kilogram

iTrader: (1)

Neither of these are original ideas, but maybe they're worth talking about since I haven't seen them mentioned here. It's also worth mentioning right away that these are ONLY useful if one is tuning with Uprev, since the changes must be compensated in the software. Both will also require retuning, the same as switching to a GT or HPX MAF, but these are both damn near free. The first is more theoretical than the second, but I'd like to discuss both so I'll start with the first:

1. Nevermind. Forgot to account for laminar vs. turbulent flow. Only works for laminar flow.



2. Voltage manipulation.

This is probably the more useful of the two, and the method I'm going to use. Turns out the Hitachi slot-style MAF is a 0-6v sensor, not a 0-5V sensor. In fact, it will keep reading airflow until the internal amplifier hits its output limit. The limiting factor is the ADC inside the Z's ECU, which can read a max of 5.1v. This is a very common sensor that many OEMs are switching to now because it's cheap and it offers a wide flow range with very smooth output.

Here's a graph of voltage vs. flow for the Ford version of the sensor (same unit, different connector):

We're looking at the 07GT Spanish Oak plot. This plot agrees with my own datalogs in terms of g/s vs. voltage, so I'm considering it reliable enough for illustration purposes. Look at the direction it's going when it hits 5v and imagine what the curve looks like if we extend the axis to 6v. That's a lot of unused airflow measurement- roughly the neighborhood of 550whp.

All we need to do to utilize that extra signal is to reduce the MAF's output voltage by about 20% to keep top end flow under 5v and recalculate the MAF table so that previous values sit at 20% lower voltages. More on that in a minute.

The easiest way would be a simple voltage divider. It's possible it could be done with a single resistor, since there's already a 199.5KΩ pull down resistor on the MAF signal line within the ECU. That gives us a resistor value of (199,500/(5/6)-199,500) or 39,900Ω. A single 40kΩ could do it. I've got a removable trim pot that I plan to test before making the install permanent.

So all that's left is to recalculate the stock MAF table and move the values accordingly. Here's an Osiris ROM for an 04.5 DE engine with the MAF table recalculated and extrapolated to higher flow values (copy/paste into your own tune's MAF table): MAF Worksheet

For that MAF table, you'll need to multiply the K value by 1.4, since the values are cut by ~40% to allow for more flow up top. The bottom end of this table looks kind of like the Uprev GT MAF, honestly, but it's a bit fatter in the middle.

I basically moved the stock values to 20% lower voltage points (interpolating when necessary), then extrapolated the top of the map where the original values flattened out. This will need to be tuned if anyone wants to try it. I will be moving to this setup in the next few weeks in preparation for a Vortech install. I'll update it with a better model when I get a chance to tune it.

As I said, this is nothing new. Other platforms have done this. The MAFia tuner for Mustangs essentially does just what I'm proposing here- cuts the MAF voltage by percentages to utilize the upper range of the sensor. The big advantage here is that you do not have to buy a $350 "upgraded" MAF, nor do you have to shell out that amount should the MAF ever fail. Just replace it with a new, OEM-grade unit. This could be very good for the HR guys, since this could potentially save them $700.

Thoughts? I'll keep this updated when I have time. I know there isn't much traffic in this particular forum.

I've also been thinking for years that the MAF sensor could be replaced completely by averaging the voltage outputs of a MAP sensor and a frequency (rpm) to voltage converter, since the MAF is essentially just directly measuring a function of airflow that is dependent upon manifold pressure as a result of the throttle position and the engine's rpm. It would require retuning for any change in VE though, unlike a MAF sensor. That's beyond the scope of this discussion though..

(Also, yes I'm aware of today's date. No, I'm not f***ing with you.)

kilogram

iTrader: (1)

Update:

Finally got around to running the voltage divider setup and it works as expected. Where I used to hit 4.09 on the MAF I'm now hitting 3.53. I now have quite a bit more headroom for boost.

I made a quick and dirty excel sheet to convert a MAF transfer function to the new function, and it's posted here (download it).

There are headers at the top for where to paste the old MAF table and where to copy the new table. It's an open office document, and for some reason it puts apostrophes and spaces when I copy/paste the MAF table (it pastes the hex values), so you may need to clean it up a bit before it will spit out the right table. You'll also have to multiply your K value by ~1.4, then tweak the K again real quick on the first start up..

I copied/pasted a table I'd been tuning for a while into it and then flashed the car with the new ROM and it started right up and ran very well. Once the K was zero'd in, it runs almost identical to how it did without the voltage divider in place. I'll need to tune the MAF table again a little bit, but nothing major.

The method I used was 4 10KΩ 1% resistors in series spliced into the MAF signal line (the orange wire). These resistors act with the 199.5kΩ inside the ECU to drop the MAF's voltage by approximately 15%. This allows the ECU to read the full 0-6 output curve of the MAF sensor. It should be good to about 500whp, according to the graph of the transfer function.

djamps

iTrader: (1)

After installing the divider, couldn't you use the multiplier tool in rom editor to increase the MAF table values by 15%? Couple of clicks and done...not sure why any interpolation or fudging is required as this is a linear change, not a different transfer function altogether. Switching to HPX or GT actually changes the transfer function completely and requires alot more work.

kilogram

iTrader: (1)

I thought it was possible that way at first too, and I wish it was because that would make it much simpler. It isn't actually a linear conversion though. We're changing the signal voltage, but signal voltage doesn't linearly translate into the data word for the MAF table. It's a polynomial of several degrees.

We technically are creating a new transfer function- we're compressing the table a bit and shifting the values down so that (for example) the value at a new voltage of 2.5V corresponds with the data word the old function had at 3V so that the table is delivering the correct value for the new voltage curve.

Part of the issue is that the data word has a max hex value of FFFF or a decimal of 65535, which is why the table has to be divided down so that the new max fits within the max data word and the K has to be multiplied accordingly. I've only extrapolated where the transfer function goes after the old max 5v value, so I may not have it quite right yet. Looking at the

When I look at the multipliers that generate the new table from the old in the excel sheet, they start off pretty much lower, come close around 1V, but the percentage difference increases as airflow increases, up to about 6% at higher airflow. It's not a constant percentage across the board.

Once I get it tuned a bit more I'm going to compare it to a stock transfer function and see how it scales out. I'd love for it to be a few clicks and done, but I don't think it's that simple. I could be wrong. I hope I am.

Like I said, I'll keep updating this as I know more. I do know that it works the way I thought it would though, which is good news for anyone who would rather buy 4 10 cent resistors than a $350 MAF.

kilogram

iTrader: (1)

... Nothing to see here

npr350z

iTrader: (1)

http://www.treadstoneperformance.com...+Hitachi+Style



There's a fellow around here that used one of these on his vortech to extend the maf range.

kilogram

iTrader: (1)

Yup, the 3" tube will give you about 19% more flow. That's certainly one method. A lot of the Mustang guys do it that way.

I think Treadstone makes a flange for the Hitachi MAF as well that, if milled flat, would work as a spacer. Not sure how thick it is though. The voltage divider is probably still the best choice for big power or for HR guys who need to buy 2 MAFs or tubes.

I did another datalog/tune today and I'm surprised that the car actually drives better than it did on the map I spent a few weeks tuning.

npr350z

iTrader: (1)

You should...just for funsies make some maf scaling tables for the 3" and 3.5" tubes. Cause you're just a swell fellow!

kilogram

iTrader: (1)

You don't need one for that. The reason we need one for the voltage divider is because we're changing how voltage translates to a data word.

Velocity in the tube is inversely proportional to the cross sectional area of the tube. Larger tube has lower velocity for a given mass flow rate. That change is given by the equation

r²/R²

where r is the smaller tube radius and R is the larger tube radius.

We're not changing how voltage translates to a data word, we're changing how a data word relates to airflow, and that change is constant across the airflow of the tube. Therefore you only need a multiplier for K or for the MAF function. No extra table needed.

As I said in the first post about the velocity method, I wonder how big you can go before you slow the airflow across the sensor to the point where it causes issues with idle and low load.

The pro to this method is that it's probably easier to tune. The con is that you only gain about 20% overhead with a 3" tube and it's still a bit more expensive than a few resistors. Depending on how high the sensor will still read airflow, you could conceivably get to 600whp with the resistor method.

I just did it that way because I enjoy electronics and math.

350z-Helsinki

iTrader: (1)

This is great stuff. I just maxed out my MAF and this could be the solution. So what I need to do is cut the orange wire and put a 40k resistor into it. Then just copy the MAF-values you posted and multiply the K-value by 1.4. Am I right?

kilogram

iTrader: (1)

Use the spreadsheet I linked to in post #2. Put your current MAF table into it and paste the resulting table back into uprev. I suggest you make a copy of your current ROM so you can revert back if you need to, just in case.

That will get you most of the way there. You may have to fine tune a few areas again, but nothing too big. The K will likely be off a little bit, just do like you did the first time you tuned it: start the car and adjust the k value until the idle AFR is good with the corrections within +/-5%.

I do recommend using metal film resistors if you can find them. They're blue-body resistors as opposed to the brown bodied carbon film resistors. They have a much lower temperature coefficient, so their resistance doesn't change as much with temperature. Carbon film can be as much as 500ppm (typical max is ~350ppm though), while metal film are typically >100ppm. A 40kΩ resistor with a temp coefficient of 350ppm will change by 700Ω over a 50°C temp increase, or 1.7%, and they typically decrease in resistance. A metal film with a temp coefficient of 100ppm will only change by 200Ω or 0.5%.

Probably not a huge deal, but the less things change with temperature the better. I haven't had any issues in ambient temps from about 15°F to 85°F so far, but I have been monitoring it.

350z-Helsinki

iTrader: (1)

Worked like a charm. I could only source brown bodied resistors but I will change them later. Thank you!!

jdm-v35

iTrader: (7)

implemented this mod today and it seems to help. Yesterday at 67% load and 4800 RPM I was at 5.07v. with the mod today at 53% load, 5288rpm I was at 4.23v. I have not tested it enough to be sure but it seems to be helping. However by those standards and my estimated power level I would guess this to max out at 430-450whp considering how close it was to 5v so far. Thanks for the info, and I hope I get to really put it to the test soon.

Conway_160

iTrader: (1)

Any Updates to this? I maxed out my MAF with my prochager today at 5300 at roughly 10PSI. I would like to try this before buying a new MAF.

kilogram

iTrader: (1)

I've been running it for several months now. I see about 4.1v at 7psi at redline right now (belt slip or boost leak, haven't had time to sort it yet). Given the logarithmic nature of the MAF scale, I'm pretty sure I'll have enough headroom to get to 12psi. And if I don't, I can up the resistor value and retune. If you pay someone to do your tuning, you may want to go a little larger on the initial resistor value (51kΩ is a common value) just to avoid the headache of having to retune. I do my own tuning though, so it's no big deal for me.

I also just had my Z emissions tested and the results were no different than the factory tune did 2 years ago despite the supercharger, larger injectors, and self-tune. I think I could probably tune it to pass without cats... Driveability is as good as stock, now that I got rid of the fuel system resonance.

jdm-v35

iTrader: (7)

I also have done this and it works great. You do need to know how to tune though and also a bit about hex to get the table working correctly.

Kilogram, in the excel table is there a field to modify the resistor value to go higher than 40k ohms? I am currently running the 40kohm resistor and used the table modifier and just had to fine tune dialing it in as you said, but I am looking to go to 51k or maybe higher as I would like to aim for 13-14psi now, and have room for more power later.

kilogram

iTrader: (1)

I think that moving to a 51k (~26% attenuation of the signal) you could almost just adjust the k value to get it idling again and then do a few datalogs at low load to see how things look. It's possible you won't have to adjust much at all. It doesn't take long to nail down the MAF table with djamps' tool.

From this, doing it by altering K to idle would get you pretty close, probably over-rich in the 1-2v area and closer to normal as airflow increases. I think it'd be easier just to do it and tune it and not bother with the table. Upload your ROM to google drive or email it to me and I can upload your ROM with the new MAF function when you're done for others to use, if you don't mind.

As I said, we still don't know what the flow limit of the sensor is, but we know it reads to at least 6v. It may go as high as 7. That's a LOT of airflow.

Conway_160

iTrader: (1)

What wattage of a resistor do I need? I got 1/4w for now.

kilogram

iTrader: (1)

1/4W is fine. It's a signal line, not a power line.

For anyone who wants to understand how to calculate this, power dissipation (W) for a resistor is V²/R. Maximum V the resistor will see is 6v, and total R is your resistor value + the 195kΩ pull down resistor inside the ECU. So, for a 40kΩ resistor, total R is 235kΩ. 36/235k= 1.53mW. 1/4W is 250mW, so even a 1/4W is way overkill for a line like this.

jdm-v35

iTrader: (7)

thanks for the info. is there a method you used to calculate the exponential change that you used to make the excel table if I wanted to take it a lot further than 51kohm so it would be able to handle even more flow/power?