Harvey - Solenoid Fault Checking ?

[Trevor]

Quote:
Originally Posted by oab_au Reference Post - Duty Solenoid B.

“The first thing to do is establish if the torque converter is locking up, like we suggested in the last post.

The code means that there is an electrical problem in the circuit. It is either drawing too much current, or too little. The Solenoid may be working OK. So check it to see.”

Harvey,

I have noticed on several occasions, including the above, that you have indicated trouble codes in respect of an electric solenoid, can be assumed to only register an electrical problem and not a mechanical malfunction.

Could you please explain exactly why this is so, as it will be of assistancs for those undertaking trouble shooting to be fully informed.

[oab_au]

The TCU monitors the current flow, that the solenoid is drawing. If it is 'out of limits', too high or too low a current, it posts a code. This indicates that the electrical circuit is open or shorted.

Harvey.

[Trevor]

Originally Posted by oab_au

The TCU monitors the current flow, that the solenoid is drawing. If it is 'out of limits', too high or too low a current, it posts a code. This indicates that the electrical circuit is open or shorted.

Harvey.

It is obvious that the trouble sensing circuitry can monitor cuurent, but why is it that a mechanical fault can not be detected ?

Trevor.

[oab_au]

There is no feedback to the tcu, to know any other condition, or fault with the B solenoid.

Harvey.

[Trevor]

Originally Posted by oab_au

There is no feedback to the tcu, to know any other condition, or fault with the B solenoid.

Harvey.

You escape the obvious in that that a highly inductive load is being switched and the inductance will vary greatly, depending on the position of the armature, as is present in a solenoid valve; i.e. in - high inductance, out - very low inductance. A simple bridge circuit can be set up to measure inductance and an indication signal acquired. There will be a considerable level of acceptable tolerance.

N.B. The Subaru Technical manuals state that a solenoid trouble code “detects open or shorted drive circuit, AS WELL AS VALVE SEIZURE”. I am of the opinion that that the truth has been published and some such circuitry is incorporated in the system.

I have noted over the years, that you always put yourself forward as an authority and state categorically, often repeatedly, what is alleged fact, rather than voice an opinion. Once again you do so and it is in the interest of others that your error is pointed out.

Trevor.

[oab_au]

Your book must be different to mine.





Harvey.

[Trevor]

Originally Posted by oab_au

Your book must be different to mine.





Harvey.

It would appear that I do have the same manuals as you, written in good English.

Try page 3-2 [TD21] and you will find the exact words I have quoted, repeated five times, in respect of all solenoids, A, B, 3, 2, 1, -

i.e. “ Detects open or shorted drive circuit, AS WELL AS VALVE SEIZURE.”

The page you have scanned, simply indicates what one should check to discover if the output signal (drive circuit), or solenoid, is open or short circuited, as a possible cause of a problem. There is no reference whatever to the extent of the diagnosis offered by the trouble code circuit, or of the alternative, which could be the existance of a mechanical problem.

In order to prevent the opportunity for another post designed to deflect and confuse the issue, I point out that the wording definitely refers to both in car and select monitor diagnosis, all as is described in the preceding text.

In the event that my honesty is in doubt, I will be pleased to add to the drama by scanning the quoted page. Currently I await a more objective reply.

Trevor.

[SilverSpear]

Cheezz, what's with you two attacking each other all the time???

Ok guys, we respect you both, now stop it

[Trevor]

Originally Posted by SilverSpear

Cheezz, what's with you two attacking each other all the time???

Ok guys, we respect you both, now stop it

BTW, i have a small question, I adjusted my brake band the first time, it went ok for a few Kms but then I felt shift taking long to lock up. And now on each gear I need to push gas to a high rpms (3000 and above) and ease my foot a little bit for the shift to lock in place... Do I still need to adjust the brake band as you advised Harvey? or this is something completely different?

Greetings Danny,

It is important in the interests of all, that technical information given here, in absolute terms, is correct.

There is no reason why a vigourous debate should not take place. There will be no nasty words and we are both adults. It is important that the truth regarding this matter is sorted out as it effects those involved with fault finding.

Please refrain from taking the thread off topic. An answer is required here for the good of all.

Courteously, Trevor.

[SilverSpear]

Ok I deleted my statement from the above, I will RE post it again in my other thread...

I am looking also for your suggestion in this issue Trevor

[oab_au]

This forum is for posting technical items. If you have some thing to say, just post it.

I agree with what you say. I had not seen that page in the book, the TCU can detect, if the
armature is not responding in accordance with that it should be doing.

The feed back necessary, would be the change in the inductance of the solenoid, as the
armature moved into/out of, the coil. As you say.

The change in inductance would cause, a corresponding change in the Rise Time of the
Duty cycle. The further on the solenoid valve, the higher the inductance, the more turned
off, the lower the inductance. I reckon the TCU would recognise this by the change in the
Rise Time of the Duty cycle on time, that feeds the solenoid, the two would have to stay
in step. So the TCU sending a 50% Duty cycle signal to the solenoid, would expect to see
a rise time, that corresponds to the armature being half way along the coil.

It would compare the rise time, against the time that it has written in the ‘look up tables’
in memory, if it was out of limits, shorter or longer, the code is set.

Good to find that, but surly you could have just posted this as a post, and save all the jibes
and dramatics. Did you expect a Fanfare, a Drum roll.

If you want to have a discussion, try the guys in “Not Exactly SVX “.

Harvey.

PS, All the above is my opinion.

[Trevor]

Originally Posted by oab_au

This forum is for posting technical items. If you have some thing to say, just post it.

I agree with what you say. I had not seen that page in the book, the TCU can detect, if the
armature is not responding in accordance with that it should be doing.

The feed back necessary, would be the change in the inductance of the solenoid, as the
armature moved into/out of, the coil. As you say.

The change in inductance would cause, a corresponding change in the Rise Time of the
Duty cycle. The further on the solenoid valve, the higher the inductance, the more turned
off, the lower the inductance. I reckon the TCU would recognise this by the change in the
Rise Time of the Duty cycle on time, that feeds the solenoid, the two would have to stay
in step. So the TCU sending a 50% Duty cycle signal to the solenoid, would expect to see
a rise time, that corresponds to the armature being half way along the coil.

It would compare the rise time, against the time that it has written in the ‘look up tables’
in memory, if it was out of limits, shorter or longer, the code is set.

Good to find that, but surly you could have just posted this as a post, and save all the jibes
and dramatics. Did you expect a Fanfare, a Drum roll.

If you want to have a discussion, try the guys in “Not Exactly SVX “.

Harvey.

PS, All the above is my opinion.

I will ignore your sarcasm, but will point out that once again you can not resist, other than to go one up, as a method of endeavouring to indicate superiority. This characteristic has annoyed others previously posting technical advice, to the point that they are now absent.

Your opinion is not credible in theory and wrong in fact. Your confused rendering may have served to impress the non technical but others will find evidence of a lack of basic knowledge and more particularly, logic.

An on/off valve should be first considered. (1) In what way could the inrush/rise time, be measured as becoming variably out of step (sync. ?) with the operating signal (effective duty cycle), when this is simply, on/off ? (2) What two parameters exist which the TUC could recognise as being in or out of step, as you predict. (3) Where, when and from what, could there be a time base established ?

You appear to confine your proposal to a valve employed as a variable pressure or volume control device.

Here I guess what you are suggesting is that, the rise time will in some way will not be constant in the event of a square wave monitored signal and something detectable will change in accordance with a varying inductance. I gather as time appears to be your basis, you are hoping for variation in frequency as the measurable factor, rather than wave form. (4) Just why would the applied frequency vary as a result of a change in inductance? (5) How can “look up tables in memory be brought within the picture? (6) Exactly what two variables do you expect to compare on the basis of time?

Most alarming is that you indicate that some form of 50% duty cycle signal will maintain the solenoid stationary in mid position, to be measured accordingly. (7) Are you suggesting that for practical purposes, a solenoid can be electrically held in a mid or intermediate position ?

The more I peruse your “opinion”, the more ridiculous and confused it becomes. As I stated the most likely method of sensing a fault in a solenoid is to use a comparative bridge arrangement to directly measure inductance. Before making this suggestion, rest assured, for good reason I ruled out the possibility of anything along the lines which you now put forward.

Your answers will be awaited with interest, Trevor.

[oab_au]

The main function in the TCU, knowing if the solenoid valve’s armature is moving, as
the duty cycle commands, is in the inductance that the coil exhibits in its operation. So
before answering all your questions, we need to explain what we are talking about, for
others to follow the post.

When we turn a solenoid on. the current that starts to flow through the solenoids coil is
resisted by a reverse current flow. As the current flows into the coil, it produces a
magnetic field, that rises from the iron core, through the coils windings. As it does it
generates a voltage that is in the opposite direction to the initial flow. So while this
magnetic field is rising, the flow of current into the coil is impeded by this reverse flow.
The higher the mass of iron in the core of the coil, the higher the inductance, hence, the
slower the current flow into the coil is.

So we know what the actual bits look like, this is the C solenoid opened up.



You can see the small armature with its return spring, and the ball that it pushes, into the
valve hole to turn the oil flow on and off. The return spring holds the armature out of the
main section of the coil, as the Duty cycle produces the magnetic field the armature is
pulled into the coil, this increases the mass of iron in the core, to increase the inductance.
When the armature is fully in the coil, the inductance produced is at its highest. During
operation the core, that the armature is in, is full of oil that damps its movement.

As the current to the coil is switched on, the flow is impeded by the inductance. The time
for the current flow to reach its maximum, is delayed, depending on the position of the
armature and its impedance.

This Diagram shows this effect.



From this we can see that there is an event that the TCU can use to recognise the position
of the armature. This effect is the time it takes for the applied current to rise to the
maximum current of 1 Amp.

We don’t need any special circuits to measure this. With a computer all that is needed is
a sub routine, to start a counter when the current is switched on, and stop it when the
current has risen to the 1 Amp. As the position of the armature changes, the inductance
changes, the Rise Time changes, the counter number changes also. This then will give us
a counter number that represents the Time Rise for all the positions of the armature in the
coil.



As the Duty cycle is increased the armature is positioned deeper into the coil, so the Rise
time increases. We now have a feed back that can be compared to the signal that drives
the armature, to see if the armature is responding to the drive signal. This is done in the
programming, by writing a set of ‘look up tables’ in memory that lays out what each Rise
Time is, compared to the level of Duty cycle applied. Like this;

% Duty cycle = 10% , 20%, 30%, 40%,,,,,,,,,,,,,,,,100%.
Rise Time = 10ms, 20ms 30ms 40ms,,,,,,,,,,,,,,, 100ms.

The program then runs a subroutine each time the current is turned on to the coil, the
computer measures the Rise Time, compares it to the % of Duty cycle in the ‘look up
tables’ to see if the Rise time, is as it should be, for that level of drive. If the Duty cycle is
changed to 90%, but the Rise Time is measured at 20 ms, it means that the armature has
not moved, and is still stuck at the 20 ms position. Fuggan marvellous arn’t it.

This should answer your questions No. 1,2,3,4,5,6. The last question, No.7, Quoted here;

“Most alarming is that you indicate that some form of 50% duty cycle signal will
maintain the solenoid stationary in mid position, to be measured accordingly. (7) Are you
suggesting that for practical purposes, a solenoid can be electrically held in a mid or
intermediate position ? “

Yes of course that is what I am saying, why else do we operate it, with a Duty cycle,
instead of a fixed on/off current. A fixed on/off current would only fully open, or fully
close the oil flow. In the case of the C solenoid that operates the AWD, it would be 100%
drive to the front 0% to the back, when turned off. 50% / 50% front/rear, when turned on.
The Duty cycle allows us to position the armature to any position, and maintain that
position, to provide any % of oil pressure, to the transfer clutch for any torque split to the
back wheels, that is needed.
This is a Diagram to show how the Duty cycle changes to control the armature position.



If we look at the top one, the solenoid is turned on for 10% of the cycle, and off for 90%
of the cycle. The short time that the coil is turned on for, is too short to allow the
armature to move very far, before the current is turned off, as it is pushing against the
spring pressure and is damped by the oil in the core. So as soon as the current is turned
off, the return spring starts to push the armature back again, but before it can move back
too far, the next cycle turns it on again, so it starts to move in again, this is repeated about
50 times a second, so if the frequency of the duty cycle is chosen to suit, the armature
mass, return spring pressure and oil damping, the armature will stay in that position.

In the 50% cycle graph the current is applied for longer % of the cycle, to allow the
armature to push deeper into the coil against the rising spring pressure, to reduce the oil
flow through the valve, before it is turned off again, to give a 75/25 torque split .

The 90% 10% graph shows that the longer on time, allows the stronger magnetic field to
compress the rising spring pressure for the armature to fully enter the coil, pushing the
ball into the valve hole to shut the oil flow off to give a 50/50 torque split.

This the way the A,B,C, 1,2,3. Solenoids in the box are operated, along with the power
steering solenoid, and the Air conditioner swash plate solenoid.

Hopefully this makes my last post clearer.

Harvey.

[b3lha]

Harvey,

I'm not wishing to get in the line of fire here, but could you clarify further.

Suppose the solenoid starts off closed and is then activated at a 10% duty cycle as per your diagram above. One cycle being 1/50th second = 20ms.

From my reading of your description, the initial 2ms pulse is enough to open the solenoid slightly, opposing the spring and the oil damping. But the remaining 18ms is not enough time for the spring to even slightly close the solenoid because of the oil damping? And then the next pulse hits and keeps the solenoid in its current position, not opening it any further?

It sounds like the spring tension and the level of oil damping must be very very accurate for this to work. A 2ms pulse at 1A opening the solenoid by exactly 10%. A worn spring or thinner/thicker fluid could cause significant variations. Does the TCU self-calibrate in some way?

From my purely theoretical standpoint, I can understand the idea of the solenoid oscillating such that averaged over any given time period, it is fully open for 10% and fully closed for 90%. But I can't yet grasp how a duty cycle can maintain the solenoid at an exactly defined partially open position.

Phil.

[Trevor]

Originally Posted by b3lha

Harvey,

I'm not wishing to get in the line of fire here, but could you clarify further.

Suppose the solenoid starts off closed and is then activated at a 10% duty cycle as per your diagram above. One cycle being 1/50th second = 20ms.

From my reading of your description, the initial 2ms pulse is enough to open the solenoid slightly, opposing the spring and the oil damping. But the remaining 18ms is not enough time for the spring to even slightly close the solenoid because of the oil damping? And then the next pulse hits and keeps the solenoid in its current position, not opening it any further?

It sounds like the spring tension and the level of oil damping must be very very accurate for this to work. A 2ms pulse at 1A opening the solenoid by exactly 10%. A worn spring or thinner/thicker fluid could cause significant variations. Does the TCU self-calibrate in some way?

From my purely theoretical standpoint, I can understand the idea of the solenoid oscillating such that averaged over any given time period, it is fully open for 10% and fully closed for 90%. But I can't yet grasp how a duty cycle can maintain the solenoid at an exactly defined partially open position.

Phil.

Thank you Phil,

I pointed out all this out to Harvey years ago, making the crucial point that even if it were possible to adjust things as he says, (It wuold be in fact be impossble with production components given the accuracy required.) Wear and tear, together with different oil viscosities involved, hot-cold, different manufactures, puts the proposal in the doubly impossible basket.

The valve armature continually moves in sync. with the fixed cycle time of the modulated supply. It may not in fact bottom in either direction due to the action of spring pressure and the controlled fluid, but it most certainly will not under any circumstance, unless fully open or fully closed, remain stationary. The resulting fluid output in fact has a wave form of varying magnitujde, which requires to be smoothed. This is carried out by the pressure modifier so as to prevent pulsations in the final line pressure.

What Harvey has put forward as his idea, could not be acheived in the manner he suggests. A transducer to sense armature position would be required, in order to produce a control signal for a feed back loop, in the manner of industrial control systems. Even this would be impossible in this application, other than under laboratory conditions.

Cheers, Trevor.