eng92

iTrader: (3)

The data in the graph below was obtained by scoping the crank and cam signals at various rpm points and measuring the timing changes between them. The overall range of approximately 0-30 degrees is in good agreement with what is stated in the FSM.

EDIT - Before anyone rags on me - The y-axis should designate CRANK angle - NOT CAM angle



The cams in the 287hp Z33 are the same and it also has the same 30 deg crank angle range. The difference (if any) would have to be the slope of the ramp down from 30 degrees and what rpm level they are brought back to 0 degrees at.

The vtc solenoid pulse signal is displayed below in blue with the ignition signal to one of the coils in red for a relative frequency comparison. Data was taken at WOT at approximately 5Krpm.

SR20DEN

iTrader: (7)

The VTC gears are supposed to have a maxium sweep of 40º.

So how did you monitor the actual cam offset, in relation to the RPM and pulse?

eng92

iTrader: (3)

Top waveforms are crank and cam signals at about 6200 rpm (VTC solenoids are no longer being pulsed)

Crank is blue. Cam is red.

Bottom waveforms are at about 3000 rpm. The cam is approximately 30 degrees advanced relative to the crank.
]

The graphs above will probably only make sense to someone who is very familiar with the crank and cam signal wheels.

eng92

iTrader: (3)

My FSM shows an expected operating range of approximately 0-30 degrees when revved from idle to 2K.

I monitored the cam and crank sensors directly by tapping the wires coming into the ecu.

I was not too concerned with the pulse signal this time around. However from the waveforms above, you can see that the "ON" time is shorter than the "OFF" time which indicates that the cams are being retarded which is what is expected at 5K.

SR20DEN

iTrader: (7)

Interesting. So the crank signals are reading the gaps as pulses and the cam signals are reading the lobes. Odd.


Do you have longer graphs? I take it you've actually seen the cam pickups.

eng92

iTrader: (3)

No they both read the gaps. You have to think of the linear velocity of the crank signal plate vs that of the cam. The speed at which the gap passes the sensor is proportional to the diameter. Also, the crank is turning at twice the angular speed.

SR20DEN

iTrader: (7)

Which is why i asked for a longer graph, to see the bigger picture. (one full cam rotation). Which should equal six of the crank reference pauses.

eng92

iTrader: (3)

This is a little longer than one full cam rotation. It is the full waveform from the 3K WOT run above. You can't really tell much from this which is why I zoomed in for the pictures above.

SR20DEN

iTrader: (7)

Sweet, now do you have this same one at idle?

eng92

iTrader: (3)

This is about 1-1/3 crank rotations at idle.



As the measurements are essentially taken from screen shots, I have better accuracy if I use a smaller time scale.

DandyMax

iTrader: (7)

Nice work Dave; good info. Will have to look at this again more when I'm not falling asleep.

eng92

iTrader: (3)

Tell me about it. I have a 3,4 and 7 year old that will be getting up in about 4 hrs.

SR20DEN

iTrader: (7)

Yeah thanks for the graphs, I can read and compare them easier that way.

SR20DEN

iTrader: (7)

That tells me they aren't going to their full range. I just did a few rough measurements of the VTC sitting on my computer desk and everything looks to be closer to a 20º cam sweep than the 15º you're showing.

The Nismo VTCs are advertised at 55º (crank) which is of course 27.5º (cam), and I have seen them listed as being an additional 15º (crank) over the OE VTCs.


Now I think it would be cool watch the cam pickup with a tiny camera, strobed with an ignition timing light to see it in action.

eng92

iTrader: (3)

The FSM indicates that at idle, the cam timing should be in the range -5 to 5 crank degrees and 0 to 30 degrees at higher rpms. That gets us a little closer at 35 degrees total.

Also, it is quite possible that the IVT control maps are set up to keep the vtc working within its range so it is not driven up against a hard stop at either end of its travel.

SR20DEN

iTrader: (7)

It is designed so that one lobe hits it's mechanical limit before the other three do, which gives the others plenty of chamber space to be filled with oil and backed out. And it can be backed up with simple parasitic drag of the cam.

Full retard, as it is in the locked position, the master control lobe is pushed against it's stop.

IceY2K1

iTrader: (3)

Know anyone with a Z33/G35 you could test?

eng92

iTrader: (3)

No, I live in GM country. I have seen one 350Z in town in the 3-1/2 years I have lived here.

It is kind of amusing really. Instead of civics, all I see are sunfires and cavaliers with the huge spoilers and the fartcan mufflers to match.

IceY2K1

iTrader: (3)

Are you planning on a TS flash?

I'd really like to know if they deliver what they promise regarding tuning the VTC maps.

eng92

iTrader: (3)

You will have to wait on Vasily for that one.

With some off the shelf components and a little bit of electronic creativity, one can create their own independent VTC controller with feedback. All the necessary sensor inputs are already present and I know what form the required outputs need to be. I just need to build a little black box to make the magic happen.

SR20DEN

iTrader: (7)

I think a feedback box would be great to interface with a EB or EU, so those devices can be used to control the VTC solenoids, via additional injector drivers.

My method is going to be a litte more crude than yours. I want to set mine up to sever the ECU link @ 90% throttle so I can write my own map with the EU during the remaining 10%.

eng92

iTrader: (3)

It is good to start off crude because it often helps you to figure out what will and won't work before a whole lot of time and money is spent.

I was thinking along your lines initially with using a relay energized at >90% throttle to switch the VTC solenoid grounds between the ecu and an external controller.

Now that I can monitor the crank and cam positions directly, I will do a little bit of experimentation first using the EU as an open loop VTC controller. I do not have any experience with pwm solenoids so I am a little concerned about trying to operate them with such a reduced frequency signal (especially at lower rpms)

IceY2K1

iTrader: (3)

A feature on the EU that might be useful, but not sure....

1)Set CHA/CHB to activate per CH1-CH6 and use the timing offset.

Maybe using these in some way or combining the two channels together to operate both VTC solenoids might be a way around the low PWM signal when using injector maps to drive them.

eng92

iTrader: (3)

Possibly option 1) if the three pulses every 2 rpm are at equal spaced time intervals (240 degrees apart). Unfortunately there are no timing options with this setting otherwise you might be able to combine the 2 channels to get 6 pulses.

Option 2) would only give 1 pulse every 2 rpm.

IceY2K1

iTrader: (3)

What if you paired the injector wires to the EU and change the # of Input to 3 and the # of Output to 6, so you get....

Ch1/Ch6
1-0
1-120
1-240
1-360
1-480
1-600
2-0
2-120
2-240
2-360
2-480
2-600
3-0
3-120
3-240
3-360
3-480
3-600

eng92

iTrader: (3)

It is a little early so maybe not all of my neurons are firing yet. Are you talking about using injectors 1 thru 3 as inputs?
So on a V6 wouldn't 1-0 = 2-120 = 3-240 ?

IceY2K1

iTrader: (3)

What I'm trying to do is pick your brain or I'm hoping SR20DEN chimes in since he knows more about injection/ignition timing events then I do, how we can somehow combine the 6-inputs from the ECU to the 8-channel outputs from the EU.

Basically, we need to still fire the CH1-CH6 sequentially, but somehow incorporate the CHA/CHB to fire more often for the VTC solenoids when in RPM mode. Just wondering if it is possible to keep the order/timing, yet combine the inputs so the EU fires the injector outputs more frequently and out of phase by the correct amount.

Just brainstorming...I need to do some digging/reading, since I don't understand the timing events clearly.

Alex_V

the way I understand it the EU can fire the 7th and 8th outputs as 2 extra injectors. Which would mean you could fire them as much as you want anywhere in the rpm range as they would be mounted in the intake plenum. VTCs would use these 2 injector drivers, with a "fuel map" with a strange shape. But thats just an outsiders view...

~Alex

eng92

iTrader: (3)

I just wanted to make sure that we are all on the same page in terms of control of the VTC solenoids.

If this is a repeat for some of you, you can put your head down and have a nap.

Below is a picture I marked up out of one of my vendor's catalogs. It clearly illustrates the three types of pulses needed for control of pwm solenoid valves.



It is actually the inverse of my vtc solenoid waveforms as the ON pulse occurs when it is grounded (ie 0 volts)

You can see the dithering waveform present in the scope screenshot I posted previously. This keeps the valve oscillating a very small amount to prevent it from sticking. It also significantly reduces positional errors due to hysteresis (equal advance and retraction pulses not ending up at the same spot)

In terms of pulse duty cycle:
DC > 50% advance cam
DC = 50% hold cam position
DC < 50% retard cam

To use the EU for control, you have 16 different rpm points available in any injector map. In order to use intermediate cam positions, you would need to use adjacent columns with rpm points that are very close to one another. The left column would be to advance or retard and the right would be to hold. The rpm difference required will need to be experimented with as I currently do not know how long it takes the vtc to move full stroke.

Wait a minute, what is going to happen when you shift? Presumably, you will be up near redline and the cam is pretty well fully retarded and you shift but you need to get the cam to go back completely the other way and start retarding again. I guess some judicious choices have to be made when choosing rpms. If the leftmost column is selected at an rpm that will always be below your landing rpm when you shift then you can set it to >50% DC. The column immediately adjacent to that would need to set at 50% and at an rpm that is always above your shift landing rpm.

You can buy pwm solenoid valve drivers that you can adjust frequency and already have built in dithering functions. As input, some of them take an analog voltage or current. Varying the input changes the duty cycle of the output. I will have to see if I can find some cost effective units.

-EDIT- IceY2K1, This was not in response to your post. I started typing over an hour ago and got called away on something else.

eng92

iTrader: (3)

Because they are intended for injectors, their firing cycle is based on 720 degrees of crank rotation. Each of those 2 channels will only fire 3 times per 2 crank revs (on a 6 cyl engine)

IceY2K1

iTrader: (3)

So, assuming the EU injector pulse type will work, the Sub injectors at 100% duty cycle would be equivalent to 50% PWM.

That could at least hold the cam angle when the ECU backs down to zero by 6000rpm, correct?

eng92

iTrader: (3)

1 crank rotation at 6000 rpm takes 10ms. In the context of an injector which is based on 2 crank rotations, 100% duty is 20ms.

To hold position at 6K, the on/off pulses would be equal at 10ms (50% duty)

At SR20DEN's 7.5K redline, 50% duty would be 8ms.

The problem would be more at lower rpms because the pulses at a given DC become longer. At 3K, 50% duty is 20ms which is the upper limit for IPW with the EU.

This may be a concern at the lowest rpm in the map because you want a high ON pulse width in order to advance the cams as quickly as possible after your shift.

Of course this is assuming that you would be making the VTC maps for rpms from roughly 3-4K+ where the stock ecu starts to retard the cams.

eng92

iTrader: (3)

Just a re-post of the graph from post#1 with some more data points added.

I did several more logs in the 3.5-3.75K region because the dip in the first data set looked suspicious. I did four independent data logs at 3500 rpm and they were all within 1.2 degrees. The three data points at 3750 were within 0.7 degrees so I guess the dip is real.

I can't quite explain why the cam would be retarded over a 500 rpm range and then advanced again.
Maybe it has something to do with working in conjunction with the VIAS

SR20DEN

iTrader: (7)

That is exactly what it is.

But now we (I) need you to take measurments with a greatly reduced MAF voltage. You don't have to swap out injectors with the EU. You can cheat by using the additional injection map, effectively making the injectors larger.

eng92

iTrader: (3)

I presume that is to see the effect of load on vtc operation so that we will have a data set between the full and no load cases.

Scratch that. Now I remember. You are running larger injectors with the scaled back MAF to bump your ignition timing.

That may have to wait for a little bit. I have not been running my injectors through the EU for a while. It runs quite rich (~10-10.5:1) at idle and I was not satisfied with the recommended solution that had no explanation. My input and output IPW is the same on my data logs so I am not quite sure where that problem lies. The only obvious thing I can think of is that one injector is grounding out in my EU harness somewhere.

Back on topic:

As far as the ecu is concerned, if you can snap the throttle open to say 50-60% to go into open loop, what is the difference? The ecu sees a reduced MAF and the only other difference that I can think of is the throttle position.

Can I not just stick a block under my pedal to simulate this or does the ecu need to see that WOT TPS as well?

What % reduction in MAF voltage are we taking about?

eng92

iTrader: (3)

Well if you didn't like my proposal above, you should get a real chuckle from this one.

A couple of years ago I was experimenting with my ignition timing and what the impact of engine load had on it. At the time I didn't have any equipment to change MAF voltage, all I had was an OBD-II scanner to monitor timing.
I was amazed how much of the filter I had to cover up before I could feel and measure a difference.



The duct tape is wrapped around a piece of plastic. It is not stuck to the filter.

eng92

iTrader: (3)

Well I have had some preliminary success with VTC control using the EU sub-injector maps. It appears that the 50% duty cycle balance point doesn't mean squat when you are using such a low frequency to drive the solenoids.

Here is a comparison of the PWM VTC waveform generated by the ecu (red) with the EU sub injector output (blue) This data was taken at around 2000 rpm.



20ms is the longest pulse width duration the EU will allow you to enter in an injector map. At 2000 rpm, this would only correspond to an IDC of 33%.
Yet by 2K, my intake cam was fully advanced at approximately 40 degrees over the VTC off condition. (Yes I found another 5 degrees of cam retard compared with the idle values I measured earlier.)

Here are two crank (red) and cam (blue) data sets taken at 1500 rpm (top)and 2000 rpm. (bottom)
The larger space between the red pulses can be used as a reference point for both traces. The two leftmost blue pulses (cam lobes) have shifted to the left in the 2000rpm trace. This corresponds to an advance in cam timing of approximately 39 crank angle degrees.




Well I guess the cams advancing at such a reduced duty cycle is beneficial in that it allows control at significantly lower rpms. However, it is going to take a lot of experimentation to figure out how fast the cams will advance and retard with a given duration. The fact that the pulse frequency varies with rpm adds yet another variable in to the mix.

As an alternative, I found this PWM driver/controller:
http://www.appliedprocessor.com/PWMC400.html
Key features are that the pwm frequency is programmable and the output duty cycle may be varied using a 0-5 volt analog input.
Price for the kit is about $250. The 10 unit price is $189.

eng92

iTrader: (3)

I should note that if you go with an external pwm driver for vtc control at wot and retain ecu control for part throttle you will need to use a 4PDT relay or 2 DPDT relays to connect everything.

This is necessary because the pwm driver will vary the signal output to the vtc solenoid while the ECU or EU will vary the ground connection.

eng92

iTrader: (3)

Well the "magic" IDC seems to be in the range of 32-33%. Below that I measured no cam advance and above it the cams advance fully (given enough time)
To determine this I set all the IPWs in the sub-injector map to the same value, say 7.5ms. I then worked up slowly through the rpm range while monitoring the crank and cam sensors to see at what rpm the cam would start to advance. This was repeated for IPWs of 7.5, 10, 12.5,15,17.5 and 20ms. The results were consistent throughout the rpm range.

Right now it would be easy to create a map that would give two cam positions (ie. fully advanced or fully retarded), but that would be far from optimal.

The IDC needs to be fine tuned further to see what level it needs to be at to hold a certain amount of advance. This way it may be possible to step the cam advance down slowly in the mid to upper rpms.

Hoooper

i dont have anywhere near the understanding of everything that went on here, but if you kept the advance further along the RPM range instead of going to zero wouldnt that be a huge help in getting rid of the drop at 5500? i would say the dip at 4k is the VIAS, and maybe you could experiment with eliminating the dip to see any effects. is there a timing adjustment with the EU or is that for the future? can we get a comparison of cam angle and ignition timing per RPM?