Automotive 6 Volt Generator Transistor Voltage Regulator

[Danwvw]

Wait! Yes the AD22057 is working now I wired it more as suggested and with a 10mf cap on the A1, A2 inputs to ground and an output gain network like in figure Figure 3 of the data sheet. Still I am confused about how to wire the inputs as they seem to be responding to noise rather than DC.

 

[Danwvw]

I think the AD22057 only functions with an AC input signal.

 

[rjenkinsgb]

From the datasheet, it's fully DC coupled, it should work fine?

 

[Danwvw]

Well I soldered a 1.3 volt battery across a 250 ohm pot and connected the (+ and - inputs) of the AD22057 one to the wiper one to Neg of battery and pot with the rest wired like figure 10 (Current Sensing Interface) Even loaded the output at 1k and did the low pass filter with .22mf to ground off of A1, A2 junction. With a 100K value for R. It did not measure the DC however would go full scale DC output for a moment if I touched one of the inputs. The .22 mf made it hold the DC output it produced for just a fraction of a second while the 10 mf was holding it longer. It behaves like a Peak AC detector that converts it to DC X its Gain.

 

[rjenkinsgb]

That device actually has an internal feedback network on the second amp, which is not obvious from a quick glance at the data sheet.
Just connecting A1 &A2 with no external resistor gives a fixed gain of 20.

To use with your single cell test, I think one side of that needs connecting to V+ or the offset terminal would need a bias voltage, to keep everything within the overall common mode range

 

[Danwvw]

What Connect +Vs to an input? I guess it will be that way on the car. I will try that and with it set to a gain of 20.
It this doesn't work maybe there is a way to use the LM311 Comparator to do this job of interfacing High Side current across the (6 volt generator electronic voltage regulator) shunt to the (el cheapo digital amp readout). AD22057 For some reason this Electro Tech web site is very slow today!

Like this?

 

[rjenkinsgb]

Yes, that's what I was thinking of.
From the internal schematic, it appears to need a voltage on the inputs to provide some voltage at the first opamp, so that can function.

 

[Danwvw]

Ok found my problem I think! I had an input pin on the IC not soldered. Responding to DC adjustments Now!

 

[Danwvw]

Now I just wired up the test circuit with the repair and it's working right!
Here is a revised test circuit drawing of what I did for the test. The Gain Set to 1 now by eliminating the 100K output gain resistor and reducing the default gain of 20 by adding a resistance of what turned out to be about 4K ohms. between stages to ground.
(Note it's not drawing much current it's just that the 9 volt battery is running a little low on voltage) Also a LED was added to simulate lover supply voltage than input voltage.
It is ready to test on the car with the display. Hoping to use the 4K gain adjustment to calibrate the current readout.

4K = Gain of 1

 

[Danwvw]

After installing it on the car it doesn't work, Same voltages it was reading the first time on the car.
However I think I have figured out the problem. It's the el-cheapo digital current display! When powered up it produces 3 volts on it's red current sense wire, It can be shorted out to make the amps read zero however it takes a much lower impedance than the 1K ohm output of the AD22057 chip.
The e-cheapo digital current display circuit needs it's shunt and probably 50 amps to drive it.

 

[rjenkinsgb]

When powered up it produces 3 volts on it's red current sense wire, It can be shorted out to make the amps read zero however it takes a much lower impedance than the 1K ohm output of the AD22057 chip

Doh... That's an odd design.
Is it worth pulling it apart to see if there are any input resistors that could be removed?

 

[Danwvw]

Yes, My statement above was wrong. I had removed the 1 K ohm load because of something I read in the AD22057 Data Sheet.
I have removed it all from the car and have the digital ammeter here and am testing it more with the AD22057. Something I should have done earlier.
It turns out that the 1 K ohm is enough to get that 3 volt input voltage (on the digital ammeter) down to 6 mv which is reading just 4 amps on the digital display with it's internal pot down all the way.
I think it's doable just need to come up with a resistor network that makes it match and span properly. The digital current readout reaches its full 50 amps with just 77 mv across the AD22057 outputs 1K ohm RL resistor. The span need to be figured out so that it can handle over 140 mv which is with the headlights on.

With the voltage I had the regulator set to the shunt was reading 60mv without Headlights and 140mv with headlights. So If I figure the difference 140-60=80mv is the amps of the headlights which is 20 amps for the pair of 60 watt headlights at 6 volts. I should be able to figure this out?

80 mv / 20 Amps = 140 mv / ? (cross multiply and divide) answer 35 amps with headlights so it's 15 amps without headlights. I think?

80mv/20Amps=?/50Amps (cross multiply and divide) answer 200mv would be full scale 50 amps. I need to reduce the gain and offset the 4 amps.

 

[rjenkinsgb]

Rather than reduce the gain, can you use a resistive divider between the IC and meter? That will reduce the offset and lower the impedance seen by the meter as well.

 

[Danwvw]

Good Point! I should have thought of that first. Just went to 2K on the Gain set resistance between stages. The AD22057 has a gain of 10 in the 1st stage and a gain of 2 in its second stage. I could figure out how to use the AD22057's built in offset feature I suppose? Yes 4 amps offset though with 0 volts into the AD22057 . Voltage wise it's right at a gain of 1/2 with the 2K instead of 4K's gain of 1. My volt can't even measure the offset it's so low less than a mv.
This quote from page 3 of the data sheet in the offset discussion concerns me and it's way I wanted to run it without the 1K RL: "Also, the output can run down to within 20 mV of ground, provided it is not called on to sink any load current"

 

[rjenkinsgb]

provided it is not called on to sink any load current

A resistor to ground means the output is sourcing current - that helps reduce the minimum output.

If the minimum low output is a fixed value, having a higher output gain / range then dividing down, will minimise any stray reading.

 

[Danwvw]

I see, Yes, it's going to have to be on the output isn't it. I did a test where I reversed the 1.3 volt test battery that supplies input to A1 the first stage of the AD22057 and it doesn't go below that same 4 amp reading. I guess it makes since as the AD22057 is a single supply device. Well I think? It does say in the datasheet its inputs can operate below ground? Maybe a slight negative voltage needs to be supplied at the output, I will try it.

 

[Danwvw]

Yeah, That works! I tried some things that didn't work, like a diode between the ground of the digital ammeter display etc.. that caused a lot of drift.
Anyhow I went to a gain of 10 which I found with a 200K ohm pot set to 71K between The A1-A2 connection and ground.
And for the voltage divider on the AD22057 Output I just used a 10 turn 1 K pot set to 961 ohms Between the AD22057 output and the Digital Ammeter input and the ground leg of the pot is measuring 39.8 ohms. This is with a 1K ohm resistor across the input of the digital ammeter.
Digital Ammeter current reading will now go down to about 1 amp.
with 200 mv input to the AD22057 it's putting out 2 volts which thanks to the 1K ohm pot makes the digital current meter read 50 amps.
Checked a few other voltages to make sure it was linier. Looks good.
0mv = 1 amp
60mv = 15 amps
80mv = 35 amps
200mv = 50 amps

 

[Danwvw]

Some Photos:

Idle:

Revved Lights Off:

Revved Lights On:

Forgot and left the lights on for 45 minutes then started it and it goes up to 50 Amps if revved. The Digital Ammeter Pegs at 55 amps. There is about 3 amps noise though!

 

[Danwvw]

Today I installed the Boost Converter which only worked for a few minutes before burning out. It didn't help much and would not start unless I removed all the protection diodes etc I had on it. I learned some things though. I don't think it's the way to solve the start charging problem however I believe the right chip and inductor would work. I had it on the DF the field but the field is a very complicated signal to condition without any voltage loss as well as keeping it below the boost converters absolute maximum ratings mainly that input must not exceed output!.

Photo before removing the Boost Converter: (Note! The little blue transformer is a choke I tried on the feed to the Un-regulated however it was removed in favor of the working Prototype 2 schematic using the STPS oring diodes with feeds to D+ and the 6 volt ignition coil power for startup!)

 

[Danwvw]

There has been a new problem! Forgot and left the key on while it was not running. It sat for an hour that way before I discovered it. When I went to put the car in the garage I noticed the voltage was too high. Sure enough it Generator was receiving a signal for Full charge. The Mosfet was locked on.
So this morning I started to troubleshoot why.
MOSFET Ok,
E7 Test Point stuck Hi. Schematic
Turned out it was intermittent too. I could wigle the current pot etc the current Ic that part of the circuit board and it would start regulating. Went over all the connections etc.. Now it won't act up!
This leaves me with the concern that there needs to be some kind of built in fail safe against circuit malfunction, Needs to be an alarm or a generator shutdown or something I can count on so it can't burn out the generator.