Paul_H

It struggles in that LM393 doesn't snap to low output for the new fuel cut.

Paul_H

Datasheet says LM393 input common mode voltage range V+ -1.5V. Input voltage may be larger than V+ without damage.

V1 setpoint works OK
V2 gets to about 4V then freezes - doesn't go to low output until about 5.6V is applied

I think we need a different comparator if we want to keep 5V supply, which ideally we do.

Paul_H

Doesn't MAX973 have input range V+ -1.3? so still has same issue with input range with 5V supply?

john banks

Why not just run the comparator off 12V? We could put a 5V zener in to keep the reference pots right. Plus a small but of positive feedback on V2. Job done?

Paul_H

No reason why not as they are rated to 36V, just trying to keep it to 5V if possible. Will try it.

john banks

I don't think to process a signal which can potentially go within 0.5V of the rail is easily possible without reducing the signal (which causes problems with V1). So I think running the comparator at more than 5V is required.

john banks

So, I think we are here then? Typo - the 10K in parallel with the zener should be 5K (for a 50% divider - down to 6-7.5V and zener will pull down the rests) as the two pots are in parallel!




[Edited by john banks - 2/24/2002 5:05:23 PM]

Paul_H

I tried running it off 12V, also with feedback resistor on V2 (I used 470K - all I had). It made no difference at all. OP2 still will not snap to low at the V2 setpoint. I've checked my build up for errors and can't see any.

When the input voltages are where you would expect the output to drop, the output "wobbles" to about 10V and the voltage at the input FALLS. Weird.

I'm using a PP3 & pot to fire voltages in & an wondering if the resistance of the pot is upsetting it?

Paul_H

Meanwhile - unrelated to the above circuit -

This circuit works effectively:

http://www.teamfc3s.org/info/articles/fcd/diy_fcd.htm

Note that this is a fuel cut ONLY. It is NOT a lifter.

Looks like this and costs less than £5:



The output tracks input very closely, here is the plot:





[Edited by Paul_H - 2/25/2002 1:30:47 AM]

[Edited by Paul_H - 2/25/2002 1:42:28 AM]

john banks

Works well Paul. The other circuit sounds as though it behaves oddly. I have had some more suggestions from a few guys on the newsgroup - will digest and post.

john banks

http://www.national.com/pf/LM/LMC6772.html

This is about £2 from Farnell and is a rail to rail single supply dual comparator with open collector outputs. Seems ideal for running off 5V supply?

john banks

The above is as far as I have got based on some suggestions sent to me from Xavier from the French Subaru board. I have emailed him with the above attempt so far and await his reply.

"I think you only need 1 comparator, 1 amp OP and a variable voltage
regulator (LM317 for example). I explain : The ampOP is used at G=1 and
powered with the variable voltage regulator. Adjusment of the first
threshold is done through the the variable voltage regulator (you tune the
output voltage with a varistor) so that the AmpOP saturates at the desired
input level. The second threshold is made by the comparator on the input
signal (threshold is tunes with a resistor bridge, one of the resistor is a
varistor). The comparator output drives the variable voltage regulator so
that when the comparator's output switches to high level, the voltage
regulator outputs a higher voltage, then the ampOP no more saturates and
its output goes to the input signal, which is over the ECU's cut threshold.
As you can see, it's very simple and no switches are needed.
You have to choose a rail-to-rail ampOP which accepts input voltages higher
than its power supply (most of the AmpOP accept this)."

john banks

Would it work?

T-uk

john banks,

have not read all the pages so forgive me if it's been suggested, but would putting a restrictor in the line to the map sensor not do the same?I noticed that cosie convert was running a second Dawes type valve to the map sensor at crail allowing him to run higher boost but shut down if it went too high.I just wonder if your restrictor valve would do the same.

Paul_H

John

Compare these two circuits - this one I know this works well:



Compare with this:



When comparator output is low D4 is reverse biased.
Comparator output is high at new fuel cut only, when D4 conducts.
Op-amp U1-2 follows voltage at pin 5 when defending.
When D4 conducts, output of FCD rises - triggers new fuel cut.
R5 sets new fuel cut level. R6 sets supply voltage for comparator.
FCD output will follow supply voltage when comparator goes high. This voltage must be sufficient to trigger a new fuel cut, but below 5V.
LM358 needs more than a 5V supply (input range supply -1.5V) which limits defend point to about 3.5V. Works well at 12V although I'm going to try a 7806.

All this is theory only - any comments before I try it?

Paul_H

T-uk

A bleed valve or similar will work but will result in MAP sensor giving wrong info to the ECU all the time with possible implications for ignition advance. We only want to feed it duff info part of the time

john banks

Sounds good.

Consider though that the voltage supply to the comparator will depend on its state as the current it draws will vary - I have the same issue with the last circuit I posted and Xavier suggested another buffer, with the modification below.



You could also consider using rail to rail comparator/op amps.

Keeping the component count down and the need for a reference would be nice.

John - Andy's system uses a ball-spring T'd off the map line as a relief valve, but the amount it can bleed is limited by a restrictor as well. The disadvantage is that you have to find the new fuel cut to work out where it is and it will also vary with temperature, as well as being considerably more expensive as a DIY project, although certainly easier than coming up with the right circuit.

john banks

Paul I think on the diagram the + and - inputs to the comparator have been transposed unless I am misunderstanding?

john banks

I think our designs are converging together.

Your comparator would have to be a "beyond the rails" input type if you are changing the supply voltage.

I struggled enough to find the LMC6772 to go up to the rails.

An op amp to go beyond the rails seems to be easier to find.

Keep us posted - we WILL get there and it will be worth it! 18 PSI does make it rather a beast.

[Edited by john banks - 2/28/2002 10:45:54 PM]

Cosie Convert

Paul

T uk was referring to a Dawes valve not a bleed valve. With a Dawes 'type' valve in the MAP line the pressure is only bled off above the valve set point ie 15 psi.
The MAP will see true pressure up until 15 psi and thereafter it will only see a percentage of the additional pressure rise, this percentage can be pre set by fixed restrictors.
My unit is set up for 20% ramp post 15 psi.

Therefore when true manifold is 15psi, MAP sensor input to ECU is 15psi

16 = 15.2
17 = 15.4
18 = 15.6
19 = 15.8
20 = 16.0 (Boost I'm using)
21 = 16.2
22 = Fuel Cut

All very simple and reliable although maybe not as much fun as electronics

cc

john banks

cc you might convert me to your method yet. You have a nasty habit of being rather right

Paul_H

CC

Wouldn't the ball in a Dawes stay closed until the setpoint - so MAP sensor would see no pressure until the valve opened? and how would it cope with vacuum in the manifold e.g. on the overrun..

I would love you to convince me, then I can just go throw all these electronics in the skip

Paul_H

Should have read above post more closely.....Dawes & bleed valve on T-piece....understand

Paul_H

Cheers John, yes - inputs transposed

Plan "B" (or should that be "F") here, with input to comparator after the buffer.


I think a rail to rail comparator will be OK e.g. if new fuel cut required at 20 PSI, set R5 to 4.36V and set R6 to just a little more. Anywhere up to 5V in fact. I'll let you know!

john banks

Xavier has come up with this and I think it will work very well. Need to use rail to rail input comparator and op amp.

BugEyed

John

I've been following this with interest for a while, and am impressed with the progress that you are making. Well done.

Am I right in thinking that although this is aimed at doing what it says in the title - stopping the fuel cut - it could also be used to cure the MY01's aversion to boost pressure. IIRC effectively the MY01 PPP (ignoring the mechanical changes) is "just" (a simple but effective solution as a more complicated one was not required) a boost clamp on the MAP signal to the standard ECU map preventing the ignition being retarded excessively and a revised boost controller that takes the correct signal. If this is correct, presumably combining both the Fuel Cut Defender and a separate boost controller would achieve the same effect?

Duncan

john banks

Duncan, the PPP on the MY01 is a smaller restrictor along with alteration of the MAP signal. There is no separate boost controller - it relies on the ECU aiming for a higher target as the MAP sensor output has been altered. However, the MAP sensor is remapped by RPM too, whereas the above design with a restrictor would presumably give you a bit more across the range. The boost controller seems superflous.

john banks

Yet another circuit. Cost about £2.34. Damn, I had better actually try one of these! Just throwing in lots of ideas!

See below for latest.

The comparator has open collector output and when activated lets the zener clamp the voltage. All runs of 5V. Both ICs are rail to rail input/output and 8 pin DIL packages. Any ideas on how the zener will work?


[Edited by john banks - 3/2/2002 11:22:38 PM]

john banks

Maybe the 1K would be better at the output of the op amp to limit the current through the zener?

john banks

With some hysteresis on the comparator, and more sensible potentiometer value. Could add the capacitors and diodes on the input like the fuel cut defender circuit.