Alfa Romeo needs Vacuum Sensor Device

[stevewaclo]

Just read the MAP link you sent and learned the difference between absolute intake manifold pressure and differential pressure. Interesting, because my VSD is essentially a differential device since one side of the bellows is at atmospheric pressure.

The Alfa's L-Jet ECU also responds to a barometric sensor, but since the two ECU's only share TPS info, I don't believe it's an issue.

Also interesting that the MAP sensor described recognizes positive manifold pressure as well...for un-naturally aspirated engines, of course. My Sierra Duramax has a boost pressure gauge installed on the A-pillar and when I floor the accelerator, boost can go to 20# and after a brief delay, that 5500# truck jumps!

 

[KeepItSimpleStupid]

Most gauges are differential. They read 0 at atmospheric pressure, not 14.7 PSI. The real terms are PSIA and PSIG for absolute and Gague.

A column of mercury for barometric pressure is measuring absolute pressure. i.e. inches of mercury.

Don't use the MAXIM analog switch.

4 bits might be too crude and 8 might be too much. 7-8 though I think is where you might want to be. A book on engine control would be useful.

I suppose you don't have a working sensor right now either?

So, are you guessing that the vacuum sensor primarily controls the timing advance? My old carburated Toyota engine 22R had two advance diaphragms on the engine. Some of the advance wierdness was temperature based and for starting issues too, I believe. It could have had one curve when cold and another when the engine was hot.

 

[stevewaclo]

Thanks for the tutorial on gauges!

Anticipating arrival of my repaired VSD any day. Will advise.

100% certain the IECU controls timing, period and only shares TPS (1/0) with the L-Jet ECU.

Interestingly enough I recently replaced the dual-diaphram, distributor advance device on my 320k, 22R, carbed, Toyota pick-up. Wife said I can keep it till it seriously breaks. Little does she know....

Best wishes

 

[KeepItSimpleStupid]

Your welcome.

I had an 82 Celica. Did 99% of the maintenance myself including putting the AC in from a box, replacing the steering rack and a clutch. After 17 years I gave it away and it was used as a getaway car twice. 220K. Still had the same front wheel cylinders.

Why don't you think about:
1. Getting a cheap scope
Say 75% of WOT is 2 mh and 50 ohms (It's not, but as an example)
2. Pick up a 2 mh inductor @ 50 ohms
3. Pick up some, say 3, 10 ohm resistors, so you can add a few in series
4. Pick up a combination or other 2 mH choke with a 10 ohms series resistance.
5. Again the idea would be to check with 0, 10, 20 and 30 ohm resistor in series.
6. Then pick up inductors say 50% and -25 and +25% from the 50% value
7 So if the 50% of the range is 1 mH 50 ohms, pick up a .75mH 50 ohm and a 1.25 mH 50 ohms

8. Pick up a OPTOMOS relay. I can help select. Will probably only need 1 mA so R=12/0.1 amp.
9 Try to see if you can use the OPTOMOS relay to select two values (Midpoint - 25%, midpoint +25%). Breadboard. A little switch of some sort. Measure the series R of the switch.

I think Inductance, series R, frequency, and Advance are the most important parameters. Use 1% resistors and inductor tolerance like 5% or so.

Once you have that data, you can probably figure out feasibility.

I'd try for a 7 bits of inductance and an 8 bit A/D converter.

Selecting the resistors and how many would depend on other things, including your results. The resistor selection might just be a series combination and you have to short a value out. The problem is that then the ON resistance of the OPTOMOS relay would dominate, so the series selection might be more complex.

You can't really do anything without some amount of data. The reason not to use the Maxim Mux is that it's very dependent on the supply rails. The OPTOMOS relay totally eliminates that problem.

Designing a circuit for automotive is not as simple as just slapping a 5V regulator and go. The automotive environment is plaqued with +50V spikes and -200 V spikes. It also has to survive a double the voltage battery jump. There are some nice parts from www.linear.com that I'd recommend for that.

Take a look at www.picaxe.com for a simple to program processor. Another suggestion would be an MSP430. I have a few, but havn't done anything with them. Like you, I've done some embededed stuff and assembly programming including playing compiler, but skills are nearly 35 years old. My first C program was a simulated operating system and my first Fortran program was a Direct Digital Control System that did PID temperature control for 7 heaters using inter process communication. Ask me about bitbanging I2C and I'm lost.

If you want you can reply via PM or email concerning anything that you may not want to post publicly. Could you estimate a market for a replacement and a target replacement cost? Order of magitude even. in increments of $100.

I saw some of your post elsewhere and my questions seem to be right on in terms of current and how is everything connected and a waveform.

Simulating a variable inductor seems possible, but unaffordable. Substituting a binary coded inductor seems possible.
Repeatability is probably more important than accuracy in this application. It's a control system, so most of the problems should come out in the wash. As I remember from my old days, vacuum advance affects starting, gas mileage and power.

 

[KeepItSimpleStupid]

Steve:

A long shot and I don't think it would work, but there are outfits that repair ECM's and they want $495 to repair an 1987 Spider (not necessarily yours). If they have info to repair, then... they are probably unwilling to share.

At the same time you could DISCUSS the possibility of a different sensor. Unlikey to work at all, but remember me. Come up with anything and everything.

As I said earlier, the actual OPTOMOS relay needs looking at better. Here is one that is Normally Closed with a 0.6 Ohm ON resistance, so it would not cause any problems. https://www.mouser.com/ds/2/205/LCB710-22935.pdf

 

[stevewaclo]

Let's go PM after this.

KISS,

Thanks again for sharing your time and knowledge!

Re: post 25 about ECM repair. You are suggesting I do 411 mining from someone who is offering to repair ECU's and obviously has the skills, knowledge and most importantly, the docs to get the job done. I recall seeing an ad for those services and my Alfa was listed. Will locate an ECM repair outfit and see what I may learn. Regarding a substitute sensor, never hurts to ask but unlikely any such highly specialized device exists.

Re: 24, wow, where do I begin? Excellent concepts but pushing my envelope.

Please remember, my overall mission, and apologies for not making it clear, is to develop a reliable substitute for the failed VSD which could be economically replicated by a reasonably capable Alfa owner. With your assistance and advice, I'm optimistic I could put such a device (system?) together and document a parts list and procedure. Keep in mind VSD's from the dwindling stock are currently available for about $80 (a very low bar) but once they are gone, or become increasingly rare, who knows? And this is not about personal gain. Just the challenge and maybe a feeling of accomplishment. Being the first kid on my block with such a device would be way cool. At this point, I'm down with proceeding.

That being said, what is your estimate of $$$, (+/- $25) of the parts you itemized above? That should give me a reality check for that direction.

While you are thinking about that, I should have more info on internals of the IECU, have my own functioning device back and a scope for gathering data on the system when it's working as designed. Also, I'm going to work on finding that VV carb device for possible use as replacement for broken bellows.

Going away Thursday for long weekend and booked up tight till I leave. Let's talk again early next week via PM.

Best wishes.

 

[stevewaclo]

Waaaaite a minute!

"After 17 years I gave it away and it was used as a getaway car twice"

How could you possibly know this?

 

[KeepItSimpleStupid]

The cops showed up at my door twice. I gave the car to the dealer in trade and didn't tell motor vehicle that I "sold" it. In reality I should have at least sold it for a $1.00 so there would have been a real paper trail. One time the thieves stole VCR's and the other time lawn mowers. I used the metric hp/lb of automobile when I bought the next car.

 

[stevewaclo]

That'll learn ya !

 

[KeepItSimpleStupid]

Yep, live and learn.

There might be an entirely different, easy and simple way of doing the control, but again, it does require measurements and would require a custom part. Frequency and current need to be known as well as the variance. Sometimes custom isn't too bad. Had a transformer done once. About $120 for a 4 x 35 V at 3 A toroidal transformer. "comparable" off the shelf core about $50.

Incidentally, speaker inductors https://www.madisound.com/manufacturers/madisound/inductors/aircore.php are in the right range, thus the RC servo mechanism could work, possibly with a moveable ferrite core.

Ideas, ideas.

Do you happen to know the max torque of your engine @ RPM? Trying to get a feel as to what's significant. e.g. I know what 80 ft-lbs feels like

______

First order:
You absolutely have to get at idea of frequency and current. The voltage and current waveform. The current waveform can easily be done with an easy to build I-V converter if the frequency is low enough.

Second order
Some sense of linearity. Frequency vs inductance.

Third order
Need the max torque of the engine at RPM or any values you have.

Fourth order
Some idea of the response time. Not sure what that means.

Fifth order
How much of a change is discernible. Friction probably modifies the response time.

Sixth order
I don't like the fact that you don't have a spare computer, nor can you open one up. I did find a ECM repair outfit that did an 87 spider for $500 which is steep.

Seventh order
Topology, or as much as can be gleaned from reverse engineering or contacting a rebuilding firm. You might be able to convince them to provide a snippet of the circuit surrounding the oscillator. I have use this successfully and unsuccessfully for other products.

---

An RC servo converted to a linear actuator and a speaker inductor with a series resistor and possibly a ferrite could actually be the safest. Kinda thinking about a hollow core ferrite that's linearly moveable.

__

I do have something else up my sleave. I do have a nice handheld Agilent RLC meter.

_____

I'm starting not to like the random connection of inductors. Why? Because the current in an inductor can't change instantaneously and that basically means a voltage spike at each switch.

 

[alec_t]

Instead of using a set of fixed inductors perhaps use a variable inductor such as in
https://www.google.co.uk/url?sa=t&r...JZgCrbM4syE5ZmQ&bvm=bv.42080656,d.d2k&cad=rja
That way the inductance is continuously variable, there's no switching, and the resistance will be a fixed value.

 

[KeepItSimpleStupid]

@Alec:

That was the one I had up my sleave. It's not an off the shelf component in the mh range. Frequency and current have to be known. Can't seem to find any decent design info.

 

[crutschow]

My gut feeling is that the inductor resistance is not critical as long as it is no greater than about 50 ohms. A lower resistance would just increase the Q of the circuit, but I doubt that will have an adverse affect on its operation.

 

[stevewaclo]

catching up on messages

thanks all...will respond properly next week.

 

[cowboybob]

I have to wonder if you couldn't replace the copper bellows with a stock vacuum advance unit (VAU) used on non-solid state distributors?

Such as this one:

**broken link removed**

(or something similar).

Mate and secure the VAU vacuum tube connection into the the hole in the bottom of the bellows canister, adapt and attach the VAU linkage to the inductor frame and away you go.

Might, obviously, need some calibration as to exact postioning, but...

 

[crutschow]

RE cowboybob's suggestion, you would likely need to add a spring to get the proper amount of motion with vacuum ( I don't think the VAU has much of a built-in spring). This would likely require some experimentation to get the proper spring size and tension.

 

[shortbus=]

A vacuum advance actually has a pretty heavy spring in it. The thing the O/P needs is a differential pressure device, vacuum to atmosphere.

 

[KeepItSimpleStupid]

Might be tough. Having rebuilt multiple bellows sealed valves, the bellows provides one end of support and a nylon bearing the other part of the support. I haven't got an answer for how much displacement is needed.

In reality, I think a response to absolute pressure would be best based on ECM requirements and the use of a MAP sensor rather than differential. I think it would then make the device more responsive to altitude changes. That's the idea behind the MAP sensor.

 

[cowboybob]

A vacuum advance actually has a pretty heavy spring in it. The thing the O/P needs is a differential pressure device, vacuum to atmosphere.

Which is what the VAU is.

 

[cowboybob]

All of these:

Absolute pressure sensor
This sensor measures the pressure relative to perfect vacuum.

Gauge pressure sensor
This sensor measures the pressure relative to atmospheric pressure. A tire pressure gauge is an example of gauge pressure measurement; when it indicates zero, then the pressure it is measuring is the same as the ambient pressure.

Vacuum pressure sensor
This term can cause confusion. It may be used to describe a sensor that measures pressures below atmospheric pressure, showing the difference between that low pressure and atmospheric pressure (i.e. negative gauge pressure), but it may also be used to describe a sensor that measures low pressure relative to perfect vacuum (i.e. absolute pressure).

Differential pressure sensor
This sensor measures the difference between two pressures, one connected to each side of the sensor. Differential pressure sensors are used to measure many properties, such as pressure drops across oil filters or air filters, fluid levels (by comparing the pressure above and below the liquid) or flow rates (by measuring the change in pressure across a restriction). Technically speaking, most pressure sensors are really differential pressure sensors; for example a gauge pressure sensor is merely a differential pressure sensor in which one side is open to the ambient atmosphere.

Sealed pressure sensor
This sensor is similar to a gauge pressure sensor except that it measures pressure relative to some fixed pressure rather than the ambient atmospheric pressure (which varies according to the location and the weather).

can be meet, once the VAU's vacuum port is sealed to the hole in the bottom of the canister, by adjusting the restored canister's ambient pressure situation accordingly.

My guess, given the age of the Alfa and the level of engine control sophistication at that time, the whole rig was designed around a differential sensor assembly.