Generators & solid state regulators

I know for motorcycles there are solid state regulators also made to do away with the old point systens used on the Lucas generators / regulators used on pre altenator British bikes

Coke-in-MN wrote: I know for motorcycles there are solid state regulators also made to do away with the old point systens used on the Lucas generators / regulators used on pre altenator British bikes

Ah, and that brings up another option for the "little" generators like on the "B".

Accel and a couple of others make solid state regulators for the big Harley and American V-Twins. I haven't done it yet, BUT: Harley used a Delco, it was an "A Circuit" like Allis uses, and various models produced 10-20 amp, also in line with Allis.

The really good part is, if true - Accel says the have a unit for ALL Harley models - so 6V may be an option with them.

Considering that a flaky regulator will quickly cook or sulfate a battery - the extra up front cost - plus longevity gets paid back fairly quickly.

Cool story, Gerald!

While not for an A- or B-circuit, I was faced with an unobtanium circumstance with a Suzuki GN-400 motorcycle that I'd converted to a lightweight cafe-racer used for emergency transportation in my service truck.  It's a kick-start single, had a 6v weenie-battery to make the lights and horn work, but the reg was fried, battery long gone, and I wasn't on a $300 replacement budget... so I made a replacement 'shunt' regulator (solid state), set it up for 13.8v, and changed the head and taillights to 12v (a GL1000 headlamp bucket looks pretty cool on that little thumper), pitched the turn signals, and called it good.  Kick it hard, rev engine, got 12v, and that's that.

My D17 seems to be charging awfully high, and it's a worker, so it'll get the solid-state kit.


Edited by DaveKamp - 29 Jul 2010 at 12:00pm

Maybe some day I'll post the circuit of my steam turbine regulator. It would work with a negative ground generator where the field is connected to the positive output terminal and the regulator grounds the field for maximum output. And a variation on a circuit I published last week in the proceedings of the Central States VHF Conference (met in St Louis last weekend) in an article about Power Amplifier Protection would work when the field is permanently grounded. Though that circuit takes two MOSFETs and a little IC, plus a couple zener diodes to protect the MOSFETs and would need a big diode to protect the main FET from the inductive kick of the field coil. As a simple regulator for a DC supply the circuit has been tested. It has possibilities anyway.

Gerald J.

Are air cooled VW Beetles with generators A or B circuit?  I do know the Bosch makes a solid state regulator for the VW generator.  I believe it is made in Mexico, about $40.  Considered trying one for an old inboard/outboard boat I own, bet switched to alternator instead.  Below is a website with info.

http://hillmanimages.com/912/voltage_regulator.html

In my slow speed analog MOSFET regulator circuits, I protect the gate insulation of a normal level MOSFET with a 12 or 15 volt zener diode, cathode to gate, anode to source for a N type, opposite for a P type. The normal level MOSFET can only stand 20 volts gate to source, then it breaks down the insulation and wrecks the FET. But it only takes 8 to 10 volts gate voltage to saturate the MOSFET with the source to drain showing .02 volts drop typically. Any more gate drive doesn't improve the saturated voltage drop. Recently a fellow tried my new circuit (uses a P type MOSFET for the pass transistor) and didn't understand that requirement though I told him repeatedly in a 42 volt solar regulator/switch and every time he put -42 volts on the gate with respect to source, his MOSFET shorted.

There is quite a bit of gate capacitance and part of it is coupled to the drain with its wide swings so it takes a small value resistor gate to source to turn the MOSFET off rapidly.

As for the VW regulator, it was 1969 when I built that one using bipolar transistors. I don't remember the circuit details any more. My second one was for a Pinto wagon alternator, probably about 1977 and it used a UA723 regulator chip and a transistor. My modern regulators use a TL431C for the regulation and MOSFETs to interface to the power supply or generator or alternator field.

The Delco alternator S-10 with internal regulator that I put on my 4020 came from the local auto electric shop, custom built with the proper width pulley and to make it work easier the smallest available diameter pulley. Still takes about 1600 engine rpm to start charging then it charges down to 450 rpm idle. I added an ammeter (using #10 silicone and fiberglass insulated wire to handle engine temperatures and because I had a roll) so I can tell what its doing. The generator and regulator bounced the ammeter from peg to peg so I couldn't tell what it was doing. The alternator just charges as needed  and tapers properly.

Gerald J.

Gerald & Dave,

Because I am a lazy person intellectually, I went the path of approximating what a mechanical regulator does. I viewed it as a combination chopper/DC-DC design, operating between 100-400hz (not khz). More than one source (three if you count my eyes) confirms this.

Because of this, a mechanical regulator is not required to sink any current, which allows the contacts to live.

The single biggest issue in regulator design at the 20 amp & higher is current limiting to protect the generator. It's output is fairly linear until the ferrous elements reach saturation.

Who doesn't like the 2n3055 and the 555? So that's what I started with - but since I was shooting for 1 device to provide voltage AND current protection - the 2n3055 came up short.

An IGBT gets around most issues of shorts, spikes. That's what Toyota uses to feed/control the Prius motor/generator.

Anyways (whew) mechanical generators sink very little current, and that's what I was hoping to do.

Regardless, someone has done all the hard work - and I'm always in favor of that.

Hey post yer circuits - I'm sure they'd be novel and educational.




Edited by CJohnS MI - 30 Jul 2010 at 5:47am

Butch, AMEN!

There's a web site run by a fella who's passion is scanning every darned Chebby manual he can get his hands on. I pulled down the complete Delco set covering generators, starters, distributors and coils.

http://chevy.oldcarmanualproject.com/electrical/49dr324/index.htm

For those who find it difficult to use/save the files to their computer, I can bundle them up in a zip file and e-mail them.

Battery chargers have come a long way too. For years, from childhood on, I kept a charger that used a rectifier that screwed into a lamp base, wire clipped on top, two large black painted wooden knobs to set amperage and how many six volt batteries were strung onto it.

And it always ALWAYS would bite me when I handled it - so had to unplug it before doing anything. But it always worked. Sadly, I have moved too many times and cannot remember its fate.

Delco seems to think little generators don't need current limiting, big generators do. That's the only criteria I have to go on.

A 2n3055 class transistor would easily handle the switching rates a generator needs - as per Delco - range was 100 to 400 hz.

You are right - that's exactly what it was. The original rectifier died after 30-40 years, I got a metal-shelled replacement - could no longer see the glow from the house coming from the equipment shed (I was always "fixing" dead batteries the neighbors gave me).

Gerald, you ought to lend your talent to a new & improved design. I just discovered this device by National called a LM5118 & will tinker with a 1/2 dozen or so (how many I fry generally).

I tend to stay in Dave's Kamp - simple is good, simpler is better.

Shoot, Brian... we're gonna hafta build one using a 6L6 and an 813...

One transmitter I worked on at Collins used a pair of 4CV100,000C for the final. They'd work as regulators, but the filament power would be more than the tractor could supply. Maybe a 304TL would do, it was designed as a high current pulse modulator for WW2 radars.

I looked at the LM5118, my first reaction is that its probably overkill, I'd think only using it in bucking mode setting the clock frequency to a few hundred Hz and using the field coil on the alternator for the inductance. Might have to drive it from the external clock input to get it that slow. Since one side of the field coil of the alternator is usually grounded you couldn't get to boost mode with that for the switching inductor.

But if you used the typical few hundred kilohertz switching frequency and the small inductor it needs, then you could use boost mode probably only for starting the charge when the engine is first started. You'd drive the field coil with the output, but sample the battery voltage rather than the regulator's output voltage. What that would do for the loop stability will take experiment and probably slowing the transient response by a factor of 100 or 1000 or more to keep things from oscillating.

In my TL431C plus FET regulators I had to put a feedback capacitor around the TL431C to keep it from being faster than the power MOSFETs then I put a big capacitor (6000 mfd) on the regulator output to supply faster load changes, with a 10 mf tantalum for faster load changes and a 0.1 monolythic for the really fast load changes. With those I can't see load changes with my 200 MHz scope. Probably over kill and then some.

Brian, you are right generally, though I just used a big diode in parallel with the field to handle the inductive kick. It has to be rated to carry the maximum field current. You could use a 555 for switching clock and as a voltage comparitor. I did that once in a computer power supply that was a preregulator so it didn't need precise voltage and I ran it open loop. I got plus and minus 15 volts from a 9 cell gel cell battery that ran 21 volts when charging and 17.5 under load. That computer ran 18 years with only 40 hours down time total. Most of the down time came from a power supply failure where  wire got pulled loose under the chassis. I had to replace the batteries once without powering it down. You might want to use a dual 555 called a 556 so you can clock and compare easily at the same time.

I like the TL431C for that task because it has a built in reference. Its sold as a programmable zener diode, but its really an open collector op amp supplied through that open collector. With that internal reference it compares to the "control" input. When the voltage on the control input exceeds 2.500 volts it pulls down on the output. It acts like a zener diode rated at 2.500 volts if the output is tied to the control pin. Except that unlike the sloping knee of the low voltage zener its voltage holds constant from less than 100 microamps up to 20 or 30 milliamps. If you put a resistive voltage divider on the control pin you can get higher voltage zeners up to a maximum of 35 volts. There is one in every computer power supply sampling the 5 volt output and sending controls back through an optical isolator to the line side switching regulator. A very popular chip.

NiCads in a battery pack have different capacities. Especially if one leaks some electrolyte. Then when the pack is discharged, the weak one gets charged with reversed polarity. That makes the metals grow whiskers and those whiskers are the shorts. Sometimes you can burn the whiskers off and have a good battery pack again, but the root cause, that weak cell, will limit the capacity of the pack and will short again. I don't think its worth the bother though I've done it many times. I prefer to open up the pack and use a good single cell with the same ratings as the one that is shorted, then I'm less likely to make it expand uncontrollably (with a bang) because the energy I apply is limited.

So long as the regulator regulates the battery voltage it makes up for the voltage drop in the diode and the diode makes a better cutout than the best relay cutout. I found that with my VW regulator with diode cutout back in 1968 or 1969. Since that diode drops as much as a volt at 30 amps, it needs a good heat sink to keep it cool. Can't be just the wire leads to keep it cool. Unless they are made of 2" copper water pipe a couple feet long.

I believe Steve in NJ sells diodes for cutouts packaged to look like relay cutouts. They would work superbly with a current regulated 3 brush generator.

Gerald J.

Gerald, Dave Brian,

You guys keep it up, ya'll might just come up with sumpin.

Would it be nice if a "B" owner could finally have a generator that didn't cook/kill the battery?