Automotive 6 Volt Generator Transistor Voltage Regulator

[Danwvw]

Reversed ACS-773 pins 4 and 5 got it working and using the 4.7K on the ACS-773 output as a pull up to the 50 ohm resistor on B+ that also pulls up the gate. Current Drain is only 1ma.
It looks like current limiting is going to kick in a little too easy now!

LT-6700-I Prototype Spring UpDate with corrections above.

 

[KeepItSimpleStupid]

A capacitor between D+ and B+? e.g. 2 mylar 0.1 uF in parallel,

 

[Danwvw]

Yeah I could play with that. The 0.1uf D+ to ground doesn't really help I could move it.
One thing is, when I had the voltmeter on the Hall Effect output and tried the headlights etc. that Hall Effect output seems nice and tight, It would run an in dash ammeter display.

 

[KeepItSimpleStupid]

It would run an in dash ammeter display.

Remember it is ratiometric, but it probably doesn;t matter.

 

[Danwvw]

Just completed a 1:25 minute headlight drain then started the car and noted Dash voltages. It is for sure current limiting the voltage won't go over 6.25 Volts but it gains a 0.4 volts and builds when the headlights are turned off.

 

[KeepItSimpleStupid]

It is for sure current limiting the voltage won't go over 6.25 Volts but it gains a 0.4 volts and builds when the headlights are turned off.

Not sure what your saying, but the ratiometric to 1/2 the supply voltage could be causing it. Temperature dependent crap might haunt you.\The 3.3V chip is a regulator, but not a reference.

 

[Danwvw]

Yeah, The 3.3 volt regulator is great. I tested it hot, cold, Idle, revved, headlights on etc it's solid at 3.3 Volts.
The Hall Effect ACS-773 Ratiometric output is 1.6 volt at idle and drops nicely from there when the engine is revved. The access to reading it is under the VW voltage regulator cover and I did not measure it after draining the car battery for and 1 Hour and 25 minutes but my thinking is that only seeing 6.4 volts at the dash is because the 1.6 volt output is dropping below the 0.4 volt LT-6700-1 + input threshold. And it's operating nicely too because the 6.4 volts at the dash is nice and stable. Turning off the headlights reduces the current 20 amps and this brings the dash voltage up a little while the ratiometric output voltage builds as the dash voltage normalizes eventually reaching 6.8 volts which is it's normal operating voltage with lights on and a fully charged battery. Normal voltage lights off battery charged is 7 Volts. The 0.2 volt difference is the voltage drop across the cutout diode.
I need to test and make sure this design is not current limiting too much to recharge the low battery with the headlights on, It's kind of a draw back of the current limiting design. But the work around is simple just turn off the headlights for a couple of minutes to let the regulator catch up the battery enough to not be in current limit when the headlights are turned back on.
I think it's about right, as 1.6-0,4=1.2 Volts and 1.2/0.0264 tells me it's current limiting at 45.45 Amps.
Not quite sure how to adjust current limiting though. Thinking an additional pullup 4.7K resistor or perhaps a little less 3.9K from the Ratiometric output to the unregulated power supply.
Actually I am very pleased with how this LT-6700-I prototype is working.

 

[KeepItSimpleStupid]

pin 3 f the ACS, is low Z, I think
pin 3 of the 6700 is a high Z input.

You can use high vaue resistors for the voltage divider into the 6700. The battery monitor circuit on page 1 of the datasheet,.

 

[Danwvw]

780 ohms in reality based on voltage measurements.

 

[KeepItSimpleStupid]

Your schematic doesn't really show a divider? I didn't see an output Z listed for the sensor either.

 

[Danwvw]

The 780 ohm Z was calculated with the car off by measuring the voltage drop across the 4700 ohm resistor then calculating the current then measuring the voltage from LT-6700 input to ground then using the formula for R=E/I it turned out to be 780 ohms = input voltage/current. I did this because I needed to check to see what resistor would bring the input high enough to get it to start charging. The 4700 ohm resistor brought it up to about 0.8 volts. I forget the exact numbers, just that R calculated to be 780 ohms. I know it wasn't a Dynamic Measurement and is meaningless. I suspect the ACS-773 output Z may be somewhat lower. The Data Sheet did suggest 4700 load resistors, 500 ohms was given for minimum pullup resistor value to VCC 3.3 volts. The info is under the ACS-773 Data Sheet / Absolute Max Ratings / Output Sink Current /Notes
Further down the data under "Common Operating Characteristics" it list Rout to be 3.3 ohms.
But your right there is not a voltage divider, that wouldn't help it needs a voltage gain to increase the effective current limit. Probably just go with the 100 Amp ACS chip to do that. I think it's going to be ok as it is.

 

[Danwvw]

Guess I could voltage divide to the 3.3 volt supply to increase current?

 

[Danwvw]

Added 2 0.4F @4.5 VDC SuperCapacitors in series to the power at the ignition switched 6 volt where the Alarm Circuit is and it worked nicely to eliminate the noise which was like little crackling sounds the alarm circuit was making. in the 6.8 to 7,2 volt range. Note the car does not shutoff for about 1 second now when the key is turned off. The voltage reading however are a little higher. Just wondering though. Would it be better for the SuperCapacitor if it were on the Un-Switched side of the ignitions switch? It would then stay on all the time but may be harder on the Ignition Switch? Also just wondering, do SuperCapacitor's ever catch fire?
View attachment 121766

View attachment 121767

Just posted this link to Page 9 just to have it handy, Trying to troubleshoot my Dash Voltage Display it keeps loosing a couple of volts then it will be ok again, It's powering off of the switched 6 volt fuse that runs a lot of stuff. I think it's the alligator clip connections.

 

[Danwvw]

It was a rainy day so didn't get a chance to try my Idea. The ACS-773 Data Sheet shows it being used with an output filter and a by-pass Cap on the 3.3 volt supply. Doubt it needs the bypass because it's already got the 220uf on D+ and it's not a high frequency application. But the output filter might give me an added 5 amps or maybe even more power output from the Dynamo.
UpDated Schematic:

 

[Danwvw]

March 25th UpDate, Got it installed and working with the Hall Effect output Filter circuit above only I used a 15K ohm resistor instead of a 10K, instead of the 4.7k it had to start it. (Turns out that with the 10K calculations show current limiting would stop working if the Output Voltage went up to 8 volts or more.)
It's pretty well behaved. Starts Charging good. ACS-773 output at Terminal Block ie 6700 + Non-Inverting input:
Start-up = 0.8 Volts
Charging fast idle no lights fully charged battery = 1.6 volts
Headlights on revved = 0.9 to 1.2 Volts (Battery is fully charged)
Headlights on revved = 0.6 volts at first after leaving the headlights on with the car off for 10 minutes.
Seems to bareilly Build Dash Voltage charging with the headlights on after 10 minute headlight battery discharge.
Just need to Drive it and see how it behaves. Looks fine!

 

[KeepItSimpleStupid]

When your all done and tested: 1) Write up a theory of operation 2) Post the final design in Articals with a link here. 3) Use www.easyEDA.com and get a few boards made.

 

[Danwvw]

No plans to ever develop a board for these regulators, I just tinker with them for fun. The LT-6700-1 can use standard 6 pin breakout board nicely. I think I am pretty close to being satisfied with the design work with the 350 prototype being the simplest and best solution while the 6700-I prototype offers great current regulation at the sacrifice more voltage loss with the headlights on however with a fully charged battery it's almost par with the 350 Prototype.

 

[Danwvw]

New Problem with the LT-6700-I Design, Further testing actually about 40 miles of driving around, it's over charging the battery! Seeing 7.5 volts at the battery 7.3 at the Dash Headlights Off, Battery fully Charged, Engine Hot and revved. It's not regulating well at full charge and high RPMs. Voltage keeps climbing.
My thinking is it's related to the 220uf Electrolytic capacitor fix to the surging problem the LT-6700 chip seems to have.
I may try a 0.1uf cap on the - input to ground or even increasing the 100 ohm resistor to ground. A better Temperature Controlled 6.8v Zener may be needed. There could be any number of things going on to cause this though, maybe the 1n5817 Diode makes a poor EMF catcher across the Field winding, It could have excessive generator arcing at a brush or something whatever it is the LT-6700 could be seeing a little less voltage at the - input under these conditions. Zener?

 

[KeepItSimpleStupid]

Do you see an end in sight?

 

[Danwvw]

I thought I had reached it!
Today I tried a few things, did some voltage test and some load testing. It turns out it's only charging to about 6.8 volts and the rest is generator output noise which on a fully charged AGM Battery shows up at B+ at up to 7.5 volts DC. It only takes park lights to eliminate it. Not sure why it's happening. A 220uf Electrolytic Capacitor from LT-6700 + input to ground lowers the output noise reading to 7.25 volts dc at the battery. I plan to try a different flywheel diode across the field before switching back to the previous design prototype 350 to see if does it too.