Automotive 6 Volt Generator Transistor Voltage Regulator

[berntd]

Hmm, yes, a battery isolator switch then comes to mind. However, it is actually better for the battaery to have some small current flowing as it stops the plates from suphating.

 

[Danwvw]

Yeah, just build that bosch circuit.
I keep thinking about the https://www.mouser.com/datasheet/2/827/DS_UJ3N065025K3S-1530382.pdf Jfet still. Mouser has them in stock at $15 each. Maybe I will send for one and try on the 6 volt generator. I just took the 12 volt gen off of my test engine and replaced it with a 6 volt gen to try the housed 6 volt Electronic regulator on.
Actually a 5 volt regulator like the LM7805 might be choice better than zeners both power wise and regulation.
But to do the JFET above still need -15 volts. not too many solutions perhaps an isolated power module find those on e-bay Isolated +- 15 volt power module it's huge though!
03/01/2020 UpDate Just realized a 7805 won't work well on 6 volt regs. Probably be ok on 12 Volts.
Schematic:

 

[rjenkinsgb]

Hi Dan / Berntd,
still watching this thread with interest!

You can get around -8V to -9V from a MAX660 with a few extra components.
See this example:

https://www.circuitstoday.com/voltage-doubler-cum-inverter

Swap the diode polarities and common from 0V for negative out rather than positive with that concept & use schottky diodes for efficiency.

You could possibly add a stage or two of voltage multiplier to get -15 to -20 or so, using Pin 2 as the "AC" input?
The MAX660 itself is rated at 100mA so 10 - 20mA should hopefully be possible after 2 - 3 stages.

Voltage multiplier - Wikipedia

en.wikipedia.org

Another thought:
It could be possible to use the charge lamp circuit to turn power to the regulator on and off with the ignition, with a bit of extra circuitry?
eg. Connect it to a resistor and two series 1N400x diodes to ground, with another resistor across the diodes, so the whole circuit only takes a few mA, not enough to light the lamp - but puts 1.2V on the diodes while the ignition is on.

Then add a resistor from the top of the diodes to an NPN transistor base, that in turn switches a P type device and applied battery power.

Another transistor across the lamp series resistor (with its emitter or source to the top of the diodes) could be used to modulate the lamp brightness for feedback.

It's more complexity, but more overall functionality & without a flat battery.

 

[Danwvw]

Hello rjenkinsgb glad to see you posted. I was wondering about using a voltage doubler I have played with them a bit years ago but using 117 VAC. We need it to be fast to come up though! a few ms.
I could take another look at using the indicator light power to start it. I had it that way with the 1st prototype for a while it sort of worked and the Gen. light would just barely glow when on. (Isolating the internal supplies from D+ with the Oring Diode worked best for doing this but then the Gen. Lamp doesn't come on.)
The added connection to the cars ignition coil and Oring Diode with the MOSFET Design in prototype 2 actually works very well. And. That is on the car engine now.
Just mainly curious about using The 25m ohm Normally On JFET myself. Interestingly the 9m ohm Motorola MOSFET s I have been running will stay where I put them (On or Off) without a gate connection. Perhaps that could be used somehow.
But
I will have to draw out what your talking about with the indicator light power to see if I messed something there.
berntd/rjenkinsgb do you know anything about the Buck Converters? Could I make my own and reverse the diode to make negative or add more windings to make an AC supply I don't know what kind of device they use as a driver?
Berntd what are your plans?

 

[berntd]

Yes, a switching regulator can definitely create a negative voltage. I would look a LT Linear Technology devices for this. I think they (Linear) have also been taken over by someone else...

 

[Danwvw]

6 volt (Prototype 1) Housed inside Old Bosch Mechanical Regulator:

 

[berntd]

Great job!
You have a test engine on the bench!!
Mega cool actually!

 

[rjenkinsgb]

I could take another look at using the indicator light power to start it.

The new approach I was suggesting was to use that purely as a control signal to operate a separate power switch, rather than as power itself.
(Plus a transistor and resistor to separately control the lamp brightness).

You can use you original Mk2 circuit with no negative voltage or obscure parts, just an extra power feed diode from this.

A couple of quick scribbles:
The power switching concept - use the ignition-controlled lamp connection to switch battery power to the regulator on.

As long as the lamp feed is dead, there is no battery current at all - but the lamp feed being on can enable full battery power to the electronics without the engine running, so you always have full field control. The series resistor needs to be high enough value so the lamp is not visibly glowing from that current alone:



And as above with an extra transistor across the series resistor in that circuit (with the switch parts not shown) to recreate the original generator warning lamp function.
Using an emitter follower allows the lamp wire voltage to track the armature output voltage, but with the lamp current still drawn through the diodes so maintaining the "switch" control voltage they produce:

 

[Danwvw]

Yes thanks this makes since now. One advantage if it was made high current enough is your start circuit could start a generator that had lost its residual magnetism like the one in my car. (If it sets a week without starting it takes longer to start charging even with the MOSFET ON full. )
Much less current is needed if it just powers the supplies. I will try it on my next Prototype! Also a 5 volt regulator IC like the LM7805 or the LT1086-5 at least for a reference voltage.

Won't the forward bias P/N Emitter/Base drop. the collector so low it can't be used by the power on circuit? The Armature is effectively a Ground. The first circuit alone without a resistor to the lamp could light the lamp with 4.6 volts, that's probably enough.

 

[Danwvw]

Maybe This: (It will be like a VW Bosch Alternator circuit in that it will need the incandescent Gen Lamp Good to start charging!)

 

[Danwvw]

Need to find. some transistors for this. application. IF a 6 Volt lamp pulls 308ma at 4.4 volts then thats 14.4 ohms on resistance however these lamps. range from 2 to 5 ohms cold. A small sized high base current low gain etc... Bu2508 has been ordered for the NPN one. Its got high base current and a high VBEO rating.

 

[rjenkinsgb]

Won't the forward bias P/N Emitter/Base drop. the collector so low it can't be used by the power on circuit?

That's why the transistor has a base resistor, so it can turn the transistor on but not sink enough current directly drive the lamp or affect the sense circuit.
The bulk lamp current will be through the collector and diodes.

Edit - remember that has to be a PNP type.

 

[Danwvw]

My thoughts on the logic are Q1 NPN turns on the gen lamp and Q2 PNP turns on the Voltage Regulator Power.

Green=(Key on not started yet)
Blue=(Generator Charging)
Red=(Key off not running not charging)

Key on followed by engine started and the conditions before it starts charging.
Green 6 volts through the Gen Lamp through the Base Emitter of Q1 turns on the Gen lamp,
Q1 to conducts full on,
6 volt batt voltage flows in EB of Q2 and the resistors through CE of Q1 and into the D+ (gen armature) which is effectively ground because it's resistance is so low,
Q2 on because of its EB is forward biased thus powers up the regulator.

Next D+ goes from 0 to 7.4 volts as the generator starts charging
Blue (generator charging)
all the points in the start circuit go to about 7 volts and little or no current flows through it or the gen lamp. (The regulator now runs on voltage from D+ not shown here.)

Next the engine is turned off
Red (off)
Positive Batt voltage on the Emitter of Q2 reaches the Base of Q2 and Collector of Q1 but there is no current flow because of the NP junctions.

Thinking (500 to 1000 ohms) for R1 and (40 or 50 ohms) for R2 (Probably work without them though) Q1 BU2508 NPN and Q2 PNP MJD32C Transistors Ordered.

 

[Danwvw]

I wonder if this would resolve all the start-up issues Just one NPN BU2508 transistor?

 

[berntd]

Very interesting.

 

[rjenkinsgb]

OK, I missed the change in the first diagram, with what I was using as 0V being shifted to D+

Very neat, that bypasses the need for the extra transistor to control the lamp!

I'd still keep the two diodes to handle the bulk current from the lamp though, with the bypass and base resistors, as the full lamp current is likely to be too much for any normal transistor base.

Also, stay with the two transistor version and only switch on the control voltage - if you try to power D+ from the battery you will need a massive transistor and heatsink, it could take 20 - 50A or more, and it will be draining the battery all the time the ignition is on without the engine running...

 

[Danwvw]

Thanks for the replies. Yeah, I'll try it both ways to find out how they work.
The BU2508 should handle it mounted to the main heat sink used in prototypes 2,3 or even the B+ terminal of a mechanical housing. (Probably will have to add a external heat sink anyhow to the B+ terminal to handle the main schottky barrier diodes).
I calclaste this circuit to pull 3.125 Amps through the BU2508 and drop 1 Volt across it directly into the armature (Calculated the 1.6 ohm armature resistance using its rating of 7 volts at 45 amps but it's DC resistance could be much lower I should measure that to make sure first.) The Lamp is something like 2-5 ohms cold 32 ohms lit. I have some 2.5 ohm varistors , one probably should be placed series with the lamp.

 

[Danwvw]

Just measured an other 6 volt generators' D+ at 1.667 ohms using some batteries and resistors an the FLUKE.

 

[berntd]

Just check that the BVebo cannot get violated when the ignition is switched off to stop the engine while the system is charging. They give it as 7.5V min so it could be close if shutting off while revving.

The cable 15 side of the lamp will be at a low resistance to ground due to the other loads on that rail.

 

[Danwvw]

Wow your noticed these things! I did too. Yeah most transistors BVebo are like 5 volts the BU2508 though is like 8 volts one reason I chose it. Still probably be wise to put a diode in series with the Base?
Been playing with the ideas a bit and after close examination I realize the Single Q direct to D+ won't even work right.
rjenkinsgb I think your right about needing 2 transistors it might be possible to just power the supply with a 1 transistor start circuit?

This maybe, This is without the current limit, a little simpler. (The BU2508 is a TV Flyback driver transistor and has the. 33 ohm resistor and a reverse diode built in.)