Confused with fuelling when switching to semi-seqential

[themouse]

I have been doing lots of googling to how the fueling works on semi-sequential as opposed to batch fire but cant seem to get any real answers. I have a Mazda 1800BP turbo with High-Z 550cc injectors (old RX7 ones) currently running in Simultaneous with 1 squirt, untimed injection on a MS2 with a v3.0 board. The upper end of the tune is great but I have never been able to get a great idle/economy. I used the squirt switching method (1 to 2 squirts ) to find the injector dead-time which came to 1.7ms which would seem to be right for old school high imp injectors. Now after doing lots of reading it seems semi-seqential for me would be a great idea as I do have large injectors and will be able to control the timing of the injectors as well as only fire half at a time but for some reason I am unable to use this mode at all. When I switch over (power cycle after making the changes needed) the fueling is so rich the car can hardly run, this makes no sense as now only half of the injectors are firing as opposed to all and the fueling constant is half of what it is on 1 squirt Simultaneous. With a dead-time of 1.7 I can not get the afr to normal levels at all, I had to drop it to 1.5 to be able to get any kind of idle, once off idle once again the whole map seems to be about 8 AFR (normally 14.7). Can any-one please explain what could be happening so I can get a better understanding of whats going on? Thanks

[racingmini_mtl]

With semi-sequential, each injector fires once per rev (twice per engine cycle). So you need to set the req fuel with 2 squirts/simultaneous (or 4 squirts/alternating). The injectors will fire in the correct alternating pattern regardless of whether you use simultaneous or alternating (the parameter is ignored when in semi-sequential mode) but the req fuel value is computed in TunerStudio and needs to be correct (half of 1 squirt/sim).

If that's already what you have, then your dead time value is way off. And I must say that 1.7ms seems very high even for old injectors. Around 1ms (or less) would be closer to what others have seen.

Jean

p.s.:

[themouse]

Hi, thanks for the reply. I have set req fuel with 2 squirts/sim, I just dont get how I can get more fuel than 1 squirt/sim untimed when they are alternating not all four at once. As for the deadtime, when set to 1.7 the afr was constant when switching between 1-2 squirts (I was expecting around 1). I have a good earth, the voltage is constant and I also ran a good size cable for the power to the injectors, I have a new fuel pump also with a good size power feed so I am at a loss why its high. I even ran it on a bench PSU so I could control the voltage (went from 11 to 14.5 v) to make sure the correction is right.

[racingmini_mtl]

Think about it. If your dead time is actually 1ms but you define it as 1.7ms, for each injection pulse you're actually injecting 0.7ms worth of fuel more than what the MS is computing. That will make it run very rich. So your dead time is likely way off and too high.

The switching method for evaluating the dead time is at best a coarse guess because there are too many uncontrolled parameters. The only way to get the real value is to measure it directly by using sets of known pulse widths and injecting into a graduated container.

Jean

[BlackBird_SR71]

OP. Just for an FYI. I'm tuning a Datsun L28 Turbo with those injectors right now. I couldn't find a reasonable answer on DT's and do not believe the large Dead times ( 1.2 to 1.5ms plus ) as quoted by some Forums. It doesn't make sense with 3 injector events per 1 revolution of the rotor. You can't have slow injectors with those many injector pulses.

A lot of the injectors built even way back in the early 90;s were a LOT faster than people give them credit for. Even the saturated injectors .

FWIW we have the RX7 550cc idling dead smooth at 14.0 on batch fire with default DT of .800 ms and .200 Batt correction . I can lean it out way further. Can't exactly remember the PW, but I think it was around 2.3ms. Standard 3 Bar ( 43.5 Psi ) Fuel pressure

I'll update the PW later. I'll be remote tuning this car later today. Haven't even bothered to optimise DT or battery correction yet. Just in initial setup stages

I've tuned several Datsun Turbo engines with these injectors ( and the Low-z version ) . And .800 ms seems to work well as a starting point

[themouse]

The switching method for evaluating the dead time is at best a coarse guess because there are too many uncontrolled parameters. The only way to get the real value is to measure it directly by using sets of known pulse widths and injecting into a graduated container.

Jean

Yeah, I had a feeling I have to do that but that is not an easy thing to do at home in my garage, it worries me trying to hold the injectors in the rail under pressure and trying to inject into a measuring jar. I just struggle to understand how even with DT too high when firing two instead of four injectors its that much richer but maybe its beyond by comprehension

OP. Just for an FYI. I'm tuning a Datsun L28 Turbo with those injectors right now. I couldn't find a reasonable answer on DT's and do not believe the large Dead times ( 1.2 to 1.5ms plus ) as quoted by some Forums. It doesn't make sense with 3 injector events per 1 revolution of the rotor. You can't have slow injectors with those many injector pulses.

A lot of the injectors built even way back in the early 90;s were a LOT faster than people give them credit for. Even the saturated injectors .

FWIW we have the RX7 550cc idling dead smooth at 14.0 on batch fire with default DT of .800 ms and .200 Batt correction . I can lean it out way further. Can't exactly remember the PW, but I think it was around 2.3ms. Standard 3 Bar ( 43.5 Psi ) Fuel pressure

I'll update the PW later. I'll be remote tuning this car later today. Haven't even bothered to optimise DT or battery correction yet. Just in initial setup stages

I've tuned several Datsun Turbo engines with these injectors ( and the Low-z version ) . And .800 ms seems to work well as a starting point

Thanks for the info, I'm keen to hear how it goes. My idle PW seems to be around 1.65 (DT lower than 1.7 as voltage 14.2 so has corrected .2 ms off) but I really struggle to get any consistent afr unless its around 12. I have seen posted DT of 0.73ms for the RX7 550 but that was the Low-Z type, I have found nothing on the saturated purple top type.

[themouse]

Been thinking about my idle issue and idle afr, when I ran the car off the bench PSU, the only part that was running the motor off the car battery was the coil and fuel pump. The rest, injectors, wideband o2, idle valve etc was off the bench supply. I did notice that when I turned lights on or radiator fan would come one the engine afr would go lean even though the bench supply was nice and steady, so maybe the issue was never the DT but something else, will the added load on the motor via the alternator make the afr go lean? I am missing something and it makes the car idle like a pig!

[themouse]

OP. Just for an FYI. I'm tuning a Datsun L28 Turbo with those injectors right now. I couldn't find a reasonable answer on DT's and do not believe the large Dead times ( 1.2 to 1.5ms plus ) as quoted by some Forums. It doesn't make sense with 3 injector events per 1 revolution of the rotor. You can't have slow injectors with those many injector pulses.

A lot of the injectors built even way back in the early 90;s were a LOT faster than people give them credit for. Even the saturated injectors .

FWIW we have the RX7 550cc idling dead smooth at 14.0 on batch fire with default DT of .800 ms and .200 Batt correction . I can lean it out way further. Can't exactly remember the PW, but I think it was around 2.3ms. Standard 3 Bar ( 43.5 Psi ) Fuel pressure

I'll update the PW later. I'll be remote tuning this car later today. Haven't even bothered to optimise DT or battery correction yet. Just in initial setup stages

I've tuned several Datsun Turbo engines with these injectors ( and the Low-z version ) . And .800 ms seems to work well as a starting point

OK for anyone interested or wants to know the Denso P/N 195500-2020 I am using that came from a S5 RX7 with the testing I have done today have a deadtime of 1.5ms @ 13.2V verified with real fuel measurements I took this afternoon and plotting on a graph. This backs up what I found using the 1-2 squirt switching which gave me 1.6ms (0.1ms out is not bad for a crud way of doing it). All the go-ogling I did suggested 0.75ms but that must be for the peak and hold version not the saturated type.

11.2v 1.65ms
13.2v 1.5ms
14.2v 1.4ms

Hope this helps someone, now hopefully I can start with getting a more consistent tune.

[BlackBird_SR71]

I think you need to read Jean's answers again and have a re-think. I believe you aren't fully understanding the relationship between Dead Time, amount of fuel required and Pulse Width.

Bottom line. Can you get a smooth and steady idle AFR of 14.0 or lower at 1.5ms DT? Have you even tried 0.8ms to .750ms as a DT on your engine with the correct Injection squirt settings and correct Req-Fuel settings? It would help if you posted your .MSQ.

Right now I have my Customer running batch fire on an Datsun L28 Turbo with Hi-Z 550cc RX7 injectors. DT is set to 0.800 and Battery VoLtage id set to .200 at the moment. Further testing of Volatge compensation is required, and based on experience, I think we will probably end up around 0.150 Volt Compensation. Pulse Width is rock steady as well.

Bottom line:. We have a rock steady idle at 850 rpm ( acn go lower if we wish ) , pull 20 in/hg Vacuum ( Which is excellent on a Turbo motor with only 7.4 CR ) and have a AFR of 13.8 to 14.0 .... which is just where a Batch Fire is happiest. A batch fire can't run as lean as Sequential at idle.

I can lean it out further, even as far as 14.7, but I lose manifold Vacuum and the idle starts to get a bit erratic. So the engine and injectors are telling me they are happy at these settings.

One Caveat. Owner says these are 550cc RXZ Hi-Z injectors. They Ohm out as Hi-Z. But they may be off some other model of Car. Like a Toyota. I'm going to get him to take a picture of Injector tops and check the color. I'll confirm colors later. There is a possibility that they may not be the same Denso P/N 195500-2020 as you tested. A small but possible important difference. My customer is 3,000 miles away from me. So I have to go on the info I am given.

But right now I'm hesitant to change my settings on a system that is working correctly. With respect, based on unconfirmed methodology of someone who is having problems with their idle. I think you can see my point. Right now I'll stick with what is actually working well for my customer.

Edit: I've also seen injector driver circuits that were built differently or with mistakes. That adds to the problems.

[BlackBird_SR71]

FWIW. Here is a link to an excellent site with Injector dead times for many popular OEM injectors. Research done by Injector Rehab. Unfortunately the RX7 injectors are not shown.

But what is important is that there are no Hi-z injectors in this listing ( and there are a lot pf data points ) that are even remotely close to a 1.5ms Deadtime. Highest is around 1.17. 1.5 ms is very laggy, and quite frankly, Nippon Denso has better technology that that. Even in the 80's.

http://injector-rehab.com/shop/lag.html

Start comparing the Japanese manufacture 440cc to 550cc Hi-Z injectors. Many are Denso even if the engine manufacturers is only listed ( Toyota, DSM, Mitsubishi ). Some have surprisingly low DT's. And the Bosch Hi-Z injectors have really low DT's.

[TurboBob]

That's great info. I would consider it just a rough guideline since the driver topology is not known.

Bob

[themouse]

I am quite happy to post my finding in case they help someone else, but also that is on my setup with my drivers fuel rail etc. I did tests at 2.0 3.0 4.0 6.0 8.0 10.0 and 16.0ms and there is a distinct curve (from what I have read that is normal. The curves were the same for the three voltages I tested at 11.2, 13.2 and 14.2 but offset from each other which I would expect). Once the graph was plotted I intersected were the graph was 0 and thats the results I got. I then tested those results on the bench and found that at under 1.5ms the injector tone changed and there was no fuel.

I do think though that if the results are that close to the 1-2 squirt switch method that it does verify the results for me. I am more interested in getting a stable idle with electrical loads changing which I could get before I could get a stable idle but only about 13.0AFR. Also I not sure how this affects the VEAL calculation but with DT set to around 0.8 it was unable to tune correctly (Megalogger didnt work either where as I have not had a problem with a higher DT). On the miataturbo forum one of the guys there will not run under 1.2ms on the 450cc version of the High-Z injectors as he has found tuning issues also with anything less.

But once again these are my results for the setup I have. I am an electronics technician so I don't think I have built the drivers incorrectly, I also had it checked by a NZ megasquirt guy who said it was fine.

[themouse]

OK back to my original question, with un-timed bank injection (all four injectors) the car wants to idle with the PW around 1.7, with timed semi-squential it wants to idle around 1.4ms which is in the uncontrolled/unreliable area of the injector. So how can two injectors be putting out more fueling than 4? Even the required fuel value has to be half of 1 squirt simultaneous. I really dont get it and no one seems to be able to explain it. Maybe I should give up on the idea of running semi-sequential (my car did used to run on it with Evo 2 510cc Low-Z injectors but my tuner put it to batch).

[themouse]

BlackBird_SR71 I have read and reread Jeans answer many times over and he doesnt really answer my question. If I understand 'why' I get more fuel using two injectors at a time instead of 4 then I have a better chance of working out the 'how' to get the right result. When I was testing the injectors when the PW dropped below 1.6ms the sound the injectors made was really strange and all I got is what looked like vapour. I had to drop the DT to 1.2 in semi mode to even get any kind of idle but it was really unstable AFR sometimes 13.0 other times 17.0, I believe this could be due to the inconsistent fueling at that low PW. I brought my MS2 v3.0 when they were relatively new. Perhaps the drivers have changed over time to a better FET. Who knows, but I have had a constant battle to get a decent afr at ldle, good job I dont live in the UK or it would fail its test as they require 14.7 I believe. I am basically now at the point of giving up, and without understanding whats going on I dont stand a chance. I will look at posting my MSQ later on.

Think about it. If your dead time is actually 1ms but you define it as 1.7ms, for each injection pulse you're actually injecting 0.7ms worth of fuel more than what the MS is computing. That will make it run very rich. So your dead time is likely way off and too high.

The switching method for evaluating the dead time is at best a coarse guess because there are too many uncontrolled parameters. The only way to get the real value is to measure it directly by using sets of known pulse widths and injecting into a graduated container.

Jean

This is what I dont get, I understand what is being said here except if the DT was way too high then the issue should be worse with 4 injectors firing at once than 2, esp when the req fuel is double in batch mode too. But I took the advice and did the graduated container test but really didnt give me a DT that much different.

I think you need to read Jean's answers again and have a re-think. I believe you aren't fully understanding the relationship between Dead Time, amount of fuel required and Pulse Width.

Bottom line. Can you get a smooth and steady idle AFR of 14.0 or lower at 1.5ms DT? Have you even tried 0.8ms to .750ms as a DT on your engine with the correct Injection squirt settings and correct Req-Fuel settings? It would help if you posted your .MSQ.

Right now I have my Customer running batch fire on an Datsun L28 Turbo with Hi-Z 550cc RX7 injectors. DT is set to 0.800 and Battery VoLtage id set to .200 at the moment. Further testing of Volatge compensation is required, and based on experience, I think we will probably end up around 0.150 Volt Compensation. Pulse Width is rock steady as well.

Bottom line:. We have a rock steady idle at 850 rpm ( acn go lower if we wish ) , pull 20 in/hg Vacuum ( Which is excellent on a Turbo motor with only 7.4 CR ) and have a AFR of 13.8 to 14.0 .... which is just where a Batch Fire is happiest. A batch fire can't run as lean as Sequential at idle.

I can lean it out further, even as far as 14.7, but I lose manifold Vacuum and the idle starts to get a bit erratic. So the engine and injectors are telling me they are happy at these settings.

One Caveat. Owner says these are 550cc RXZ Hi-Z injectors. They Ohm out as Hi-Z. But they may be off some other model of Car. Like a Toyota. I'm going to get him to take a picture of Injector tops and check the color. I'll confirm colors later. There is a possibility that they may not be the same Denso P/N 195500-2020 as you tested. A small but possible important difference. My customer is 3,000 miles away from me. So I have to go on the info I am given.

But right now I'm hesitant to change my settings on a system that is working correctly. With respect, based on unconfirmed methodology of someone who is having problems with their idle. I think you can see my point. Right now I'll stick with what is actually working well for my customer.

Edit: I've also seen injector driver circuits that were built differently or with mistakes. That adds to the problems.

[themouse]

Here is my MSQ

[themouse]

And here's my Injector flow graph ms on the x axis

Screenshot_20180218_203009.png

[Dennis_Zx7r]

If this graph is correct and I'm understanding it correctly (y-axis ^= injected volume), these are the most non-linear injectors I've seen so far. I also don't see how you derive a DT of 1.5 from this data, and the R² must be very low.
Could you upload your raw measured data from the DT test?
As a comparison, this is typical DT data for some of the hi-z injectors I measured so far.

DTExample.png (14.73 KiB) Viewed 837 times

[themouse]

Sure here is my results :-

ms squirts ml@11.2v ml@13.2v ml@14.2v
2.0 4000 34 50 56
2.5 3000 43 54 58
3.0 2600 52 60 64
4.0 2000 60 66 69
6.0 1500 75 80 82
8.0 1000 70 73 74
10.0 800 72 74 76
16.0 500 75 77 78

[themouse]

Opps I stuffed my graph up im sorry everybody, it actually came to about 0.8ms. The youtube video I followed was so blurry I got the calculation wrong. now it looks like this

Screenshot_20180219_072204.png (25.98 KiB) Viewed 825 times

[BlackBird_SR71]

Opps I stuffed my graph up im sorry everybody, it actually came to about 0.8ms. The youtube video I followed was so blurry I got the calculation wrong. now it looks like this

Screenshot_20180219_072204.png

That jives a lot more with what I've seen. 0.800 ms is what I've run as mentioned and the cars I've Tuned have no idle or driveability issues ( Mainly Datsun L28 Turbo's )

BTW, your .MSQ has a few errors in it. Why do you have you EGO sensor disabled and Incorporate AFR Target = Don't incorporate AFR Target. There are many other things that affect the operation of the Tune other than DT. I think those are what may be causing your real issue.

Dead Time does not have to be bang on. Unless you are constantly swapping injector sizes. As long as you are reasonably close, you can dial them in with Battery Volatge correction and Dyno Tuning ( or Auto Tune will get you 90% there ) . Batt Voltage correction Table such as MS3 has is much better than the MS2 linear table, but you can make the Ms2 system work quite well.

I can probably help you out on this. I'll send you a PM.