicks in and the batteries are both flat and in good condition, they are able to accept more current than the alternator can provide at the regualted voltage, at this stage the alternator is current limited (either by stator design or by RPM) and (as you say) the rotor current is maxed so no more is avaialable regardless of any eletrickery. That's the bit I skipped over - well spotted...

 

However, in practice, this stage generally only lasts for a very short time given even moderate engine speed and once the current the batteries need to support the regulated voltage falls below the maximum of the alternator this is where the problems generally occur. At this stage, the regulator kicks in and the alternator starts to operate in constant voltage mode and here sophisticated regualtion systems can help.

 

On a conventional alternator, the regulator measures the voltage available on the diode trio and operates on the assumption that this is the required output, but, as has been commented on elswhere, the voltage drop on the main rectifier under load is significantly higher than that of the diode trio and at higher currents there is a significant voltage drop across the wire from the alternator to the batteries, dismissing for the moment any splitter diodes or FET's.

 

In consequence, in the early stages of the constant voltage mode when the output current is not at maximum but somewhere near it, the voltage the regulator "sees" is quite a lot higher than the actual battery voltage and it holds back the available current unnecessarily. Alternatively, if there is a regulation system that "sees" the true battery voltage and thus compensates for the diode error and cable losses the full output of the alternator can be maintained for far longer and during part of the charging cycle actually increased over and above the current that a conventional system would be able to produce (as described in my last post).

 

Please forgive me but the compex interaction of non linear variables that takes place when an engine with varying speed drives an alternator with varying output into batteries with varying and non-linear charge acceptance characteristics pretty much defies any verbal explanation (for me at least) and I have probably not made the topic much clearer.

 

However the effect can be amply demonstrated in practice. If a regulator with remote sensing is fitted to an engine and made to charge a battery bank that is fairly flat. Once the initial constant current mode has finished if the ouput current is monitored as the sensing position is changed from the diode trio output to the battery positive, a significant increase in the charging current will be observed. In addition it will be seen that the charging voltage on the battery terminals both rises to the actual regulated voltage and it becomes more constant at varying engine revs.

 

CAUTION - do not try this at home! to do this demonstraton you must have some resistor networks in place to prevent the sensing wire ever being open circuit. If this happens the alternator can run unregulated and if the batteries are reasonably well charged and/or the engine reves are high the voltage can achive levels that would damage equipment.

 

I hope that this clears it up slightly.

 

Regards

 

Arnot

[Arnot]

Reading this has set up a train of thought in my mind (not easy on a Sunday afternoon).

 

I have a diode splitter and an external (Adverc) alternator controller. Because of the voltage drop in the diodes, the output (B+) voltage from my alternator is nearly a volt higher than the battery voltage.

 

The voltage on D+ tracks that on B+ so my rotor is being fed from a source which is nearly a volt greater than the battery voltage. Therefore when it is being fully driven, I presumably get about 7% more rotor current than I would otherwise have had. This presumably results in an increased charging current.

 

So including a diode splitter increases your charging current. I wonder if a chain of diodes would give me even more?

 

Not quite – including a diode splitter will not actually increase the amount of charging current available to the batteries, what it does is fools the regulator into thinking that the battery voltage is higher than it actually is thus causing the regulator to restrict the output of the alternator prematurely and reducing the charging current.

 

I suspect that you are confusing rotor current and output current here…

 

Regards

 

Arnot

 

 

...but surely the charging current gain will be reduced because of the voltage drop in the diode.

I suppose also it'll depend on how close the rotor is to saturation?

 

Tim

I think you have it right on the matter of diode voltage drop.

 

The issue of saturation (I presume you mean magnetic flux saturation) is actually almost never a problem in practice, providing the rotor design is good (and it usually is) then it would require significantly more current than could be obtained from a nominal 14.4v supply to saturate the rotor. However as the temperature rises this becomes more of a possibility.

 

Regards

 

Arnot

[Keeping Up]

Not quite – including a diode splitter will not actually increase the amount of charging current available to the batteries, what it does is fools the regulator into thinking that the battery voltage is higher than it actually is thus causing the regulator to restrict the output of the alternator prematurely and reducing the charging current.

 

I suspect that you are confusing rotor current and output current here…

No, read my post again and think harder about it.

 

I have an external alternator controller which is battery sensed, hence until the pre-set voltage is reached by the battery, the alternator is running "flat out".

 

I am not confusing the two currents; what I am wondering is whether the higher rotor current (which is a consequence of the extra volt being applied across it during this bulk charging phase) would result in an increased output current.

[Arnot]

No, read my post again and think harder about it.

 

I have an external alternator controller which is battery sensed, hence until the pre-set voltage is reached by the battery, the alternator is running "flat out".

 

I am not confusing the two currents; what I am wondering is whether the higher rotor current (which is a consequence of the extra volt being applied across it during this bulk charging phase) would result in an increased output current.

 

Sorry, I ignored the the bit about the Adverk. Correct me if I am wrong here but the Adverk devices I have come across are not actually regulators in the true sense of the term but a device that overrides the original regulator for a while to boost the charging voltage temporarily.

 

I was looking at a boat the other day that had some wierd intermittent charging fault (needless to say it worked perfectly all the time I was there) with one on and the owner had merely disconnected the wire to the alternator and the original regulator continued working. From what I could see it merely kept the rotor current maxed for longer by keeping the field terminal to earth when the inbuilt regulator would have allowed it to float up. Rather like converting the alternator to a three stage charger but without any feedback loop.

 

What it did not have was any connection to the batteries themselves and thus could have no way of knowing the actual battery voltage. Of course it possibly had not been installed correctly but since the fault did not occur after an hour and the owner did not have flat battery issues I didn't poke around too much.

 

I accept that there may be other Adverk devices and if you let me know what yours is I will try to get some idea of what it does from their website.

 

However, back to your question "what I am wondering is whether the higher rotor current (which is a consequence of the extra volt being applied across it during this bulk charging phase) would result in an increased output current." The answer is yes, given sufficient speed and that the alternator is in constant voltage mode, an increased rotor current would result in higher rotor flux density and thus higher stator voltage or current at a given speed. However this is only one part of the charging current improvement, probably more is gained by the stator peak voltage being allowed to go higher than normal as per the maths bit in a previous post I made. A double whammy...

 

Regards

 

Arnot

[Keeping Up]

There are many alternator controllers available, of which the Adverc is just one. There may be simple non-sensing ones available but I've never seen one installed in a narrowboat.

 

The simplest one in common use is Acorn's Kestrel 90, which incorporates battery sensing but is just a one-shot booster. The Kestrel is generally regarded as totally obsolete nowadays. All the others - Adverc, Sterling, and so on, are true regulators which sense the battery voltage with varying degrees of sophistication (including temperature compensation etc) and generally drive the alternator as a multi-stage charger with full feedback.

 

So it looks as if here is one point to those of us who favour diode splitters over relays: unexpectedly, they can actually increase your available charging current.

[Arnot]

There are many alternator controllers available, of which the Adverc is just one. There may be simple non-sensing ones available but I've never seen one installed in a narrowboat.

 

The simplest one in common use is Acorn's Kestrel 90, which incorporates battery sensing but is just a one-shot booster. The Kestrel is generally regarded as totally obsolete nowadays. All the others - Adverc, Sterling, and so on, are true regulators which sense the battery voltage with varying degrees of sophistication (including temperature compensation etc) and generally drive the alternator as a multi-stage charger with full feedback.

 

Further to my last post I just had a look at the Adverk website but could not seem to find any installation instructions or description of the dv/dt algorithm used by their system. However in all fairness since this is commercially sensitive it’s not too much of a surprise.

 

From what I could divine, the Adverk BMS is a relatively sophisticated form of overlay alternator regulation that overrides the original regulator if it’s own algorithm decides that the output voltage should be higher. It does clearly show a sensing lead connected directly to the services battery and thus should be able to correctly control the charging voltage. I suspect that the algorithm is similar to the type used by many multi stage lead acid battery chargers and has bulk, absorption and float voltages.

 

Although in their text they mention temperature compensation I don’t see any sensing system for this and so presume that it is built in to the control unit. This is fine if it is thermally coupled to the batteries otherwise it will just be reacting to the temperature of the control unit which with a) a different location and a different thermal mass, will bear little relationship to the battery state.

 

There is another potential problem here. Some alternators have a regulator system that not only ties the field terminal low to increase the output but can also feed it with a voltage to actively lower the alternator output. If the Adverk BMS is connected to one of these there will occasionally be a conflict. In this case the strongest device will win and the weaker device (or even both of them) may well be permanently damaged. If this happens it may not be apparent and in all fairness may have no practical impact but in a worst case scenario regulation may be completely lost and the whole system could cook.

 

One final thing, I didn’t see any method of trimming the regulation points of the BMS to suit differing battery types but then I didn’t have the instructions so there may be some way of doing this. For example, long periods of 14.4v applied to engine heated cheap flooded cell batteries will shorten their life significantly whereas the same voltage applied to thin plate, pure lead AGM batteries will take an age to get them fully charged.

 

Having said all that, I like the approach of Adverk and a lot of what they say on their site not only makes good sense but also betrays a profound understanding of what is going on. Reading between the lines they have the same difficulty I have found in explaining a complex non-linear relationship in simple terms.

 

So it looks as if here is one point to those of us who favour diode splitters over relays: unexpectedly, they can actually increase your available charging current.

Absolutely true, and as an interesting extension of this, there will probably be an optimum for the voltage drop in the wiring between the alternator output and the batteries. Thus although common sense dictates that the thicker the cable the better, this may not actually be the case...

 

One word of warning here, when a diode type split charge is used in this type of control system the additional quantity and duration of the charging current will increase the thermal ouput of the splitter diodes and they may well get very warm. It is thus a good idea to make sure that they are mounted so that the cooling fins are vertical and preferably in some sort of forced air stream.

 

My preference would be for a relay type split charge system and optimised cable resitances to give about 1.0v drop at full alternator output. This is the system I have used for the last thirty something years and it seems to work just fine

 

Regards

 

Arnot

[Keeping Up]

Although in their text they mention temperature compensation I don’t see any sensing system for this and so presume that it is built in to the control unit. This is fine if it is thermally coupled to the batteries otherwise it will just be reacting to the temperature of the control unit which with a) a different location and a different thermal mass, will bear little relationship to the battery state.

The battery voltage sense lead has a thermistor at the battery end

Edited May 4, 2008 by Keeping Up

[Arnot]

The battery voltage sense lead has a thermistor at the battery end

Ah hah! that explains it - thanks for the info.

 

That probably means that with a bit of modification the charging voltage can be trimmed as well.

 

Regards

 

Arnot

[Gibbo]

Reading this has set up a train of thought in my mind (not easy on a Sunday afternoon).

 

I have a diode splitter and an external (Adverc) alternator controller. Because of the voltage drop in the diodes, the output (B+) voltage from my alternator is nearly a volt higher than the battery voltage.

 

The voltage on D+ tracks that on B+ so my rotor is being fed from a source which is nearly a volt greater than the battery voltage. Therefore when it is being fully driven, I presumably get about 7% more rotor current than I would otherwise have had. This presumably results in an increased charging current.

 

So including a diode splitter increases your charging current. I wonder if a chain of diodes would give me even more?

 

It's a nice idea but.................

 

With no load on the output the rotor magnetically saturates at *very* low rotor voltage. The only limiting factor on rotor current (with no load on the output) is the impedance of the rotor. With zero rotor impedance even 0.0001 volt would cause magnetic saturation. Magnetic flux density is proportional to current not voltage.

 

As the output current is increased the back EMF from the stator reduces the flux denisty in the stator and the rotor. This allows more current to be forced into the rotor by increasing the rotor voltage. But the flux denisty remains roughly the same as with a very small load. There is the rotor current forcing it one way, and the back EMF from the stator forcing it the other way. They balance exactly which is one of the inherent current limiting effects of alternators.

 

If you put lots more voltage on the rotor this (obviously) causes an increase in flux density (assuming it isn't saturated which it won't be if there is a load on the output) but as soon as the output current increases the back EMF reduces it back to where it was.

 

I had a similar thought a few years ago and thought about producing a controller that overdrives the rotor for low revving engines. You get a *bit* of an increase but only maybe 1 or 2%. It took me some considerable time to work out why.

 

Now you could increase the size of the rotor and reduce the resistance of the rotor and stator to allow this idea to work but then you've got a bigger alternator anyway!

 

Alternators are inherently self current limiting.

 

Gibbo

Edited May 5, 2008 by Gibbo

[Keeping Up]

Thanks Gibbo. I just knew there had to be a flaw in the logic somewhere, but I just couldn't see it.

[Dylan]

Snip

"Alternators are inherently self current limiting."

 

Gibbo

 

[sueanddaren]

Similar in its effect, just quite a bit easier to do ( and sans blue LED,s ) it is in effect an adjustable, battery sensing internal regulator. The kit consists of a new regulator brush box that just replaces the original. The replacement item has another wire that has to be connected to the positive of the battery that is used most (probably the domestics) through a calibration resitance to set the desired voltage. It can even be made variable at the panel if you wish but I wouldn't reccomend it.

 

This then battery senses the charging system and at the same time allows the regulating voltage to be set to whatever is appropriate. Very useful if you use AGM or Gel batteries.

 

Typically the ultimate output of the alternator only rises by about 10-20% but the real benefit comes at low engine speeds. At 50 to 100% above the speed where the alternator commences charging the increase in current can be over 100%. There is a technical justification for this but I won't bore you with it unless you really want to know.

 

I first developed this system in the 70's when I did a couple of boats with seven 2V fork lift truck cells to provide about 14.5v even with the engine not running and plenty of capacity. This was to get round the problems of voltage drop for the early pumps fitted at the front of a 70ft hull. The regualtion system allowed me to run the 160A alternator at 16.8v charging voltage at the battery for a quick recharge and good control, I seem to remember it increased the current output to about 210A but it was thirty years ago...

 

At the moment I only have an updated version for the A115 and A127 but am working on varieties for the Iskra unit fitted to Beta engines and both 12v and 24v for the Leece Neville bigguns.

 

By the way, if you really want loads of output at low engine speeds on one of the Gardners, Rustons or other vintage engines, then Leece Neville is the way to go. Just a bit expensive.

 

Hope this helps...

 

Regards

 

Arnot

 

 

OK now the big question, how much?

 

"14.4v applied to engine heated cheap flooded cell batteries will shorten their life significantly whereas the same voltage applied to thin plate, pure lead AGM batteries will take an age to get them fully charged."

 

6 hours with only 25% current from 100% discharged (C10) sounds quick to me.

 

Daren

[Sir Nibble]

Glad we've got all that sorted out!

Does occur mind, sensing off the battery +ve is a fine idea, but without sensing off -ve as well, isn't it half a job?

[Gibbo]

Glad we've got all that sorted out!

 

Wishful thinking

 

Does occur mind, sensing off the battery +ve is a fine idea, but without sensing off -ve as well, isn't it half a job?

 

It is indeed. I seem to remember one of the Sterling alternator controllers used to have a -ve sense lead as well?

 

It always seemed a bit pointless to me without that. Though I suppose the main reason for it is to compensate for diode drops.

 

To me it seems far more sensible to size the cables poroperly so there aren't any obnoxious voltage drops through the cabling rather than trying to compensate for it with the regulator.

 

Gibbo

[Sir Nibble]

Interestingly, when I was mucking about with the low speed configured alternator, I calculated that the maximum current would be the original 65A over root 3 which gave me around 37A. In practice I got 40A. Fair unuff, manufacturing tolerance etc, But when I installed the battery sensing I experimentally installed a switch and having driven the 40A output into a flat battery I switched in the batt sensed system. Blow me! the current leapt up to about 55A! Now I found this difficult to get my head round, and my theory is that since inductive reactance XL = 2pi x fL, the lower value of f due to the lower speed will have caused a reduction in overall impedance.

Comments are invited.

[Gibbo]

Interestingly, when I was mucking about with the low speed configured alternator, I calculated that the maximum current would be the original 65A over root 3 which gave me around 37A. In practice I got 40A. Fair unuff, manufacturing tolerance etc, But when I installed the battery sensing I experimentally installed a switch and having driven the 40A output into a flat battery I switched in the batt sensed system. Blow me! the current leapt up to about 55A! Now I found this difficult to get my head round, and my theory is that since inductive reactance XL = 2pi x fL, the lower value of f due to the lower speed will have caused a reduction in overall impedance.

Comments are invited.

 

It's possible but there's an alternative theory.

 

Your mod increases the DC resistance of the stator. It will therefore conduct through the diodes over a wider angle giving a higher average current.

 

I have no idea which is correct. It could be a combination of both.

 

Gibbo

[Arnot]

Interestingly, when I was mucking about with the low speed configured alternator, I calculated that the maximum current would be the original 65A over root 3 which gave me around 37A. In practice I got 40A. Fair unuff, manufacturing tolerance etc, But when I installed the battery sensing I experimentally installed a switch and having driven the 40A output into a flat battery I switched in the batt sensed system. Blow me! the current leapt up to about 55A! Now I found this difficult to get my head round, and my theory is that since inductive reactance XL = 2pi x fL, the lower value of f due to the lower speed will have caused a reduction in overall impedance.

Comments are invited.

Thank you Snibble, this is the phenomenon I was trying (inadequately) to describe in an earlier post and my explanation is this;

 

If you draw on paper (or a computer if you are that way gifted) the waveform produced by a three phase stator with a peak of about 20v and then draw a line at the DC voltage that a battery would start charging at (say 13v) then the difference voltage divided by the internal resistance of the battery will give the current. The current multiplied by the difference voltage will give you the power transfer.

 

If you then do another drawing with the peak voltage increased to say 23v (as in battery sensing) and do the same equation you will find that the current and more particularly the power transfer has increased significantly.

 

Now if you do the same two calculations with a single phase alternator, you will find that because of the lower duty cycle of the waveform the difference is even more pronounced and the increase from 40A to 55A becomes pretty much predictable.

 

What you actually did was just what I suggested although the increase on a standard alternator would probably not be so great.

 

Did you notice the more rapid warming of the stator?

 

It is my guess (and I have not done the calculations) that the inductive component of the stator is relatively insignificant compared to the effects of the winding resistance and won’t have much impact on the output. If the inductance did have a marked effect then I would anticipate that the output of alternators would fall off as the shaft speed and hence frequency increased, and it doesn’t seem to on test.

 

Question – did you try the same switching of the sensing point at different speeds? My prediction would be that there would be little or no difference. After all the frequency at 2500rpm would only be about 42Hz, there are very few windings and quite a lot of air gap.

 

Please bear in mind that in doing this sort of thing, I don’t take a very academic approach more an empirical one. As long as the effect is explicable and more or less predictable, I prefer to leave the absolute mathematical proof to others…

 

So on this basis I will stick to my rule of thumb that the current output of an alternator is proportional to the difference between the voltage the battery starts to draw current and the peak current divided by the internal resistance of the battery. There are a lot of other linear and non-linear variables involved but even lumped together they are minor.

 

Regards

 

Arnot

[Sir Nibble]

Thank you Snibble, this is the phenomenon I was trying (inadequately) to describe in an earlier post and my explanation is this;

 

If you draw on paper (or a computer if you are that way gifted) the waveform produced by a three phase stator with a peak of about 20v and then draw a line at the DC voltage that a battery would start charging at (say 13v) then the difference voltage divided by the internal resistance of the battery will give the current. The current multiplied by the difference voltage will give you the power transfer.

 

If you then do another drawing with the peak voltage increased to say 23v (as in battery sensing) and do the same equation you will find that the current and more particularly the power transfer has increased significantly.

 

Now if you do the same two calculations with a single phase alternator, you will find that because of the lower duty cycle of the waveform the difference is even more pronounced and the increase from 40A to 55A becomes pretty much predictable.

 

What you actually did was just what I suggested although the increase on a standard alternator would probably not be so great.

 

Did you notice the more rapid warming of the stator?

 

It is my guess (and I have not done the calculations) that the inductive component of the stator is relatively insignificant compared to the effects of the winding resistance and won’t have much impact on the output. If the inductance did have a marked effect then I would anticipate that the output of alternators would fall off as the shaft speed and hence frequency increased, and it doesn’t seem to on test.

 

Question – did you try the same switching of the sensing point at different speeds? My prediction would be that there would be little or no difference. After all the frequency at 2500rpm would only be about 42Hz, there are very few windings and quite a lot of air gap.

 

Please bear in mind that in doing this sort of thing, I don’t take a very academic approach more an empirical one. As long as the effect is explicable and more or less predictable, I prefer to leave the absolute mathematical proof to others…

 

So on this basis I will stick to my rule of thumb that the current output of an alternator is proportional to the difference between the voltage the battery starts to draw current and the peak current divided by the internal resistance of the battery. There are a lot of other linear and non-linear variables involved but even lumped together they are minor.

 

Regards

 

Arnot

 

I should point out, that this whole system is a long way from standard now. Rotor excitation is from the battery and the excitation trio in the rectifier is now redundant. I was testing with too small a battery bank to sustain the test for more than a second or two before the regulator cut in and the current nose dived, I didn't actually load it to test maximum current, that's not what I was after.

The stator? Yes, part of this whole project has been frequent stripping of the alternator and the stator gets hot enough to significantly discolour the insulating varnish. I'm working on the basis that A127 stators are lying about all over the place and If I need a spare it's available. The weakest link in an A127 appears to be the rectifier, and that of course is getting an easier time of it.

[Arnot]

I should point out, that this whole system is a long way from standard now. Rotor excitation is from the battery and the excitation trio in the rectifier is now redundant. I was testing with too small a battery bank to sustain the test for more than a second or two before the regulator cut in and the current nose dived, I didn't actually load it to test maximum current, that's not what I was after.

The stator? Yes, part of this whole project has been frequent stripping of the alternator and the stator gets hot enough to significantly discolour the insulating varnish. I'm working on the basis that A127 stators are lying about all over the place and If I need a spare it's available. The weakest link in an A127 appears to be the rectifier, and that of course is getting an easier time of it.

It sounds like an interesting project, have you posted any details elswhere? I have a load bank for testing alternators which can eliminate the variable of battery charge but I am sure that this could be reasonably replicated with other more readily available loads.

 

My interest at the moment is what happens just at the point the regulator cuts in, until then while the alternator is in constant current mode there is little a control system can improve. Once the regulator cuts in and the alternator enters constant voltage mode there are considerable gains in battery charging to be had by keeping the current output as high as a safe battery voltage will allow for as long as possible.

 

I agree about the A127 being readily and cheaply obtained and with the comment about the rectifier. I have not looked at this recently but it used to be the case that many of the aftermarket rectifiers were a lot better than the originals.

 

Regards

 

Arnot

[Sir Nibble]

It sounds like an interesting project, have you posted any details elswhere? I have a load bank for testing alternators which can eliminate the variable of battery charge but I am sure that this could be reasonably replicated with other more readily available loads.

 

My interest at the moment is what happens just at the point the regulator cuts in, until then while the alternator is in constant current mode there is little a control system can improve. Once the regulator cuts in and the alternator enters constant voltage mode there are considerable gains in battery charging to be had by keeping the current output as high as a safe battery voltage will allow for as long as possible.

 

I agree about the A127 being readily and cheaply obtained and with the comment about the rectifier. I have not looked at this recently but it used to be the case that many of the aftermarket rectifiers were a lot better than the originals.

 

Regards

 

Arnot

http://www.canalworld.net/forums/index.php?showtopic=12027

it used to be the case that many of the aftermarket rectifiers were a lot better than the originals.

Yup, even down to copper heatsinks instead of ally.

I also have a loadbank and I am knee deep in A127s.

[Gibbo]

If you draw on paper (or a computer if you are that way gifted) the waveform produced by a three phase stator with a peak of about 20v and then draw a line at the DC voltage that a battery would start charging at (say 13v) then the difference voltage divided by the internal resistance of the battery will give the current. The current multiplied by the difference voltage will give you the power transfer.

 

The magnetic interaction between the stator and the rotor limits the current. It's just the way it is.

 

If you then do another drawing with the peak voltage increased to say 23v (as in battery sensing) and do the same equation you will find that the current and more particularly the power transfer has increased significantly.

 

The magnetic interaction between the stator and the rotor limits the current. It's just the way it is.

 

It is my guess (and I have not done the calculations) that the inductive component of the stator is relatively insignificant compared to the effects of the winding resistance and won’t have much impact on the output. If the inductance did have a marked effect then I would anticipate that the output of alternators would fall off as the shaft speed and hence frequency increased, and it doesn’t seem to on test.

 

Question – did you try the same switching of the sensing point at different speeds? My prediction would be that there would be little or no difference. After all the frequency at 2500rpm would only be about 42Hz, there are very few windings and quite a lot of air gap.

 

Do your alternators use a single pole rotor? Mine don't hence the frequency is somewhat higher than this. Like about 7 times higher.

 

Gibbo

[Arnot]

The magnetic interaction between the stator and the rotor limits the current. It's just the way it is.

So, let me get this clear, are you saying that the resitance and inductance of the stator have no limiting effect? The way I see it is that the magnetic interaction between the rotor and the stator actually generate the voltage and the current is limited by the size of the windings providing there is no saturation.

 

Do your alternators use a single pole rotor? Mine don't hence the frequency is somewhat higher than this. Like about 7 times higher.

 

OK so I used the example of a simple rotor and a multi pole rotor (as used on the A127) would give a higher frequency. I still doubt that it would have much effect on the output. I would like to find out what impact the frequency has on the output by controlled testing, have you tried it? Transformers seem to have a pretty linear transfer characteristic within a wide frequency band and I can't think of any obvious reason why an alternator should be different.

 

Regards

 

Arnot

[Gibbo]

So, let me get this clear, are you saying that the resitance and inductance of the stator have no limiting effect? The way I see it is that the magnetic interaction between the rotor and the stator actually generate the voltage and the current is limited by the size of the windings providing there is no saturation.

 

No that's not what I'm saying. They have a huge effect obviously. The magnetic interaction between the rotor and the stator DO actually generate the voltage. You are quite correct there. But they are also responsible for the major inherent current limiting effect of alternators. It's one of the (many) reasons they are so damned reliable. If the stator resistance alone was resonsible for the current limit then they would simply burn out when run into a short circuit. They don't. Because the back EMF from the stator reduces the flux as the output current tries to increase thus giving a negative feedback type of current limit.

 

OK so I used the example of a simple rotor and a multi pole rotor (as used on the A127) would give a higher frequency. I still doubt that it would have much effect on the output. I would like to find out what impact the frequency has on the output by controlled testing, have you tried it? Transformers seem to have a pretty linear transfer characteristic within a wide frequency band and I can't think of any obvious reason why an alternator should be different.

 

Indeed but that's not the issue. The issue is that the magnetic coupling between the rotor and stator is absolutely terrible in an alternator when compared to a transformer. There is HUGE leakage inductance in both. As you try to increase the output current, lots of voltage is dropped across the leakage inductance of both the stator and the rotor. Voltage dropped across the leakage inductance of the rotor us just wasted volts reducing the flux density but not actually dissipating power(remember that even though driven with DC voltage when the reg is "hard on", the current in the rotor is a mirror of the output current [which is AC - but smaller obviously] and the flux density is related to the current not the voltage). Voltage dropped across the stator leakage inductance obviously reduces the output voltage (and hence current) but also reduces the flux density in the stator without affecting the rotor because the coupling is so bad.

 

If you put a fixed voltage on the rotor, and increase the load on the output you will see the current remain *roughly* constant and the voltage will decrease. You can do this right down to zero volts output. The current will remain roughly the same. It doesn't go up, it doesn't go down. Alternators are, to a first appximation, constant current output devices irrespective of voltage.

 

Gibbo