Strange setup problems

[Ilkka]

Hello,

I bought the HFS-3 system for my Saab NG900 2.0 turbo -96 with the original Trionic 5 engine management system (around 350Hp, 500Nm before the Aquamist). Spent the weekend with the installation and today started to perform final system tests. I instantly ran into strange problems.

Once I turn the ignition key to pre-crank position and wait a few seconds for the system to start, all three lights on the board lit up (constantly). At the same time the pump starts pumping and doesn't stop until I turn the ignition off. I measured the voltage on the injector- wire (soldered to the pin 3 wire coming out of the ECU) during pre-crank, and it was 0.12V. Injectors are Siemens Deka 630cc, 12ohm. The black and white wires are grounded to the chassis ground point near the ECU.

If I start the car, the pump shuts down and the F-IDC led starts to flicker and the two other leds aren't active, just as it should be. Now the injector- measures 14.1V at idle. Injector+ (red wire) and main power (red wire) measure 14.3V (connected to the same pin 60 ignition switched wire coming into the ECU). With the threshold trimmer at the minimum, if I blip the throttle the THRES led lits but nothing comes out of the nozzle. If I put a jumper to the SYS, the nozzle outputs a nice cone.

Below you can find a lot of information on Trionic 5. The ECU pinout can be found on page 125.

Here's how I connectd the HFS-3 to it:
Pin 3 - injector- (green)
Pin 26 - fail safe (cut the wire and used brown and white)
Pin 60 - injector+ (red)

http://trionic.mobixs.eu/Trionic%205.pdf

Any idea how to start debugging the system?

I have access to portable oscilloscope etc, and work daily with electronic instruments and devices.

Thanks!

Ilkka from Finland

A few pictures at different stages of the install.

http://personal.inet.fi/private/irissanen/aquamist/img%20014_s.jpg

http://personal.inet.fi/private/irissanen/aquamist/img%20010_s.jpg

http://personal.inet.fi/private/irissanen/aquamist/img%20018_s.jpg

http://personal.inet.fi/private/irissanen/aquamist/img%20017_s.jpg

[Ilkka]

Would love to get some customer support from the manufacturer of this expensive device.

Here's the wiring diagram of the Trionic 5. As described earlier, injector- (green wire) is connected to the pin 3 on the ECU (E33 wire on the diagram).

206a-d are the injectors.

http://personal.inet.fi/private/irissanen/aquamist/saab%20t5%20injector%20wiring.png

[Richard L]

can you connect the red wire of the grey harness to J67as in the schematic above.

Richard

[Ilkka]

can you connect the red wire of the grey harness to J67as in the schematic above.

Richard
I will try that and report back.

Thanks!

[Ilkka]

OK, wired the red wire of the gray harness to J67.

Now the system won't start (gauge nor any leds on the main unit won't lit) until I start the car. The J67 won't have any voltage until the car is started, is this the reason?

Once the car is idling, the F-IDC led starts to flicker and the two other leds aren't active, just as it should be. With the threshold trimmer at the minimum, if I blip the throttle, the THRES led lits but nothing comes out of the nozzle. If I put a jumper to the SYS, the nozzle outputs a nice cone. So no change from what was noticed before.

[Richard L]

Ok, we are getting there.

The reason for the full-on spray is because the (+) side of the fuel injector side is not powered up at the pre-crank position. By wiring the red wire from the grey harness to the fuekl injector (-) has sloved the "false start" probelm at the pre-crank position.


OK, the next problem...

Check the jumper "INJ" is linked and the gauge is switched on.

[Ilkka]

Ok, we are getting there.

The reason for the full-on spray is because the (+) side of the fuel injector side is not powered up at the pre-crank position.
But it was powered up before I moved the injector+ to J67. When powered from pin 60, pre-crank position gave 12 V.

By wiring the red wire from the grey harness to the fuekl injector (-) has sloved the "false start" probelm at the pre-crank position.
So that was intentional and wanted?


OK, the next problem...

Check the jumper "INJ" is linked and the gauge is switched on.
Both check.

[Ilkka]

This is only me but I'm suspecting that the 0.12V DC measured on the injector- at the pre-crank position fools the system to think that the IDC is 100%. That's why all three leds lit constantly. Could try a large capacitor (10000uF) in series with the injector- to block the DC.

But I don't understand why the system is not working when the car is started. Everything seems to be working ok but the nozzle doesn't output anything. Test with the SYS linked proves that both the pump and check valve are working.

Could it be that the output of the nozzle is so weak at low IDC that it isn't visible? I will take the car on the road tomorrow so that I can engage higher IDC.

[Richard L]

There is not reason why the system will not inject. Your reasoning for the injector at 0.12V is correct, so the H3 system thinks the injector is at 100% DC.

Most cars powers up the fuel injector at "pre-crank" position. Yours doesn't. Until of course after the engine has started.

Can you check all the jumper links are at the "default position" and the gauge is switched on.

Richard





There test pins

[Ilkka]

There is not reason why the system will not inject. Your reasoning for the injector at 0.12V is correct, so the H3 system thinks the injector is at 100% DC.

Most cars powers up the fuel injector at "pre-crank" position. Yours doesn't. Until of course after the engine has started.

Can you check all the jumper links are at the "default position" and the gauge is switched on.

Richard

I will put the injector+ wire back to pin 60 so that it will be powered at the pre-crank position. Then I will add a 10000uF capacitor in series with the injector- to block the DC.

Jumpers are at the "default position" and the gauge is switched on.

[Richard L]

Before you do that, please read the voltage at test pin16 against ground, This shows what the controller sees.

0-5V = 0-100% DC.

Richard

[Richard L]

If possible, please take a picture of the circuit board of the controller.

[Richard L]

Just realise, you have a scope!

I suppose there is no harm done with a capacitor as it has a large valve, it is unlikely to swing below ground level. It might not work because thereare no ground path on the H3 side.

Please describe the waveform of the injector pulse. Let me know if it is not a peak and hold circuitry.

[Ilkka]

Before you do that, please read the voltage at test pin16 against ground, This shows what the controller sees.

0-5V = 0-100% DC.

Richard
Pin16 against ground on the board was 5.22V.

If possible, please take a picture of the circuit board of the controller.
http://personal.inet.fi/private/irissanen/aquamist/img%20021.jpg

http://personal.inet.fi/private/irissanen/aquamist/img%20020.jpg
Just realise, you have a scope!

I suppose there is no harm done with a capacitor as it has a large valve, it is unlikely to swing below ground level. It might not work because thereare no ground path on the H3 side.

Please describe the waveform of the injector pulse. Let me know if it is not a peak and hold circuitry.
Will do that tomorrow, now it's getting too late.

The capacitor didn't do anything.

I measured the voltage between an unconnected injector- wire and ground, it was 10.8V. Once the injector- is connected to the ECU pin3, the voltage drops to 0.12V (all at pre-crank). At the same time the injector+ sees around 11.8V.

This is what I found on the Trionic 5 injection system:

"The fuel injector valves are of solenoid type with needle and seat. They are opened by a current flowing through the injectors coil and are closed by a strong spring when the current is switched off. To ensure as optimal combustion as possible and with that lower exhaust emission the injectors are equipped with four holes, which gives a good distribution of the fuel. The squirts of fuel are very exact positioned (two jets on the backside on each inlet valve). This put very high demands on the fixation of the injectors. To secure this fixation the injectors are fixed in pairs by a special retainer between cylinders 1 – 2 and 3 – 4. The injectors are electrically supplied from the main relay, while the ECU grounds the injectors."

[Ilkka]

After reading this (http://hondaswap.com/general-tech-articles/peak-hold-vs-saturated-injectors-30347/), I'm 100% certain that my injectors are saturated. They are Siemens Deka 630cc, 12 ohms.

[Richard L]

On the HFS3, you put the jumper link to FDC to see what the injectors are doing via the bar graph on the gauge. This way you don't need a DVM to measure pin16.

12 ohm injectors are usually saturated driven.

Can you measure the voltage of the injector (+) at the pre-crank key position.

[Ilkka]

On the HFS3, you put the jumper link to FDC to see what the injectors are doing via the bar graph on the gauge. This way you don't need a DVM to measure pin16.

12 ohm injectors are usually saturated driven.

Can you measure the voltage of the injector (+) at the pre-crank key position.
I've already posted it, it's around 12V (11.5-12.5V depending how long the car has been sitting since the last start-up).

[Richard L]

Sorrym I should have read it more carefully.

.. at pre-cranking key position..........
This is strange. the presence of the injector (+) should not trigger the system unless the injector (-) is >0.5V below 11.5-12.5V. This how the H3 defines a ground pulse.

[Ilkka]

Sorrym I should have read it more carefully.

.. at pre-cranking key position..........
This is strange. the presence of the injector (+) should not trigger the system unless the injector (-) is >0.5V below 11.5-12.5V. This how the H3 defines a ground pulse.
Well injector- is 10.8V against ground when not connected to ECU pin3. When connected, injector- vs. ground drops down to 0.12V.

[Richard L]

One more measurement tomorrow. Voltage difference between P60 and J67 in pre-crank position. It is possible the fuel pump is dragging down the voltage at J67, so that the voltage is more 0.5V below P60. Under this condition, the system will trgger un-unintended.

[Ilkka]

One more measurement tomorrow. Voltage difference between P60 and J67 in pre-crank position. It is possible the fuel pump is dragging down the voltage at J67, so that the voltage is more 0.5V below P60. Under this condition, the system will trgger un-unintended.
That's easy because voltage at J67 in pre-crank is 0V, as discovered a few posts back.

I don't understand how J67 has anything to do with this? Injector+ (and also main power) is connected to the pin60, and injector- to the pin3.

Also fuel pump starts only for a 2-3 seconds at pre-crank position and doesn't start again until the engine is started. The HFS-3 is still in yellow light mode during this.

[Ilkka]

Sorrym I should have read it more carefully.

.. at pre-cranking key position..........
This is strange. the presence of the injector (+) should not trigger the system unless the injector (-) is >0.5V below 11.5-12.5V. This how the H3 defines a ground pulse.

Well injector- is 10.8V against ground when not connected to ECU pin3. When connected, injector- vs. ground drops down to 0.12V.
Isn't this the problem? Injector- is 10.8V (or actually 0.12V when connected to ECU pin3) when injector+ is 12V? Higher than 0.5V difference.

I don't understand why injector- goes down to 0.12V when connected to ECU pin3. The pin3 must be grounded during pre-crank, otherwise it doesn't make sense.

What's between those two wires (injector + and -) inside the box?

[Richard L]

This is very strange.

The two wires of the HFS-3:

Red: ............ + reference of a comparator input
Green: ........... - reference of a comparator input

The trigger differential is set to ~0.5V between those two points. Ie is the voltage of green wire is 0.5V below the red wire, the system sees it as a starting point of the injector firing.

From the diagram J67 appears to be the main +12V supply to all the injectors and other devices. So I can safely assume that I can use J67 as a (+) reference for the +12V.

If you can post a larger diagram of the ECU that has p60 included so I can see what else is connected to that pin.

[Richard L]

Isn't this the problem? Injector- is 10.8V (or actually 0.12V when connected to ECU pin3) when injector+ is 12V? Higher than 0.5V difference.

I don't understand why injector- goes down to 0.12V when connected to ECU pin3. The pin3 must be grounded during pre-crank, otherwise it doesn't make sense.

What's between those two wires (injector + and -) inside the box?

If pin 3 is grounded the fuel injector must be firing at 100%, if there is 12V on the other side of the fuel injector.

[Ilkka]

If pin 3 is grounded the fuel injector must be firing at 100%, if there is 12V on the other side of the fuel injector.
Yes but there isn't 12V on the other side during pre-crank. The J67 get's powered up only once the car is actually started. And then the pin3's continuous ground switches to normal pulsing ground.

[Richard L]

Ideally, put the red wire on J67 and green on pin3. The system can only be used after the engine is running, is it acceptable?

During cranking, the system will not inject because there is a 10 second start-up timer to disable WI activating.

If you want to switch on the system before cranking, you have to disconnect the green wire from the injector(-).

[Ilkka]

Ideally, put the red wire on J67 and green on pin3. The system can only be used after the engine is running, is it acceptable?

During cranking, the system will not inject because there is a 10 second start-up timer to disable WI activating.

If you want to switch on the system before cranking, you have to disconnect the green wire from the injector(-).
Yes, that's acceptable and that's what I am planning to do next. Then I try a proper high IDC road test to see if the system is activated by the IDC.

[Richard L]

We will move on to test the working of the system next. This will allow me time to find a way around this.

[Ilkka]

I can confirm that the system is working. Even though the THRES led lit during standstill revving on the parking lot, it wasn't enough to trigger the system to spray. Once I got on the road and was able to push it harder, the nozzle started to spray.

Can you guess why the THRES led lights too early?

Now I need to empty the tank from water and put the 50/50 mixture in. Then starts the fine-tuning.

[Richard L]

You can diagnose this by putting the jumper to FDC and watch the gauge what the system is seeing.

[Ilkka]

You can diagnose this by putting the jumper to FDC and watch the gauge what the system is seeing.
What's the relationship between bars and FDC?

[Ilkka]

Okay, did the first test run with 50/50 mixture (store bought -50C windshield washer fluid, manufacturer has told that it should be almost pure 50/50).

I have one 0.3mm nozzle in each intake runner, so total of four 0.3mm nozzles. Each line has a 15 psi check valve just before the nozzle. Restrictor used in FAV is 0.7mm.

WL set to min, WH set to max. THRES and GAIN at the factory default.

SC had to be set to max in order to get enough bars on the display. During spool up 6 bars show up quickly but then stables to 4 bars. To me it seems that there is not enough flow. Should I set the gain higher?

Here's quick on the road log taken with the T5Suite program (sorry, Saab specific ;) ) and exported to Logworks3.

http://personal.inet.fi/private/irissanen/aquamist/first%20methanol%20100-200.png

[Richard L]

If you are running with one nozzle, I can reflash the flwo table sio that you can the best resolution on the gauge. You need to post the controller back yo me and It will take one day, return post an ?5.00

This is a very nice logging work. I will have a closer look tomorrow.

Can you tell me what voltage you are getting from pin16 or bar ready with FDC linked at full boost and RPM. You can do it in low gear.

try setting gain higher but avoid getting the 95% red led to come on.

Richard

[Ilkka]

Please read my previous post more carefully re the number of nozzles.

[Richard L]

I read it but didn't seemed to sunk in. My apologises. I still think you can do with a reflash because the SC is set to maximum. Are you intending to flow a bit more than your present setup? I am happy to change the 0.3mm jets over for you to 0.4mm jets if you send them to me.

[Ilkka]

Is there any way to tell the absolute amount of flow by looking at the flow meter? There is a legend on the gauge which reads 100 to 400 ml/min, but is that with what SC trimmer setting? If that's with 'centre' setting, how is the legend changed when SC is set to max?

Per your flow charts, four 0.3mm nozzles should flow enough (around 440 cc/min required with 20% ratio) compared to fuel flow (around 2200 cc/min) at the moment so there shouldn't be a need to change them. But of course if the actual flow is lower than what is calculated, it might be needed. But then the question is, why are the calculations wrong.

[Ilkka]

One more question before I continue with today's results.

What is the correct orientation for the 15 psi check valve?

http://www.aquamist.co.uk/sl/plist/pic2-08/806-249/806-249.html

Now I have them so that the 1 bar arrows are pointing towards the nozzles. Is that correct?

[Ilkka]

Can you tell me what voltage you are getting from pin16 or bar ready with FDC linked at full boost and RPM. You can do it in low gear.

try setting gain higher but avoid getting the 95% red led to come on.

Richard
With the FDC linked, I'm getting full 8 bars above ~5000 rpm. ECU calculated fuel injector DC is ~75% at 5000 rpm. I don't know if the red led is on because the box isn't visible during driving.

I also tried trimming the gain, but it didn't seem to have any effect what so ever. With the normal flow gauge enabled, I'm getting 4 bars whether the gain is set to min, centre or max. Also the gain setting didn't seem to have any effect on the AFR readings. Very strange.

Below are two runs overlaid on top of each other. Solid lines are with gain set to MAX, and dashed lines are with gain set to MIN. AFRs (orange lines) are basically identical.

http://personal.inet.fi/private/irissanen/aquamist/gain%20min%20max.png

[Ilkka]

Tomorrow I will test the larger restrictor (0.9mm), but if that doesn't help, then I guess it's time for the 0.4mm nozzles.

[Richard L]

I suggest leaving restrictor out altogether out and see if you can squeeze some more flow out of the system.

[Ilkka]

Can you please see my questions below. Thanks.

Is there any way to tell the absolute amount of flow by looking at the flow meter? There is a legend on the gauge which reads 100 to 400 ml/min, but is that with what SC trimmer setting? If that's with 'centre' setting, how is the legend changed when SC is set to max?



What is the correct orientation for the 15 psi check valve?

http://www.aquamist.co.uk/sl/plist/pic2-08/806-249/806-249.html

Now I have them so that the 1 bar arrows are pointing towards the nozzles. Is that correct?

With the FDC linked, I'm getting full 8 bars above ~5000 rpm. ECU calculated fuel injector DC is ~75% at 5000 rpm. Does 8 bars mean that the HFS-3 is seeing already 100% DC. How can that be when the real DC is only 75%?

[Richard L]

Is there any way to tell the absolute amount of flow by looking at the flow meter? There is a legend on the gauge which reads 100 to 400 ml/min, but is that with what SC trimmer setting? If that's with 'centre' setting, how is the legend changed when SC is set to max?


The pulses from the flow sensor (absolute value) can be intecepted from pin2 (yellow wire) of the blue harness, red wire is pin1. This raw signal will not be affected by the SC trimmer. Frequency to voltage converted signal (SC affected) can be obtained from test pin21.



What is the correct orientation for the 15 psi check valve?

http://www.aquamist.co.uk/sl/plist/pic2-08/806-249/806-249.html

Now I have them so that the 1 bar arrows are pointing towards the nozzles. Is that correct?

You have installed it correctly.


With the FDC linked, I'm getting full 8 bars above ~5000 rpm. ECU calculated fuel injector DC is ~75% at 5000 rpm. Does 8 bars mean that the HFS-3 is seeing already 100% DC. How can that be when the real DC is only 75%?

DFC is displayed between ~10-90% or 0.5V to 4.5V (2nd-9th bar). 8th bar shows the system is seeing ~80% F-IDC.

I have mentioned before test pin16 is what the system is seeing from the fuel injectors in voltage from 0-5V = 0-100%DC.

[Ilkka]

The pulses from the flow sensor (absolute value) can be intecepted from pin2 (yellow wire) of the blue harness, red wire is pin1. This raw signal will not be affected by the SC trimmer. Frequency to voltage converted signal (SC affected) can be obtained from test pin21.

So there is no way to see the absolute flow from the flow gauge? Why is there a 0 - 400 ml/min legend then?


DFC is displayed between ~10-90% or 0.5V to 4.5V (2nd-9th bar). 8th bar shows the system is seeing ~80% F-IDC.
I can see only 8 bars on the gauge? Where is the 9th bar?

[Richard L]

So there is no way to see the absolute flow from the flow gauge? Why is there a 0 - 400 ml/min legend then?


It is possible if you calibrate the sensor against SC to give some representation of the voltage against flow. The 400ml is just a scale or multiple of.... The flow sensor is design for use with the failsafe, it is a simple flow monitoring system, Not intended to read absolute flow. If you want a true and calibrated flow sensor, I can supply one that is calibrated and signal conditioned to your setup. The sensor is ?80.00 and calibration will be ?60.00 on top.

As I mentioned before, the raw signal (pulsed) from the sensor can be used. But you can make your own decoder to make the output absolute.


I can see only 8 bars on the gauge? Where is the 9th bar?

Correction, I meant the 1st to 8th. Sorry.
1st = 10-20%
8th = 80-90%

[Ilkka]

Here's a log from my first standing mile event. Will change to 0.4mm nozzles to get some more flow.

http://personal.inet.fi/private/irissanen/halli%20run2.png

[Richard L]

I think it is a good move to go for the 4mm. I just remembered to mentioned, if you have four inline checkvalves, it will further taxing the flow.

Do you mind if I resize the pictures for you? You do post great information!:)

[Ilkka]

I think it is a good move to go for the 4mm. I just remembered to mentioned, if you have four inline checkvalves, it will further taxing the flow.
Yes, I have four inline check valves. Probably the pressure drop through them is a little bit larger than what is claimed?


Do you mind if I resize the pictures for you? You do post great information!:)
I changed it to a link instead of a picture.

[Richard L]

I suggest that you send the jets for size change and controller reflash as soon as you can. You last plot registered some knock events.

Pressure drop is directly proportion to the "cracking" pressure of the valve. A 15psi checkvalve will reduce 15psi across the valve. There are no reason to muddle the "actual" or "claimed" pressure drop discrepancy. Aside from the natural flow resistance due to turns and frictions.

[Ilkka]

I suggest that you send the jets for size change and controller reflash as soon as you can.
I have been in contact with your Finnish distributor, Sporttiauto, regarding the nozzle change. They say that they don't have 0.4mm nozzles in stock at the moment. I suggested that could you (Richard) send them to me directly instead, but I haven't got a reply yet. I would like to test them first before sending my current 0.3mm back. What if I get too much flow with them?

Why would the controlled need to be reflashed? If I'm using four 0.4mm nozzles, that's same as two 0.8mm nozzles. You have written that the controlled is optimized for two nozzles. Is two 0.8mm still not enough flow?


You last plot registered some knock events.
Yes it did, but they weren't because of the W/M injection. Also Trionic 5 is very sensitive to register knocks. Usually they aren't real knock events.


Pressure drop is directly proportion to the "cracking" pressure of the valve. A 15psi checkvalve will reduce 15psi across the valve. There are no reason to muddle the "actual" or "claimed" pressure drop discrepancy. Aside from the natural flow resistance due to turns and frictions.
Is it possible to raise the pumping pressure to compensate the drop?

Here ("pump pressure adjust" almost at the end) it says it can be done: http://www.impconcepts.com/porscheturbowaterinstallation.htm

[Ilkka]

I think I'm having the same issue as described in this thread. (http://www.aquamist.co.uk/vbulletin/showthread.php?p=12492) If I'm accelerating with partial throttle (say 75%), all 8 bars on the gauge light up, but once I push it to the metal, flow meter drops to 4 bars. That can't be right so I'm assuming that the FAV isn't pulsing anymore with higher F-IDC and it is causing the flow meter to read wrong. Is there any way to reduce the F-IDC seen by the HFS-3 (change some components on the board)?

[Richard L]

The FAV should not be operating at the static mode. Re-trim your IDC gains until the red led does not comes on anymore. We did not put a clamp on the IDC & outputting to the FAV.

Remove the restrictor and let me know if make any difference.