# Change voltage circuit

#### Kev0511

Joined Jul 20, 2004
39
Hello all

Basically i'm trying to simulate a Gas tank pressure sensor. At atmospheric pressure the voltage output should be 2.5 volts and when the ECU go into the test procedure for testing the gas tank for leak, it should read 1.6 Volts. Now a Simple Voltage divider from a 5 volt source (LM317 / LM8705) can easily get a 2.5V signal, then add in a transistor to alter that Voltage divider to get the 1.6V isn't hard either.

The basic test procedure is as follows, the "inputs" are the inputs to the circuit, not the ECU, Also all the inputs will have a @33ohm pulled up to 12V resistor (for the ECU solenoid checks) hence why the 12 Volt rating.

I'm still working on making up a diagram (hand drawn and crappy) of how i think this may work, like using a latching circuit, that modifies the output voltage, is turned on when Signal 2 is pulled low, and is released when Signal 1 is pulled up again .

Been messing with this circuit to create what you see below But know i'm doing something wrong, or there is a much easier way, SCR maybe?

Thank you to all that looked

#### Alec_t

Joined Sep 17, 2013
14,237
I think you have a problem with the test logic.
In stages 1 and 3 sig 1 is low and sig 2 is high, yet you are expecting different output signals?

#### ScottWang

Joined Aug 23, 2012
7,397

#### Kev0511

Joined Jul 20, 2004
39
I think you have a problem with the test logic.
In stages 1 and 3 sig 1 is low and sig 2 is high, yet you are expecting different output signals?
Ahh yes, i've gotten that response before when talking to someone on the phone, but i'm also trying to simulate a test procedure. Maybe i can explain how the test is done so it makes more sense?

At a point in time, and OBD2 based car will test the evaporative emissions, and see if there is any leaks (a 0.02 and 0.04" leak test) after a number of parameters have been met.

Basically what the ECU does is close all vent valves, ie signal 1 above (Fresh air side of the carbon canister, fuel filler vent, ect) and will apply vacuum, ie signal 2 above, to the gas tank while monitoring a pressure sensor on the fuel tank, once the vacuum hits a certain amount, it will then stop applying the vacuum and just hold that vacuum for a pre calculated amount of time that was determined via Fuel temp, level, even consumption. once the test (s) pass, the ECU will open up the vent valves allow the vacuum pressure to drop and fresh air to enter into the system via the carbon canister/fuel cap

Hopefully that explains why there is two different values for the same "parameters" int he above flow chart / test logic

thank you
Kev

#### Kev0511

Joined Jul 20, 2004
39
Thank you for the drawing.

kevin

#### Kev0511

Joined Jul 20, 2004
39
Well i just quickly made up a circuit using SCR's and wondering what i did wrong? And if this would be a better choice then the Transistor way? If only i had a SCR to play with

thank you again.

#### #12

Joined Nov 30, 2010
18,224
This took me a bloody long time to think about, but here it is, in a crude way. This is an OR function until both inputs drop to zero. Then it becomes an AND function. That can be done by moving the switching point of the gate. I did it in this drawing by adding a bias current to the first transistor. If it's too picky to adjust, you can make the input a differential pair to reduce the loading on the left transistor input current.
Ponder this and see if I'm on the right track.

#### Attachments

• 9.2 KB Views: 9

#### ScottWang

Joined Aug 23, 2012
7,397
The below circuit was designed according to the table in #1.
How is the stage 4, does it just turn off the power?

#### #12

Joined Nov 30, 2010
18,224
Where input high and low are 12V and zeroV
and output high and low are 2.5V and 1.6v

Stage 1: Both inputs high ... output high.
Stage 2: Input 1 changes to low...output stays high.
Stage 3: Both inputs go low...output goes low
Stage 4: Input 2 goes high...output stays low
Stage 5: Both inputs go high...output goes high
Repeat.

#### ScottWang

Joined Aug 23, 2012
7,397
@#12.
Your numbers of stage are difference with the #1, your all numbers should be sub 1 as from stage 0.
So the next stage didn't show up is like reset, that's what your think?

#### #12

Joined Nov 30, 2010
18,224
I see. I didn't use the TS labels because I think differently from him.
He thinks stage one is when the initial conditions are first changed.
I numbered them differently because the output has to finally end at the initial conditions.
I can change my numbering...

a) normal (long term) conditions: both inputs high, output high
b) Stage 1...input 1 goes low, output remains high
c) Stage 2...both inputs low, output goes low
d) Stage 3...input 2 goes high, output stays low
e) back to normal conditions.

In the thinking of the first post, there are only 3 changes, but I see four changes.

#### ScottWang

Joined Aug 23, 2012
7,397
#@12, thanks.
I think the next stage as reset function was a good thought, otherwise the table could be mess up.

#12

#### #12

Joined Nov 30, 2010
18,224
#@12, thanks.
I think the next stage as reset function was a good thought, otherwise the table could be mess up.
Part of this problem is fighting the belief that there are only 3 changes.
There are 3 changes in the mind of the TS.
In a real car, there must be 4 changes. The output must resume initial conditions or the next test cycle will fail.

I am working on a car right now so I think in terms of the car.

#### ScottWang

Joined Aug 23, 2012
7,397
Part of this problem is fighting the belief that there are only 3 changes.
There are 3 changes in the mind of the TS.
In a real car, there must be 4 changes. The output must resume initial conditions or the next test cycle will fail.

I am working on a car right now so I think in terms of the car.
The input signal 1 just as the enable function for the logic IC, if the next stage that the input signal 2 going low then the table will be mess up, because no any possible like that, according to stage 2,the last stage is 2.5V, the output signal is from 2.5 V down to 1.6 V, but now the present stage is 1.6 V, so that's why I saying that the table could be mess up.

#### #12

Joined Nov 30, 2010
18,224
I am having trouble understanding you.
Use the table of the TS or use mine. You don't have to make them match. You just have to choose.
You can ignore my offer.
There are several methods suggested.
Post #6 won't work because output will not have a current path through the SCR when input 2 goes high.
I can't read post #8 because I don't know how a flip-flop works.
I did it in analog because that's what I'm good at.

#### Picbuster

Joined Dec 2, 2013
1,047
Maybe is missed a drawing but:
As you explained about stage 1 and 3 now you have to introduce a state counter avoiding misinterpretation. (1.6V or 2.6V at the same input condition).
The most and simple way: use pic12f675 or use counter to address gates evaluating counter against input values.
one output 1.6V+ VfDiode other 2.6V+ VfDiode and or with diodes. ( no need for dac)
inputs to be scaled down to 0-4V approx.
Define also number of stages.

#### #12

Joined Nov 30, 2010
18,224
The most and simple way: use pic12f675
Oh yeah! Programming a microprocessor is always better than using 2 or 3 transistors.
Can you list the equipment needed to do this with a PIC?

#### Kev0511

Joined Jul 20, 2004
39
Part of this problem is fighting the belief that there are only 3 changes.
There are 3 changes in the mind of the TS.
In a real car, there must be 4 changes. The output must resume initial conditions or the next test cycle will fail.
I am working on a car right now so I think in terms of the car.
Hello #12,

Yes one can technically say there is 4 stages but there is 3 stages of the test the ECU goes through to get the values it wants to see, which is what my list is basically going through, So technically the "normal/ No test" can be a stage itself.

Your schematic seems easy enough to try out, but how does the AND aspect work when one of the inputs is released, will the output be changed?

Also to get the 2.5V nominal and only 1.6V after Input 2 is triggered, i was using a 5V source (need to make up an 3.0V and a 2.7V output as while with this circuit by means of other voltage dividers) and also want a more stable power source so that the Value going into the ECU doesn't fluctuate when the Battery voltages swing from say 12.8 to 14.3 volts, and adding yet another Regulator seems to be just adding to the complexity of this circuit.

Also if i wanted to use a Micro controller, odds are i would try to use an ATtiny, since i have some, very limited, practice with Arduino. The thing is i like the idea of using discrete components / going analog

Thank you all for working through this for / With me.

#### Kev0511

Joined Jul 20, 2004
39
Post #6 won't work because output will not have a current path through the SCR when input 2 goes high.
Please explain, i know an SCR can be used a latching switch, typically on the ground side of a circuit. just a quick pulse of positive current on the gate will cause it to on. The 3904 being "on" that will allow a path the ground and when it's released the SCR will shut off, in turn return the output to 2.5V.

I'm far from an expert, but do love to dabble with electronics, and try to do as much research as i can but i know i fail more times and not.

Last edited:

#### JoeJester

Joined Apr 26, 2005
4,390
No to be picky, but, Vout in Scott's drawing would not be as labeled if there is 12V tied to R7. The voltage divider switches between 50% and approximately 33% of that voltage.

On edit:

I see the error is a typo as the voltage divider in the TS's second circuit uses 5 volts as the input to the voltage divider. I suspect it was a copy and paste error on Scott's second circuit. With a 5V input, the Vout would be as posted.

Last edited: