Alternator transformer/rectifier remote mounted (test 1.0)

[Wizard]

Bob g,

When you are getting into high frequency stuff, things get more critical and more lossy.

1. length of cabling between source and load and type of material.  High frequency stuff tend to have short wirings between source and load to keep inductive losses and inpendance losses low and rectify 500Hz to DC immediately at the source or short from the generator itself.

2. type of transformer becomes more different and specialized above 100Hz.  You're going above 500Hz means different windings and type of core material etc.

3. Construction of copper core (litz or solid, copper or aluminum etc), cabling type does matter as well.  Some impedes the 500Hz worse than other.

Cheers, Wizard

[mike90045]

It's three phase power, Mike, so each wire is carrying 1/3 the current- and that's the current at higher voltage, much less than 100amps.  That and the improved efficiency of the alternator at higher volts is the beauty of Bob's design.

Got it, I was thinking 100A at each phase !

[mobile_bob]

Wizard:

thanks for the input, let me address each as best i can

1. while this is generally true, submarines, aircraft and a few other applications use 400hz at some distances, but i am learning about inductive
losses and all the rest, and am going to try to quantify as best i can these losses.

2. transformers:  a few year ago while doing early research on alternator efficiency and the use of transformers, and also being familiar with delco
and leece neville transpacks, i decided to call a few transformer manufactures and talk to their EE's they told that for my purposes the use of 60hz
transformers could be used to 400hz no problem and upwards of 1000hz before things start to drop off. i was assured that at 5-600hz i would see no
appreciable drop in efficiency using a 60hz tranformer.

higher frequencies require thinner laminations, 60hz generally use .014" thick, the transformers i am useing have .008" perhaps not as thin as i would like but a step in the right direction.  the magnetizing current is ~.1amp  and they do not warm up at all, so i don't think that 500hz is a problem at least that i have an indication of.

3. i am using stranded 10/3 SO cable,  only because that is what i have on hand, i suspect there are losses in transmission, but so far i have not
tested/measured to determine what those losses are. i suppose i could place a meter at the alternator and another at the input of trans pack
to see what the drop is in the 10/3?

this project is a learning experience for sure, and the system would not be my first choice, as  i would rather get my batteries as close to  the
charging source as possible, however i realize this might not be an option for some folks, Lloyd being one.

until i run the BSFC test to compare against an ordinary charging system (12volt alternator going directly to a closely located battery), i can't say for sure whether it is a viable system or not.

it may well turn out to be better to generate high voltage DC, transfer it over the required 30 odd feet, feed an mx60 or similar and let it do the
charging.

in either case getting the high voltage from the alternator will require it to be spinning upwards of 5-6krpm which means no charging at engine idle
or low speed operation.

always compromises

bob g

[mobile_bob]

got in an incomplete test run today, but ran out of fuel before i could finish

max output 125amps DC

attempted to measure the voltage drop on a single conductor of the 10/3 line, at full loading i got a it over 2volts drop
and the cable never even got warm,, but this was an intermittent short duration test, so

not much was learned other than it would appear that the voltage drop across the 10/3 in transmission due to high frequency issues
does not seem to be significant,

more testing tomorrow

bob g

[mobile_bob]

got some fuel and had just enough time to run a quick test
voltage drop at max load on one leg of the 10/3 is just under a half volt

i think that is acceptable.

what i am now researching is how to accurately measure amperage on a three phase line that is running at 500hz.

i am not convinced that my clamp on is at all accurate at high frequencies, relatively speaking of course (500 being about
8 times higher than 60hz)

this is not something that is addressed in just any old book, so i have to do some digging.

bob g

[V5CVBB]

How hard would it be to measure the rectified DC and work backwards to get the 500hz AC current?  Or are you looking for more precision?

Kevin

[mobile_bob]

i could work backwards if i knew the efficiencies of the rectifiers, and the transformers running at 500hz

without those numbers i can't calculate the ac current.

but i am getting  closer,

its like formulating a logic problem, taking all the data i have and then arranging it into formula's to get those numbers
i can't directly measure.

after doing a relative BSFC test, i think i will have enough data to build enough formulas so that i can determine the AC current
and the efficiencies of the rectifiers and transformers to a fair degree of accuracy

bob g

[V5CVBB]

Efficiency of rectifier isn't a problem is it?  Measured voltage drop across diodes and DC current gives losses in watts.  Maybe I'm missing something. I haven't thought it through.

Should be able to find published losses of running 60hz trans. at 400hz that would get you in the ballpark with a little interpolation.

Good luck.  I'm just thinking out loud. 

Kevin
 

[mobile_bob]

not a problem, i like it when people think out loud

the problem with the transformers is even if i knew who made them, which i do
contacting the manufacture about specifics of the operation of the units is a
secret bound by contract.

i have been down this road before

the transformers are salvaged from APC ups units, where the oem contracts out to a transformer
manfacture to custom build units for use in their ups design, the transformer manufacture as part
of the contract is bound to never release specifics about the transformers.

however, i am glad you mentioned this, because i will try the transformer manufacture to see if
they too are bound by a similar contract.

my bet is they are, but it cannot hurt to ask

bob g

[mobile_bob]

ok, this will probably be the final test of this system, by me
i have found out what i wanted to know, and here is a rough draft of the results

when i get some time i will compile what was done and attach it to the white paper section of the forum

final test  2/21/10

alternator 175amp leece neville, stator reconnected as Y from the oem delta configuration

alternator test speed ~4850rpm

field current 14volts @ 5 amps

stator open voltage phase to phase 99vac

stator voltage under full test load 85vac

transfer line from alternator stator to remote transformer/rectifier assembly = 10/3 SO cable
length of cable 6ft

transformers used 1.25kva transformers salvaged from APC rm1000 ups systems,
3 were used configured in a three phase setup with delta connected primaries and Y connected secondaries

rectifier bridge stock leece neville 3 phase bridge mounted to breadboard

load nichrome high power resistor

engine speed 1800rpm

engine started and brought up to full operating temperature, and stable operation

apply field current to alternator and measure dc output voltage and current to power resistor load

15.1 vdc @90amps,  1359 watts
fuel consumed per kwatt hour 1294 grams

comparison test load on st head

1300watts, 984gr/kwatt/hr

note:  for those that want to analyze the BSFC of this unit, please note that this unit is a trigenerator driving several components
electric waterpump, electric fan, 555 alternator for starter battery charging etc, and the full load output is approx 7.2kwatts from the st
head,  all generators in this class will have similar poor fuel consumption numbers being run at very light loads,

from earlier calculations the st head works out to be approx 78-80% efficient, for the sake of ease of illustration
we will accept it as 80% efficient

so with the alternator consuming 1294gr/kw/hr and the st consuming 984 gr/kw/hr  under similar loading

984/1294 =.76

80% st efficiency x .76 = ~61% efficiency for the alternator/transformer/rectifier

from leece neville the alternator under full load is approx 54% efficient, so it would appear there is some improvement in
efficiency using the transformers to allow the alternator stator voltage to remain high, the alternator is much higher in efficiency
at the higher voltage and with the use of the transformers as a load matching device there is ample increase in alternator efficiency
to cover the losses associated with the transformers.

other observations

the 10/3 cable carries ~13.2amps per phase or approx 13.2 x 1.73 = 22.83amps

22.83 x 85volts = 1941watts

1359watts output / 1941watts input = ~70%

therefore the difference in the 70% and the calculated 61% is likely attributed to rectifier losses predominately
(also it is fair to note that my clamp on meter may not be accurate at 500hz measuring the AC current per phase)

more...

the 10/3 cable remains cool to the touch, as do the feed wires to each of the transformer primaries

the rectifier heat sinks heat to approx 155 degree's F, and would be cooler had i placed a fan on them

the transformer secondary windings got to 126 degrees F after several tests, without fan cooling so it would
appear there is no heating issues with them.

the alternator stator gets quite hot at ~240 degree's F, i should also note that i am using the wrong rotation fan
and there is likely not enough airflow for this operation, while 240F is within the alternators capability the higher temperature
reduces the output and efficiency of the alternator.

conclusion

i think there is more that can be done to optimize such a system, using transformers better suited to this type of operation
likely would improve the overall efficiency a bit, as would a better rectifier, and an alternator with a stator optimized for this
type of service, along with a proper cooling fan.

it would appear that there is no reason why an alternator cannot be used as a 3phase generator of AC current, transfer the power
some distance at higher voltages, then step it back down and rectify for battery charging, and in doing so not suffer losses in efficiency.

it is also likely in my opinion that the overall efficiency could be improved as well.

while there may well be some flaws in my testing procedures and measurements, it is my opinion that the rubber meets the road in
how many grams per kwatt/hr the engine burns to produce the intended voltage and amperage needed to properly charge a battery bank.

i think my numbers are good enough, the estimate of system efficiency being approx 61% is likely within +/- 2 points, and in any event
the results illustrate that the system is certainly as efficient as a stock and closely coupled charging system and likely higher in efficiency.

certainly there appears to be adequate promise with the system that ought to warrant further R&D by someone that is interested or has
a real need for what the system offers.

bob g

[mobile_bob]

i wanted to add the following

the alternator stator connections i modified from delta to Y

the voltage goes up 1.73 time, but the amperage goes down by the reciprocal
or about .57

so for an alternator that is rated originally at 175amps it would then become rated
at 175 x .57 = ~100amps

the oem rating is based on 77 degree's F, and is the cold rating
the hot rating is ~145amps originally, so

145amps hot x .57 = 82.65amps

about 83 amps should be the max rating for the stator, we are drawing approx
13.1v x 1.73 x pf = 22.66 amps assuming pf of unity (which it is probably close to at full load
on the transformers)

from an amperage standpoint, the alternator is well under its rating of ~83amps at ~23amps

i am thinking that my next move before i abandon this project will be to get an accurate measurement of the
alternator output amperage at the 500hz operational speed.

i am not convinced that i can accurately measure this high of power in hz with a clamp on amp meter.

bob g

[mobile_bob]

ok, so i can't seem to help myself

like a dog on a bone, i had to try again today

this test i ran with the field at 75% approx excitation

the results boil down to an overall efficiency of 64%, as opposed to the st head being approx 80% efficient

there might be a bit more that can be found by a bit more tweaking, and i guess i will always wonder "what if"
if i don't take the time to try everything i can think of while it is setup.

what i should probably do is run a battery of tests at various loading and plot a graph of output power vs fuel consumption
and then work to determine if there is a sweet spot where the efficiency is markedly higher than at full power, it would appear
that this is likely the case as the efficiency improved approx 2 points by running at ~75% excitation.

if i had to guess, i would bet that the alternator is more efficient than the st head at high voltage, by a significant margin.

btw, i checked the alternator dc posts which are fed by the internal rectifier bridge with updated diodes, (which i am not using
in this test) the open circuit voltage is 147vdc at full field.

with enough resistive load i should be able to test the efficiency of the alternator apart from the transformer pack by retesting
with the dc output feeding a dc load at ~120vdc and measuring the amperage

that might be interesting, because i could then calculate the efficiency of the ups transformers i am using in this application (500hz)

so i guess i am not done yet 

bob g

[mobile_bob]

i believe the system will return similar results over a longer distance at the current voltage, and certainly would work
over a quite long distance with higher voltage, at least i would think.

i found a longer piece of 10/3 and am going to put in it line with the system and retest to see if there is any degradation
with increased length.

not sure i would use this system myself, but
the lessons learned here will likely find application down the road.

bob g

ps.  just dug out the 10/3 and measured it at 40ft
will try to get a test in today using it.
actually i am kind of excited to see what the results might be.

[mobile_bob]

yes i have seen this recently on either a bosch or a denso alternator

iirc it was a denso

the 3rd harmonic is useful on higher speed alternators, but goes away at low rpm

in my testing since i am running at near 5krpm there might be some power available
there, and i plan on doing some testing to find out.

i just wish that i would have brought out the center connection of the Y stator when i had
it apart, it would be easy to add another pair of diodes and measure the current provided by this phenomena

this testing thing can turn into a full time job!



btw, i don't think you will find any of the 3rd harmonic and 4th diode set in heavy duty alternators because they
are almost exclusively delta connected stators and have no center connection, and turn too slowly in use to have
much of a 3rd harmonic anyway.


bob g

[mobile_bob]

just took the time to connect the new transmission cord from the alternator to the remote trans/rect pack

turns out it is a 12/3 forty foot section of SO cable

it carries the load just fine, 13amps at 85volts under load
with a load of 14.5vdc and 85amps resistive on the rectifier

i measured the voltage drop on one of the 12 guage conductors and it measured .85vac drop over the 40ft

this is turning out to be a useful experiment.

i would have never guess i could generate and transmit power 40ft, and have 14.5 vdc at 85amps from an alternator
that would be rated at approx 85amps in stock oem configuration, and do it with a significant gain in overall efficiency.
doing so with a 12/3 equivalent to an extension cord rather than 4/0 battery cables and fight with getting enough voltage
to properly charge the battery bank.

in my opinion we are about 90% there, the next step would be to scale up to a larger capacity alternator and work out the details

the cost of this system for the crafty DIY'er would be a small fraction of what 4/0 cable would be.

this looks very promising in my opinion for those generating power some distance from their battery bank that is 12volts,
there is likely some gains with both 24 and 48volts as well but maybe not as dramatic as with 12volts.

Lloyd i think you can get the job done with a little work, and have a system that fills your needs.

bob g