Modifying alternator for higher voltages?

[TitanDiddly]

I'm doing some research into various power systems and it often seems like it would be convenient to use an automotive alternator for charging. My system would need it to charge a string of batteries rather than just a single one, that is, 24v, 36v, 48v, whatever it is. I've done a little bit of research but I don't quite have the understanding yet to sort out what I'd need to do. I know that the windings of an alternator, if unregulated, can/will put out over 100 volts at their higher RPMs, so I know that the voltage in there. I also know that some part in there is responsible for taking that voltage down to somewhere between 13-16 volts. What if I wanted to say, double that voltage so that I could charge two batteries in series?

I presume that some/most alternators use a bridge rectifier for AC to DC, but that doesn't so anything for the voltage. (Right?) What regulates the voltage? Which parts would I need to replace? Where could I get them?

Any input on this would be great!

Edit: In the above, I'm talking about a single-wire, internally-regulated alternator.

 

[PottedMeat]

You would have to get into the voltage regulator on the alternator. The alternator needs a battery to work correctly - there are no magnets to generate a static field. Essentially the voltage regulator senses that the voltage is too low and uses the battery to supply current to the rotor. When the voltage reaches the setpoint, the regulator cuts back current to the rotor. I suppose if you got a schematic of the regulator you could put in a resistive divider from the output to the sense input to ground, by varying the divider you could get different voltage outputs.

Here's a little info:

http://www.tpub.com/content/co...14273/css/14273_50.htm

 

[TitanDiddly]

Originally posted by: PottedMeat
You would have to get into the voltage regulator on the alternator. The alternator needs a battery to work correctly - there are no magnets to generate a static field. Essentially the voltage regulator senses that the voltage is too low and uses the battery to supply current to the rotor. When the voltage reaches the setpoint, the regulator cuts back current to the rotor. I suppose if you got a schematic of the regulator you could put in a resistive divider from the output to the sense input to ground, by varying the divider you could get different voltage outputs.

Here's a little info:

http://www.tpub.com/content/co...14273/css/14273_50.htm

Thanks. I do plan to use this for battery charging, just with a string of them rather than a single one. So supplying the field current shouldn't be a problem. From what I've gathered so far, the voltage regulator more or less senses the power needed (by voltage, I presume?) and adjusts the field (current or voltage?) to match.

I found this page helpful: http://www.qsl.net/ns8o/chrysler.html
And I think this diagram would be helpful if I knew what I was doing: http://www.qsl.net/ns8o/altreg.jpeg
And his page doesn't link back to itself, so here's root: http://www.qsl.net/ns8o/welcome.html

 

[NeoPTLD]

The outer coils generate the power. It's a three-phase variable frequency(corresponding to engine RPM) alternator. Output is turned into DC using a set of diodes.

The output is fed into the regular and controls the duty cycle of commutator winding fed from a pair slip rings. I think it operates at 400 Hz or something in PWM mode.

The minimum full load RPM is the shaft RPM needed to maintain rated output when the commutator is at 100% duty cycle. You simply ought to bypass the regulator and directly feed DC 12v into the slip rings.

If you want to be able to regulate to specific voltage appropriate for charging a chain of batteries, it gets difficult. Here's a voltage regulator IC for automotive alternator.

Since Vsense and Vcc are not separated, I don't know how you can do it. If you can find an IC with separate Vcc and Vsense you simply need to provide about 12v to Vcc which will be used to provide power to commutator coil, then build a resistive divider which gives a ratio of v/n n=number of 12v batteries.
example, 48v nominal setup
Assuming your alternator is capable of making 60+V (you need 15v/battery), you'd set up a resistive divider using precision resistors, so the output is exactly 1/4 of battery voltage to make the regulator think it's working with a 12v battery.



http://www.st.com/stonline/pro...iterature/ds/11353.pdf

 

[PowerEngineer]

Why not just connect the batteries in parallel while charging?

 

[TitanDiddly]

Originally posted by: PowerEngineer

Why not just connect the batteries in parallel while charging?

Too much of a PITA for regular use.

 

[Navid]

Originally posted by: TitanDiddly
What if I wanted to say, double that voltage so that I could charge two batteries in series?
.

I have a problem with this concept!
What happens if one battery is fully charged and the other one is not?
Your system will detect a low overall voltage and send current to charge it. The problem is that the battery that is fully charged will be overcharged then.

 

[Navid]

Originally posted by: PowerEngineer

Why not just connect the batteries in parallel while charging?

Because they have different voltages!
He wants to charge them, not to watch fireworks!

 

[TitanDiddly]

Originally posted by: Navid

Originally posted by: TitanDiddly
What if I wanted to say, double that voltage so that I could charge two batteries in series?
.

I have a problem with this concept!
What happens if one battery is fully charged and the other one is not?
Your system will detect a low overall voltage and send current to charge it. The problem is that the battery that is fully charged will be overcharged then.

Don't worry, this concept happens all the time. They'll be all fully charged, then discharged/charged together so that they're never at a different state of charge. This is a normal way of doing things. Eventually, yes, due to manufacturing differences, the batteries will develop a slight imbalance, but this won't cause catastrophic failure on the first cycle or anything.
In fact, almost every battery you use is charged in series- your laptop has at least 3 cells in series(with a monitoring system), your car has six cells in series, RC cars have 6 or more cells in series, etc. As long as they're matched first, then charged and discharged together, they're fine. Dig?

 

[Navid]

You are right if the batteries are identical (Voltage, Amp.Hour, Age, temperatures experienced, humidity profiles experienced, ...).
That is the case for the six 2V units of a 12V car battery as you point out.

But, is that the case for your 48V battery and your 24V battery?

Do they have equal Amp.Hour ratings?
Do you always load the 48V battery with exactly twice the impedance you load your 24V battery with?


Edit:
It will not be a catastrophe or even a slight problem if you trickle charge them. But, then, it could take a very long time to change.
An efficient charger uses high currents first to bring up the voltage. But then, switches to trickle mode to avoid overcharging at high currents, which could decrease battery life. So, you get an overall fast charge as well as avoid overcharging at high current rates.

 

[TitanDiddly]

Originally posted by: Navid
You are right if the batteries are identical (Voltage, Amp.Hour, Age, temperatures experienced, humidity profiles experienced, ...).
That is the case for the six 2V units of a 12V car battery as you point out.

But, is that the case for your 48V battery and your 24V battery?

Do they have equal Amp.Hour ratings?
Do you always load the 48V battery with exactly twice the impedance you load your 24V battery with?


Edit:
It will not be a catastrophe or even a slight problem if you trickle charge them. But, then, it could take a very long time to change.
An efficient charger uses high currents first to bring up the voltage. But then, switches to trickle mode to avoid overcharging at high currents, which could decrease battery life. So, you get an overall fast charge as well as avoid overcharging at high current rates.

Yes, 36v and 48v single-string lead acid batteries have no trouble being charged at high currents. Just look at forklift batteries that last for years and years. I don't think you're understanding what I'm getting at - I want to have four deep cycle batteries permanently connected in series, and always charged together. Nothing to get out of phase. Golf carts have this exact setup, the only difference is that they use a transformer charger and not an alternator.

 

[Navid]

I see!
It was not clear from your previous posts, at least to me, what you were going to use the batteries for.

 

[DrPizza]

How about using a DC to DC converter?

 

[NeoPTLD]

Originally posted by: Navid
I see!
It was not clear from your previous posts, at least to me, what you were going to use the batteries for.

It's assumed that you use well matched batteries of same type, same age and preferably the same lot.

If you have a forklift with 4 sets of batteries and one dies after 6 years and you replace just that one with a new one, you can expect problems due to variations. They always recommend you replace all batteries together.

 

[TitanDiddly]

Originally posted by: DrPizza
How about using a DC to DC converter?

Well, it would need to be able to upconvert, which is difficult, and furthermore, handle around 60 amps.

 

[Braznor]

Come over to Off-topic. You will be given the answer in terms of lbs of beef!

 

[futuristicmonkey]

I suggest caution trying to get more voltage out of the alternator. There is A LOT of complicated math/physics that determines the voltage output of an alternator. If you over-excite the rotor in an attempt to get higher voltages, it will overheat and "die". There are two major considerations you must keep in mind for a project such as this: voltage and a frequency.

For electric machinery, an important value is the V/Hz, regarding rotor voltage and mechanical frequency. The lower the frequency, the lower the reactance of the coil: X_L = 2*pi*f*L. If you maintain the excitation voltage while lowering the frequency, the rotor will draw excessive current which is not proportional to the output of the alternator. Similarly, if you increase the voltage while maintaining the frequency you will have the same effect: there is not enough back-emf (back-voltage) to keep the current at a reasonable level.

There is a fine line between underexcitation and overexcitation. You'll want to research magnetization curves. Most electric machinery is operated at the so-called "knee point" of the magnetization curve, where you get the most magnetic flux per amp-turn of magnetizing force (in laymen's terms: the most bang for your buck). If you don't excite the rotor enough you'll have disappointing output. If you overexcite, you could (will) burn out your alternator.

I once had the notion to rig an alternator with my own voltage regulator, in the hopes of making a small camping generator. Turns out it wouldn't work -- there's simply not enough steel in an automotive alternator to output 120 volts at 60Hz with any appreciable current.

I hope this helps.

-ben

 

[DrPizza]

Originally posted by: TitanDiddly

Originally posted by: DrPizza
How about using a DC to DC converter?

Well, it would need to be able to upconvert, which is difficult, and furthermore, handle around 60 amps.

Yep. I didn't think about how many amps you had to handle though. Suggestion: maybe if you're really lucky... really really lucky, you can score some sort of upconverter at a junkyard from some electric car? (Do hybrids have an upconverter to go to higher voltages?)

 

[TitanDiddly]

Originally posted by: futuristicmonkey
I suggest caution trying to get more voltage out of the alternator. There is A LOT of complicated math/physics that determines the voltage output of an alternator. If you over-excite the rotor in an attempt to get higher voltages, it will overheat and "die". There are two major considerations you must keep in mind for a project such as this: voltage and a frequency.

For electric machinery, an important value is the V/Hz, regarding rotor voltage and mechanical frequency. The lower the frequency, the lower the reactance of the coil: X_L = 2*pi*f*L. If you maintain the excitation voltage while lowering the frequency, the rotor will draw excessive current which is not proportional to the output of the alternator. Similarly, if you increase the voltage while maintaining the frequency you will have the same effect: there is not enough back-emf (back-voltage) to keep the current at a reasonable level.

There is a fine line between underexcitation and overexcitation. You'll want to research magnetization curves. Most electric machinery is operated at the so-called "knee point" of the magnetization curve, where you get the most magnetic flux per amp-turn of magnetizing force (in laymen's terms: the most bang for your buck). If you don't excite the rotor enough you'll have disappointing output. If you overexcite, you could (will) burn out your alternator.

I once had the notion to rig an alternator with my own voltage regulator, in the hopes of making a small camping generator. Turns out it wouldn't work -- there's simply not enough steel in an automotive alternator to output 120 volts at 60Hz with any appreciable current.

I hope this helps.

-ben

Thanks for the good info- what if I don't care about frequency though? I mean, it's getting rectified to DC so frequency means nothing to me... unless of course it means something and I don't get it. I was thinking of just hooking up the alternator to the 48v battery and putting a dimmer switch in series with the field windings, and just hand-tweaking it all the time- it wouldn't be an automatic charger, but it would 'work' as a jury-rigged solution.

 

[futuristicmonkey]

Well, think about what is getting rectified into DC. A car alternator puts out 3-phase alternating current which is then rectified into the DC which your car uses. The frequency in this case refers to the frequency of the alternating current fed into the rectifer. This frequency is set by the rpm's of the prime mover. That reminds me, how do you plan on driving this alternator? Most lawnmower motors will goo right up to 3600rpm. If you connect the alternator via a gearbox or some other method of increasing the speed you will reduce the chance that you'll overexcite and burn out the alternator.

I may be making a big deal about something not so critical: the "speed" (rpm's) at which you drive the alternator. These things are made to function over a wide span of speeds and not burn out. Most electric machinery is not like this. The good thing is that this means you probably won't run into problems caused by overexcitation.

Your real issues are going to be how you actually decide to excite the rotor. You need some form of feedback or control mechanism to reduce the excitation when there is no load. If you don't, your voltage will spike when you remove the batteries being charged. Oh, and that dimmer switch won't work. It's made for AC only. One possible solution would be to control the excitation current via pulse-width-modulation. There are lots of tutorials around the web that can show you how to do this using only a few chips and a (big) transistor or two. Good thing is it'll only cost a few bucks for parts.

Once again, I hope this helps.

-ben

Edit: I forgot to mention that if you increase your output voltage you should increase your drive speed proportionally. I tried to imply that but I don't think it came off clearly.

 

[TitanDiddly]

Originally posted by: futuristicmonkey
Well, think about what is getting rectified into DC. A car alternator puts out 3-phase alternating current which is then rectified into the DC which your car uses. The frequency in this case refers to the frequency of the alternating current fed into the rectifer. This frequency is set by the rpm's of the prime mover. That reminds me, how do you plan on driving this alternator? Most lawnmower motors will goo right up to 3600rpm. If you connect the alternator via a gearbox or some other method of increasing the speed you will reduce the chance that you'll overexcite and burn out the alternator.

I may be making a big deal about something not so critical: the "speed" (rpm's) at which you drive the alternator. These things are made to function over a wide span of speeds and not burn out. Most electric machinery is not like this. The good thing is that this means you probably won't run into problems caused by overexcitation.

Your real issues are going to be how you actually decide to excite the rotor. You need some form of feedback or control mechanism to reduce the excitation when there is no load. If you don't, your voltage will spike when you remove the batteries being charged. Oh, and that dimmer switch won't work. It's made for AC only. One possible solution would be to control the excitation current via pulse-width-modulation. There are lots of tutorials around the web that can show you how to do this using only a few chips and a (big) transistor or two. Good thing is it'll only cost a few bucks for parts.

Once again, I hope this helps.

-ben

Edit: I forgot to mention that if you increase your output voltage you should increase your drive speed proportionally. I tried to imply that but I don't think it came off clearly.

Thanks for the good info. So far, I understand that the voltage(under load x) is governed by the amount of current(or is it voltage? power? I know they'll all change but which is the governing one?) going into the field coils. Is this RPM dependent also?

If the electrical load stays the same and the engine RPM doubles, will the field current/voltage drop to suit with new rotor speed?

Is the field current/voltage proportional? E.g. 0 field current/voltage for zero power output needed, 35% field current/voltage for a 35% load, 100% for 100%, and, at 100%, would that be the same as connecting the field directly to the battery that you're charging, or is there a current limiting resistor?

I'm thinking that if it's proportional, (as a test) I can rip out the field-governing electronics, apply 48v from the battery to the field and get a 48v charging voltage at rated current out. Correct or no? Is the field expecting AC? If so, does it need to be AC?

Also, what about self-exciting alternators? Could one of those be a 'plug' and play solution? I'd probably have to replace some of the components inside with higher voltage rated ones, no doubt.

Take a look at this, it's how I know that the dimmer switch works in at least some instances: http://classicbroncos.com/homemade-welder.shtml

 

[futuristicmonkey]

Good questions. I'll try to keep the answers concise.

It's the field current which actually excites the rotor. The voltage is what you vary in order to _change_ the current to what you need.

For all(well not ALL, but the majority of) electric machinery, if you need to excite the field it will be by using direct current.

If, for example, the excitation current was constant and the engine rpm doubled two things would happen:
1) The frequency of the output would double.
2) The voltage will double.

Now, since the AC coming off the stator is being rectified, #1 doesn't really matter.

The purpose of the voltage regulator is to increase or decrease the excitation current as one of a few things happen:

1) To maintain a certain voltage(really, power) at a lower speed, the excitation current will increase.
2) To get a higher output voltage(power) at a constant speed, the excitation will increase.
3) As more load current is drawn(at a certain voltage) the excitation needs to increase.

So, for example, if the engine rev'ed down to a lower rpm the automatic voltage regulator(AVR) will increase excitation to maintain the alternator's output (14-15V)Edit: At that power output level.

This is why you really want to have some automatic regulation circuitry -- if you have a large load connected and you suddenly disconnect it without reducing excitation current your voltage will spike!

The simplest solution would be to modify the regulator circuitry as NeoPTLD suggested:

Originally posted by: NeoPTLD
Since Vsense and Vcc are not separated, I don't know how you can do it. If you can find an IC with separate Vcc and Vsense you simply need to provide about 12v to Vcc which will be used to provide power to commutator coil, then build a resistive divider which gives a ratio of v/n n=number of 12v batteries.
example, 48v nominal setup
Assuming your alternator is capable of making 60+V (you need 15v/battery), you'd set up a resistive divider using precision resistors, so the output is exactly 1/4 of battery voltage to make the regulator think it's working with a 12v battery.



http://www.st.com/stonline/pro...iterature/ds/11353.pdf

Keep the questions coming!

Edit: What I really want to say is that excitation depends on speed and power draw of the load. If the load increases, the excitation needs to, as well. If the load draws less power, the excitation also needs to decrease, or else the voltage will increase. The inverse is true if the speed changes: lower speed, constant power => more excitation and vice versa. This latter case doesn't tend to happen at power stations, however, since the frequency must remain constant. So, at constant speed/frequency the excitation is changed according to the load.
-ben

 

[HydroBudz]

Out of curiosity, what are you trying to accomplish with 3-4 12v batteries in series?

Without modifying the alternator you could simply pick up a "Battery Isolator" to effectively charge 4 batteries off 1 alternator. It will individually charge each battery. They've been using these types of units in car audio and marine/boat use for years.

http://www.emarineinc.com/prod...rnators/isolators.html

I'm not trying to thread crap, but why reinvent the wheel?

 

[bobsmith1492]

Originally posted by: futuristicmonkey

This is why you really want to have some automatic regulation circuitry -- if you have a large load connected and you suddenly disconnect it without reducing excitation current your voltage will spike!

Even when regulated, disconnecting the load causes a transient before the system can react because of the high inductance of the coils. The output current tries to stay constant, basically, so the voltage flies waaay negative. When you disconnect the battery in a car while running, it causes what's known as a "load dump" spike. At work we're designing circuitry for car engines and it has to withstand a -600V spike for just this situation.

 

[TitanDiddly]

Originally posted by: HydroBudz
Out of curiosity, what are you trying to accomplish with 3-4 12v batteries in series?

Without modifying the alternator you could simply pick up a "Battery Isolator" to effectively charge 4 batteries off 1 alternator. It will individually charge each battery. They've been using these types of units in car audio and marine/boat use for years.

http://www.emarineinc.com/prod...rnators/isolators.html

I'm not trying to thread crap, but why reinvent the wheel?

It's a 48v system. I'm not just trying to charge 4 12v batteries at once just so I can take them off and use them separately, it's a complete system. Think golf cart or forklift.

futuristicmonkey-thanks for all the info, I just got back from camping and I'll get back to you with some more questions later.