Bench testing Carrera Idle Control Valve
 

Thougt I'd post howto bench test a Carrera Idle Control Valve (ICV). I came up with this quick bench testing procedure for these valves.

The center pin is the ground pin and the 2 outer pins are the signal, but the signal that comes from the ECM is a negative voltage -0v to -5v signal. So if you apply ground to the center pin and -5V to one of the outer pins the valve opens then switch the -5v to the other outer pin valve closes.

So you can simply Bench test this valve with a 9V battery (see attached pictures). To simulate the -5V signal you apply the '+' side of the 9Volt battery to the center pin and the '-' of the battery to one of the outer pins and then move the '-' to the other outer and the valve will open and close. I tested this on my valve, the 9volt battery is certainly more than 5V but has little current so you won't damage the valve. However NEVER try testing the ICV with 12V from the car it could (and most likely will) burn the valve and it will be toast!

I also have a fancier test if you have a variable voltage supply, you can apply the '+' to the center pin and the '-' to the outer then start bringing up the voltage from 0V to 5V as you increase the voltage the valve slowly opens, you can actually watch it. Then once fully opened, switch the '-' over to the other outer pin and once again start at 0V then slowly increase to 5V and watch the valve close. I tested this as well.

So there you have it a bench test for the ICV.

Once I get the car running I'll verify what the voltages are to the outer pins while the valve is operational as well as where the valve is parked when you jumber B&C.

Here are the pics:
http://forums.pelicanparts.com/uploa...1186442858.jpg

http://forums.pelicanparts.com/uploa...1186442875.jpg

Excellent write-up. Thank you.

excellent. thx

+1, this really helped me too. Provided additional confirmation of a bad ICV

I'm really glad this has helped folks. I spent sometime with one of these ICVs I took a good one completely apart to fully understand how it works and even measured voltage levels with it functioning in the car.

The 9volt simple test will tell you 98% of the time the health of the ICV. The next better test is to use a variable power supply 0-5volt instead of the 9Volt, if you have a 0-5volt supply you can actually move the valve open and closed very slowly and look/listen for stick points.

yes,..best to work with the actual operational ranges that would be used....variable supply.

great write up.,.,., down and dirty!!!!!!

Best,

Couldn't one buy a control/regulator at Radio shack place it on the + line?

Somewhere I read that the ICVs are controlled by pulse width modulation of the 5v and -5v signals. I guess that applying a variable voltage, 0- 5v fools the ICV into thinking that it is seeing 0-100% pulse width duty cycle.

REALLY good advice about NOT applying 12v! Thanks.

good question.. it would be nice to know if they are PWM control or straight DC control - I bet Sal knows.

might be able to test the ICV with just straight DC even if it is PWM driven... I don't know... I remember someone saying they make a buzzing sound..

The ICV is a very simple control, I have had one apart and internally it is just a simple DC Motor, the type with permanent magnets. The only thing is it can not spin because it has a internal stop so it turns in one direction hits the stop (fully closed) then can be reversed and it turns in the other direction about 360deg and hits the same stop on the other side (fully opened). The DME sends simple 0 to -5V DC to the outer pins to get it to turn the voltage is variable 0 to -5v

Applying voltage briefly will pin the motor one way or the other which made me initially think it acted as a duty cylcle device - but it doesn't. In actual operation it "hovers" partway open.
Someday when my car is running I'll have to scope it.
-Chris

I'll see if I can get some wires on mine... and then get the mobile scope on it. Do the wires go to the DME? If so I can use my breakout box... I have to go look if those signals go up there.. or at least a signal that makes sense to look at.

You might be able to back probe it at the connector (remove the rubber boot on the connector and push scope probes in from the back).
Another way is to make a little breakout connector using a female 3 pin (Eagle Day sells them) and a hard to come by 3 pin male connector. You did save the connectors before you threw out those old speed and reference sensors right? ;)
-Chris

Will be replacing my Ref/Speed Sensors, PM me and I will mail you one ;)

ChrisBennet said:
"Applying voltage briefly will pin the motor one way or the other which made me initially think it acted as a duty cylcle device - but it doesn't. In actual operation it "hovers" partway open. Someday when my car is running I'll have to scope it."

And hcoles said:
"[ICVs] make a buzzing sound.."

The "hovering partway open" and "buzzing sound" indicate PWM control. 5v applied at 60% duty cycle is seen as 3v (that's probably oversimplified) by the device so it "hovers' at a point of equilibrium. A straight 5v (or -5v) is 100% duty cycle and fully opens (or closes) the device.

I see folks are determined to figure out this ICV control and how the DME signals the ICV. So I'll put in my details here:

First, my bench test was a quick dirty way to just check the ICV and I was never really sure just how the DME signals work, for example are the DME signals analog (0-5v variable) or are they somehow a duty cycle signal like the ones that fire the fuel injectors? I think this is the core of the question. After detailed look at the DME schematic and how the ICV is wired to the DME I have concluded that the answer MOST likely is it is a duty cycle signal. The control lines (2 of them) that goto the ICV from the DME are actualy Ground signals! Here is the schematic all marked up with my comments for explanation on my theroy of how it most likely works:

http://forums.pelicanparts.com/uploa...1201623158.jpg

First thing you notice is that pin #4 of the ICV is always HOT (12Volts) once the car is running. Pin 4 ALWAYS has 12v on it! This means that the 2 lines that goto the DME MUST work by going to Ground. Now we know from bench testing that 9Volt battery applied to the ICV drives it fully opened or closed. This leaves me with a funny feeling that the DME controll lines (pins 33 & 34) are pulsed GND signal lines. But they could also be variable 0-12V analog but I doubt it. Now couple in the fact that some folks have said that the valve BUZZES when it's working points to pulsed signal.

Next, how can we prove what these signals really are? Here is what I would do.
1. start the car let it warm up fully
2. once warm let it idle till idle is stable
3. pull the wire harness/connector from the ICV
4. The idle should stay right where it is so long as nothing changes, like load or air leak.
5. Now place a scope across the harness pins like this: use a 2 channel scope. Pin 4 (center pin) is always 12 volts so we really want to observe pins 3 & 5. So we scope pins 3 & 5 and see what we have.
6. Now if the car is still at a good idle, meaning the current idle matches the DME idle value, 880RPM for 86 Carrera or 800RPM for 84. If the idle is correct the DME should send NO signals to the ICV.
7. No introduce false air by removing a vacuum line or the oil cap. The idle will imidiately go up do to the false air and this should result in a signal seen on one of the ICV lines (pin 3 or 5). All the while the harness is dissconnected from the ICV because if you do this test with the ICV connected it would instantly correct the idle and the feedback loop would cut the signal and what we want is to keep the signal alive so we can really see it on the scope.
8. Then stop the false air leak and now lets drive the RPMs down by just applying load by turning on the headlights and heater blowers, this should be enough load via the alternator to drive the RPMs down and we should now see a signal on the other line of the ICV.

So, there you have what I think would be a good test.

If I get the time I will try this, I have a really good lab quality 2 channel scope. You could also try a simple volt meter as well.

Not 100% sure but my bet is this is a pulsed GND signal.

Also here is the test procedure from the DME Test Planning manual, this is the recomended in car testing for the ICV:

http://forums.pelicanparts.com/uploa...1201624640.jpg

Great stuff Sal, will be interesting to see your scope results.

Jon

"This leaves me with a funny feeling that the DME controll lines (pins 33 & 34) are pulsed GND signal lines."

Good guess!

"But they could also be variable 0-12V analog but I doubt it."

Bad guess!

Most make this issue more complicated than it really is,
i.e. just check that the valve is clean and most importantly
that each winding (from the center pin) is about 20 ohms.

A very simple operational test:
1. Remove valve and fully close the vane and re-install
the valve without connecting the connector.
2. Start the engine and it should idle at a low RPM.
3. Connect the connector. The idle should come up to
the normal RPM.
4. Stop the engine and remove the valve and fully open
the vane and re-install the valve without connecting the
connector.
5. Start the engine and it should idle at a high RPM.
6. Connect the connector. The idle should go down to
the normal RPM.

Simple!!!

Sal, very good. Does the ICV have a spring force like center point or does it "float". Also.. I'm wondering if by slowly applying from 0-4vdc and 0 to -4vdc on the other pin and looking for non smooth movements is of any value.

Hi Loren,

I agree, the first test you should do is the one you mention or the one from the manual page I scanned above.

But if the in car testing fails it might not mean the valve is bad, it could be a wiring issue or bad DME (cold solder joints). This is why I came up with the simple bench test.

Now, what has happened on this thread is that folks wish to really fully understand how this ICV puppy works. I don't mind us doing some research if it helps better our understanding. I'm willing to entertain the idea of doing some research to learn more.

But I'm with you 100% that you should start with the in car testing first then
if you suspect a bad ICV bench test it. But if the in car fails and the bench passes you may need to know a bit more and this is where our research can pay off.

As always thanks for your input.

Henry,

The ICV just floats, meaning it sort of stays put where you leave it. If you bench test it and you can get a variable supply that is the best test. You just vary the voltage from 0-5v and the valve starts to move you can hear it plus you can blow into it while doing this. Drive it fully open then switch the leads and drive it fully closed. All the while listen and look for stick points. Remember this thread is about bench testing the valve. But you should have done in car testing as well.

So we have our answer! The ICV signal is a pulsed ground signal, someone has already done the work for us, see this thread:

http://forums.pelicanparts.com/porsche-911-technical-forum/346667-motronic-idle-surge-solution.html

Also, note that in John's post above his issue was a burned out transistor in the DME! He diagnosed the issue with a scope on the ICV control lines and found one of them completley dead while the other had signal. The signal to OPEN the valve was fine but the one to CLOSE it was dead, he replaced the transistor and fixed the issue.

It's issue like this that really do require us to understand these signals.

Many thanks to John.

I want to add this extra in car test data to this thread:

With engine at idle you can measure DC voltage between the center pin and the 2 outer pins and you should see 3.0 to 3.5vdc on each side. If both sides don't read in this range you could have a bad drive transistor in the DME or a wiring issue. Of course I assume the valve is good.

Note that the valve MUST be connected and functioning when you perform this voltage reading test. This means you MUST have a way to access the pins with the valve plugged into the harness. Easiest way to do that is pull back the rubber boot on the harness so you can expose the wires in the connector. Or use 3 test leads to connect the harness to the valve.

Great work!

As an aside, I believe there is another failure mode that wasn't mentioned: the ICV itself wears out. The ICV is a little motor and I think internal contacts fail or more accurately, the the part that they rub on (the commutator?) wears through. [1]
A mechanic friend told me they wear out, I've replaced them and had that "fix" the problem. I've also disassembled one and seen some wear but wouldn't swear that it was the problem. I might have a worn one kicking around that I disassembled.

[1] I Am Not Totally Sure On This. You won't hurt my feelings if you correct me!

Chris,

Yes, they do wear out from mechanical movement at the contacts. I have such a valve here on the shelf. You can easily spot this type of failure because once a contact no longer makes contact that winding goes dead. Depending on what winding it is it results in the valve going fully closed or fully open and you either get extremely low or high idle speeds.

In addition to this failure the valves sometimes just get dirty inside and they bind up mechanically but this can often be fixed with some carb cleaner and a few q-tips to help aid the cleaning.

I bench tested my ICV and it opens/closes fine. I can't see the test procedure noted above from SCARCELLER. Does anyone have it?

Hi Scarceller,

I just tested my ICV Voltage not at idle but with ignition on, not sure if engine must be running, but I got 5.3 and 5.8V between centre pin and the two outers. I check wiring continuity, and its fine. Does this point to possibly transistors on the board failing, and needing replacement?

EDIT: I started the car and tested voltage at idle. one sat at 3.5V the other at 9V. I also bridged the valve and got 4.0V and 9V. So looks like I need to open the DME and try pull transistors off...

Thanks

You need a scope for a proper test! Read post #20 in this thread. It's a simple test to check the valve drivers in the DME ECM.

First, basic simple test.
Remove the ICV from the car and hold it in your hand tight, snap your wrist in either direction and the valve should be able to be opened with snap of wrist. You should be able to influence the valve to close or open with simple hard snap of wrist. If that does not work the valve is mechanically binding. Also if it's dirty clean it with starter fluid and few q-tips.

Next, fully close the valve using wrist snap. You should not be able to blow air through it when closed. With it fully closed plug it into the harness in engine bay but do not install it. Now turn ignition key to RUN and the valve should be humming and vibrating. And the valve should now have opened a bit.
This first test proves the DME is controlling the valve and is capable of opening it.

Next, remove ignition key. Unplug the valve and now fully open the valve manually. With it fully open plug it back in and turn key to RUN. This will cause the valve to close some. This test proves DME can close the valve.

If those tests pass, the DME circuitry is likely fine and so is the valve.

Why do you suspect you have a ICV issue? Can you elaborate on why you suspect the valve may not be working?

Enjoy!

Great revival of an old thread. Very interesting to read and learn how to test the ICV (easy way, and proper way).

QUESTION: what is the purpose of the ICV? What does it do, or what is it really controlling?

Thanks!

Hi Scarceller,

Thanks for replying. I can confirm the ICV moves freely when flicking it, and tests correctly at 22ohm and 44ohm. Middle pin out puts 12V too. I have checked Pin 33 and 34 from ECU to plug in engine bay, and there is continuity. After completing your tests mentioned above, I can confirm valve moves to centre position in both situations. So this as you say, points to DME being capable of centering the ICV.

Why I think there is an issue is because the ICV doesnt seem to ever open fully or even pass mid way when engine is running.

After the first 45 second cold start sequence, the idle then sits at 600 rpm until car is fully warm, in which it then gets up to 800rpm.

When using the test port to centre the ICV (with engine running), it wont centre. The picture below illustrates this.

So I am just really confused as to why the DME is forcing a low idle at cold start. It could easily increase rpm, but forces the valve closed instead.

Thanks.


http://forums.pelicanparts.com/uploa...1677746892.jpg

Yes! Post #20 provided a simple test to eliminate both the valve & the ECM drivers. Did you do it?
The next test should have been to test for the idle switch on pin 2 of the ECM being closed.
I've seen idle problems the result of the A/D converter in the ECM having bad input resistors values, i.e. not sensing pin 2 properly, or a damaged A/D input.

It's NOT that difficult!

Tested all ECM pin outs, idle switch worked. Depressed the throttle by 1mm and continutity changed. When pulling the idle switch plug off when engine is running, idle speed shoots up.

How else should I be testing it?

Re-do the test in post #20. Again;

1. Remove the valve and fully close the valve & re-install and start engine without connecting valve;
a) idle is very low and almost dies
b) connect the valve - idle increases
c) idle opens properly, if not - bad valve or ECM

2. Remove the valve and fully open the valve & re-install and start engine without connecting valve;
a) idle is very high
b) connect the valve - idle decreases, if not - bad valve or ECM

Hopefully, the valve is not inserted the wrong way. The arrow on the valve indicates air flow.

It doesn't get any simpler that!

It doesn't need to be condescending.

Yes, would be nice to assume I am not a complete idiot haha

I also feel as if I am going in circles with this, not just you :)

All I can say is I have tested the icv valve and function as per post 20. It passes when engine is not running.

This is my predicament though,

When shorting the connector to centre the icv I get two different conditions depending if engine is running or not.

If I turn ignition to run, then disconnect the icv plug, then start the car, it idles at 800rpm.

If I start the car first, then bridge the connector, the engine runs at 600rpm.

I have visually inspected the icv in both situations, and when pulling the plug before starting the engine, it is centred. When checking the valve after starting the car and bridging, the icv is closed.

So I am very confused as to why, if the engine is running and icv bridged, it won't centre.

I am hoping I am an idiot and there is a simple reason for it. For now, I can't set base idle as the icv just closes when the port is bridged with engine running.

The test in post#20 requires the engine to run after manually re-positioning of the vane in the valve. Read the post again.
The idle when the vane is fully open before re-connecting should cause the idle to be close to 1500 RPM,
if the throttle bypass screw is adjusted properly. Once connected, the idle should be less than 1K, but not 600.
The other test requires the valve be fully closed manually and re-installed without the connector.
Once the engine is started, the idle should be very low, e.g. less than 400-500 RPM. When the connector
is then connected, the idle should be about 800-900 RPM.

If not, then the tests failed and there's a problem in the DME ECM or the valve.

Set the base idle to about 800 RPM with the jumper connected. Once the jumper is removed, then the valve can compensate for any engine load changes from the base setting,
e.g. the compressor or throttle undershoot. If this doesn't function, then there's a problem with valve driver circuitry or valve. Use an ohmmeter to check the collector (middle pin)
and to each of the other two pins on both darlington transistors. The meter polarity is key. You may have a shorted collector to its emitter or base.

Message for scarceller

I'm in the UK and have an 1988 3.2 Carrera Targa.

My idle was surging a few times and I naively thought it was the Idle Control Valve (ICV). I took mine apart and cleaned it and when I put it back on the car the engine continued to surge.
Whilst the engine was running I turned the body of the ICV and it stopped vibrating!
I assumed that I had broken the valve and so got a used replacement off ebay. On the car this replacement valve didn’t vibrate or work.
I then took the car to a specialist garage (shop) and they also couldn’t get the ICV working.
I purchased a new LOWE valve from Germany and that too isn’t working.

All three of the valves I have will click side to side with 3 volts applied across the centre and outer pins.

There is 12.2/3 volts across the ICV Connector centre pin and ground, plus around 10.5 to 10.7 volts across the centre and outer pins with the ignition on, car not running.
On a normal multimeter I have 12.6 to 13 volts across the centre to one outer pin with the engine running and 12.7 to just over 13 volts from the centre pin to the other outer pin.

My friend bought his scope over and it is showing a square duty signal sign wave at the connector (disconnected from the valve) with the ignition on.

We opened up the engine management box and non of the components appear blown and the transistors and diodes are not showing any open circuits.

How does the engine management reduce the 12 volt supply down to 3 to 3.5 volts to modulate the ICV as you state? If i’m getting more volts and the ICV isn’t working/vibrating what is the likely cause?

Appreciate your help
Thanks
Andrew
UK