Cool, I understand the three-resistor voltage divider network thing that sets Centre and Width voltages! Instead of arbitrarily messing around with the resistor number I can work it out instead. Cool!

 

I am aware that I'm constantly talking to myself here, but.. I just had a thought.. since I can see the relevance of a voltage divider network with these op-amp comparator whatsits, it ought to be pretty easy for me to configure a quad-comparator to do what I want. I can make up all the voltages that I need for all the vREF high/lows, without needing any of this Centre/Width stuff..

I'll give it a whirl tomorrow. It's 2am now

 

OK, sorry - the finale' of the television series "LOST" was on, and that preempted doing anything on here.

If you can figure it out, then be my guest.

However, my recommendation still stands - run the compressor anytime you have high engine vacuum, and the evaporator could use some more pressure to produce cool air.

If you add a requirement for medium RPM's in the mix, you will wind up being a speeder, or revving your engine up at a stoplight, just so that you can get cooling - which will be counterproductive.

 

Yes, the medium-RPM thing won't realistically ever happen. The RPM window is going to be IDLE -> 4,800 (max safe for compressor). So basically, it'll just disengage the compressor if the engine bogs down badly or the engine isn't actually started or something.

In reality what I want to do is cause engine RPM to raise slightly when the compressor is engaged, but I'm not sure how to do this. Since I am monitoring the throttle position opening, and it's a drive-by-wire motor, I may be able to pull the DBW signal down a little to cause RPM to go up (0.2v - max throttle, 4.8v = throttle closed), but I'd want this pull-down to only happen when there's no throttle. I wouldn't want to pull-down all the way across the throttle operation.

 

You have to ask yourself, "What might happen if my circuit failed?"

Toyota did not do this enough, and as a result, many people have been injured or killed when their engines suddenly went WOT for no apparent reason.

For obvious reasons, we are reluctant to discuss circuits that even monitor existing engine control signals, as malfunctions of circuits may result in safety issues.

 

You have to ask yourself, "What might happen if my circuit failed?"

Toyota did not do this enough, and as a result, many people have been injured or killed when their engines suddenly went WOT for no apparent reason.

For obvious reasons, we are reluctant to discuss circuits that even monitor existing engine control signals, as malfunctions of circuits may result in safety issues.

That's a very good point, and surely good enough for me to forget about any kind of throttle position signal fudging. I will talk to the LS1tech / V8RX7 guys and see if there's a safer way of telling the ECU to raise throttle, or otherwise tell it that AC is ON. In the older LS1 engines it was all easier because the ECU did AC as well, but on the LS2 engines, the AC has been moved to the Body Control Module (secondary, body/chassis ECU) which I don't have, and in fact nobody ever has when they do these engine swaps.

 

Nope, I'm struggling. I found a quad comparator in LTSpice - the LT1721 (http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1154,C1004,C1012,P1707 ), but there are two things I don't understand. Firstly, it just comes up as a basic comparator symbol in LTSpice.. it certainly doesn't show four inputs. Second, whist the datasheet does indeed list four separate inputs (and another four inverting of the same), I don't see any kind of "vRef" pins. I was sort of hoping for 4xvRef and 4xInput, and that the chip would compare input against vRef, for each of the four inputs.

I guess it's just not that simple is it ? Any hints?

 

I think this might tell me what I need.

 

I get it all now. I'll post my results for you to look at later if that's OK

 

Here's my first part.. it goes ON for "over 4v". I know I should have a resistor with the LED. I just had to put the LED there because LTSpice wasn't showing me a voltage on the Output of the comparator until I put a route to ground there (via the LED)..

Quick question.. what's with the ripple on the 4v ref/input? Because my signal (at +in) is a sine wave, it seems to cause a similar sine wave ripple on the inverting input which I am using as the Ref input... (green line) See below:

Green should be constant 4v (from the voltage divider network), but there's a ripple in it.
Blue is my sine wave signal/sensor output, 0 - 5v, 1Hz
Red is the circuit's output.. should go ON when the sensor/blue line is above 4v, and it does too

 

You're using REALLY high values for your resistor divider network, and no stabilization cap.

Usually, you'd want somewhere around 0.1mA to 2mA current flow through the resistor network, so that it doesn't take too long to establish the reference voltage, but at the same time it doesn't use a lot of power. 1K per 1v of Vcc usually works pretty well.

See the attached.

 

Nope, I'm struggling. I found a quad comparator in LTSpice - the LT1721 (http://www.linear.com/pc/productDetail.jsp?navId=H0,C1,C1154,C1004,C1012,P1707 ), but there are two things I don't understand. Firstly, it just comes up as a basic comparator symbol in LTSpice.. it certainly doesn't show four inputs.

That's to make it easier on the person who's creating the simulation.

For example, you didn't need to use all four comparators in your test; just the one.

Secondly, when an IC has a lot of inputs and outputs, it gets to be difficult to determine the best placement for the I/O pins. If they were all in a big rectangle plunked down in your schematic, you'd have to run wires all over the place to connect things up - or use lots of NET labels. Having just the basic comparator symbol with it's power, ground, inverting and noninverting inputs and output gives the designer the flexibility to put them where it would make the most sense in the schematic.

Second, whist the datasheet does indeed list four separate inputs (and another four inverting of the same), I don't see any kind of "vRef" pins. I was sort of hoping for 4xvRef and 4xInput, and that the chip would compare input against vRef, for each of the four inputs.

There are some comparators that do have built-in voltage references. However, they are not hard to make using resistors or a pot, and a small cap.

I guess it's just not that simple is it ? Any hints?

See the post I just made.

 

You're using REALLY high values for your resistor divider network, and no stabilization cap.

Usually, you'd want somewhere around 0.1mA to 2mA current flow through the resistor network, so that it doesn't take too long to establish the reference voltage, but at the same time it doesn't use a lot of power. 1K per 1v of Vcc usually works pretty well.

See the attached.

Fab, that looks great, and I'm really pleased that I understand it too

Why 2.4k resistors and not 2.5k ?

 

Fab, that looks great, and I'm really pleased that I understand it too

You can download the simulation to your LTSpiceIV directory and play with it.

Why 2.4k resistors and not 2.5k ?

Good question. Now why would I have done that?

Because 2.5k is not a standard value of resistance!

Standard E24 resistors jump right from 2.4k to 2.7k.

Here's a table of standard resistance values:
http://www.logwell.com/tech/components/resistor_values.html
Bookmark that page.

Now you say to yourself, "That silly Wook; he could put any value he wanted to in there!"

Well, that is true. However, frequently people want to actually build something that was the result of having been prototyped and evaluated using SPICE. So, why would you start out building a simulation using values that are not standard or obtainable?

If you stick with standard values in your simulation, going from a simulation to a breadboard circuit will be a whole bunch easier. Otherwise, to get 2.5k you would have to use combinations of resistors in series and/or parallel.

Speaking of which, here's a very handy online calculator for doing just that:
http://www.qsl.net/in3otd/parallr.html
Bookmark that page too. Try it out; select E24 values, and then enter 2500 for the resistance. It will show you various combinations of resistors in series and parallel to get close.

However, since I started off with standard values - it's unnecessary.

 

You can download the simulation to your LTSpiceIV directory and play with it.


Good question. Now why would I have done that?

Because 2.5k is not a standard value of resistance!

Standard E24 resistors jump right from 2.4k to 2.7k.

Here's a table of standard resistance values:
http://www.logwell.com/tech/components/resistor_values.html
Bookmark that page.

Now you say to yourself, "That silly Wook; he could put any value he wanted to in there!"

Well, that is true. However, frequently people want to actually build something that was the result of having been prototyped and evaluated using SPICE. So, why would you start out building a simulation using values that are not standard or obtainable?

If you stick with standard values in your simulation, going from a simulation to a breadboard circuit will be a whole bunch easier. Otherwise, to get 2.5k you would have to use combinations of resistors in series and/or parallel.

Speaking of which, here's a very handy online calculator for doing just that:
http://www.qsl.net/in3otd/parallr.html
Bookmark that page too. Try it out; select E24 values, and then enter 2500 for the resistance. It will show you various combinations of resistors in series and parallel to get close.

However, since I started off with standard values - it's unnecessary.

Thanks very much. I have bookmarked those sites, and have just put 2500Ohms into the resistor calculator. I hadn't considered what running resistors in parallel would achieve. Interesting.
Thanks again. I am going to spend more time on my Op Amp & Frequency/Voltage circuit soon, maybe tomorrow since it's a bank holiday weekend here
cheers,
Carl