MR2 Owners Club Forum banner
1 - 17 of 17 Posts

·
Registered
Joined
·
1,118 Posts
Discussion Starter · #1 ·
It has been a while since I posted in MR2OC, but I think it is still the best place to post fixes or common problems because it is easily searchable. FB is great for following what's going on or asking for immediate assistance, but searching on there is a PITA.

I have a stock MR2 turbo and want to keep it that way, I don't want to add an external boost gauge and adding an aftermarket boost gauge in the stock location leads to illumination mismatches (usually the gauge is much brighter than the dash, and it is not dimmable). So I thought about "fixing" the stock boost gauge to make it more useful. It really would be OK for boost levels of 15psi or less, so those of you with high boost levels - please just ignore this ;)

So, I collected some data on the stock boost pressure sensor and the stock boost gauge and it turns out both the sensor and the gauge are nice and linear. The pressure sensor maxes out at 17psi (further increase in pressure does not change voltage beyond ~4.8V) and the gauge in stock form sweeps from -5 inHg to 7psi (rather limited).

I was really impressed with the boost pressure sensor - resistance was 3.0k on the dot VC to E2 ( most resistors are +/-5% ) and the voltage coming out was solid/stable despite its 25+ year service life. Here are the numbers I collected:

Gauge reading should be interpreted based on the diagram below (these are just arbitrary labels I set on the gauge to take the data in the table):
76148


Pressure​
Sensor Voltage​
Gauge reading​
-5 (in Hg)​
2.3​
-3.75​
-3 (in Hg)​
2.42​
-2​
0​
2.61​
0​
3 psi​
2.98​
4.75​
5 psi​
3.33​
7​
7 psi​
3.59​
10.1​
10​
4.01​
13 (above the + sign)​

So what this means, is the gauge reads from -5 in Hg to 7 psi and that's why it is considered useless.

BUT, if the voltages from the sensor could be mapped to the gauge to read from 1 to 15 PSI, it would actually be pretty useful. Lets say the gauge was set to the following in PSI:

76149


I'll post if I make more progress on this project but for now, I wanted to just post the data I collected for posterity. Bottom line - the stock boost pressure sensor can be used for setups up to 17psi - it is actually a very reliable and high quality component.
 

·
Registered
1992 JDM RHD Turbo T-Top All Stock
Joined
·
64 Posts
I wonder if the little cylinder on the vacuum line going to the MAP sensor is some kind of pressure smoothing device that attributes to lazy readings on the stock boost gauge?

My aftermarket MAP sensor reacts instantly to any throttle input and I T'd it before that little white dongle-thingy.
 

·
Registered
Joined
·
962 Posts
It'd be impossible to really damp the pressure signal THAT much. I run a ~0.040" restrictor orifice between the intake plenum and my MAP sensor (has been the stock Rev 2 sensor, now using a 4th gen 3S sensor). This tiny hole only slows down maybe 1 msec, it's honestly hard to even tell side by side on a 250 Hz datalog. This does a tiny amount to smooth out cylinder to cylinder pulses in the MAP signal, but even with a .005" hole, which is about the width of a human hair (and difficult and expensive to actually manufacture like that), I doubt you'd see the pressure signal slowed down a huge amount. Maybe 10-20 msec to equalize if there was an extremely rapid change? But in reality, the MAP signal doesn't really change that quickly except going from boost to zero throttle very quickly.
 

·
Registered
1992 JDM RHD Turbo T-Top All Stock
Joined
·
64 Posts
So if the sensor reacts quickly and the boost gauge is accurate... why does it stink? Is the ECU just bastardising the signal from the MAP sensor?
 

·
Registered
Joined
·
962 Posts
I think it’s a direct 0-5V signal from the MAP sensor to the gauge. I’m betting it’s just internally damped to a high level, like how your fuel float sloshes around a lot on an empty tank, but the fuel level gauge takes a very long time to slowly move.
 

·
Registered
1992 JDM RHD Turbo T-Top All Stock
Joined
·
64 Posts
I think it’s a direct 0-5V signal from the MAP sensor to the gauge. I’m betting it’s just internally damped to a high level, like how your fuel float sloshes around a lot on an empty tank, but the fuel level gauge takes a very long time to slowly move.
How can it be a direct signal when the ECU uses this for fuelcut decisions?
 

·
Registered
Joined
·
1,118 Posts
Discussion Starter · #8 ·
The gauge gets a 0-5V signal from the turbo pressure sensor (at least in 3SGTE gen1 and gen2) - I think it "stinks" because it only covers -5 inHg to 7psi. If you have a real boost gauge, you know how quickly you sweep past the range shown in blue:

76168


As far as "slugishness", Note the pressure sensor gets its signal directly from the intake manifold. When you mash the go pedal, as the turbo is spooling up it really does take some time for the pressure to build in the intake manifold. It is an epic pressure battle as the turbo is pushing more and more air into the intake plenum, while the pistons are pulling more and more air from it as revs build. This is why boost will go up a lot faster in 5th gear starting at 3k rpm than 2nd gear starting at the same rpm. I've seen posts in the past that suggest setting boost controller levels using 4th or 5th gear rather than 1st or 2nd - because you might not reach full boost in 1st or second.
 

·
Registered
Joined
·
962 Posts
How can it be a direct signal when the ECU uses this for fuelcut decisions?
It's spliced into the signal for the MAP sensor. From the '93 BGB 3S-GTE electrical diagram:

76174


L-B is the signal wire color, and it's spliced at E17 before going to the engine bay fuse box (EA1), then goes into the cabin to the gauge cluster via IE3.
 

·
Super Moderator
Joined
·
2,852 Posts
I think the main issue is the fact that the gauge is really slow to react to the MAP sensor voltage change. I read somewhere that someone had managed to modify the little circuit board in the gauge cluster that controls the gauge to make it react like most aftermarket ones do. But if it's only getting 7psi which is approximately when the wastegate starts to open then it's still a little useless. Although with your aim to modify it to read 15psi that would help.

I'd look at the circuit board in the gauge cluster for this and see what resistors, capacitors and similar can be changed to make it faster to react.
 

·
Super Moderator
Driving MR2s since '89
Joined
·
4,279 Posts
Yes, there's bound to be a capacitor in there used for the damping. If you can figure out which one, and replace it with a much smaller value, then that would improve the responsiveness. Using a voltage divider for the input signal would get you 2x the normal 7psi range, without altering any of the other circuitry.
 

·
Registered
Joined
·
962 Posts
Yes, there's bound to be a capacitor in there used for the damping. If you can figure out which one, and replace it with a much smaller value, then that would improve the responsiveness. Using a voltage divider for the input signal would get you 2x the normal 7psi range, without altering any of the other circuitry.

Yep, that’s my recommendation on both fronts.

What about tossing in the guts from an aftermarket gauge with the stock needle? Might be less work to hit all the design goals?
 

·
Registered
Joined
·
1,118 Posts
Discussion Starter · #13 ·
Yes, there's bound to be a capacitor in there used for the damping. If you can figure out which one, and replace it with a much smaller value, then that would improve the responsiveness. Using a voltage divider for the input signal would get you 2x the normal 7psi range, without altering any of the other circuitry.
Yeah, you're right about the "damping". It is especially bad when going from full boost and letting off, the needle kind of just floats down. I think adding a voltage divider would make it even slower, an active circuit would work better - could be done with one of those tiny Arduinos. It's overkill, but they're less than $20. I'll look at it some more...
 

·
Registered
Joined
·
25 Posts
A few years a go I designed and built a replacement PCB to control the stock gauge better. I used an AVR micro (similar to Arduino) and did all the control logic in software. It worked pretty well and is still going strong today.

The issue with the stock gauge is that it's just not that accurate. The needle position varies quite a bit with battery voltage and temperature also each gauge has slightly different characteristics, so calibration is required.

I can share the files if anyone is interested?

76222
76223




 

·
Registered
Joined
·
1,118 Posts
Discussion Starter · #15 ·
A few years a go I designed and built a replacement PCB to control the stock gauge better. I used an AVR micro (similar to Arduino) and did all the control logic in software. It worked pretty well and is still going strong today.

The issue with the stock gauge is that it's just not that accurate. The needle position varies quite a bit with battery voltage and temperature also each gauge has slightly different characteristics, so calibration is required.

I can share the files if anyone is interested?
Very impressive Jon!! What is your gauge calibrated to? is it this ( 0-10psi ignoring the -4):
76228

or maybe you went with the same 0 point but had it max out at 20psi?

I'm not an op-amp guy but think I understand the basic functions of each of the blocks - you have an op-amp ckt to measure battery voltage, an op-amp ckt as a "level shifter" for the PIM input, a voltage regulator for the board - I assume it is 5V, the DW/ISP chip is a clock generator(?), your AVR micro, and the output driver to convert the 5V output to 12V - very nice!!

So did you have the board fab'd or did you make it yourself? If you have any extra boards I'd be interested in trying it out (of course with appropriate compensation), if not I can look at what it takes to fab a board - can't imagine it is cheap for 1 unit though.
 

·
Registered
Joined
·
25 Posts
Very impressive Jon!! What is your gauge calibrated to?
Thanks. I've calibrated the gauge to the min and max I see when driving. So min is vacuum on warm idle and max is full boost. About 18psi. The middle thick line is set to around atmospheric pressure.
It allows me to do a quick visual check to see if everything is normal.
I did originally spend quite a bit of time trying to calibrate it to 0psi, 5psi, 15psi, etc. But the gauge isn't really accurate enough to get repeatable results. Probably why Toyota designed their circuit the way they did!
I'm not an op-amp guy but think I understand the basic functions of each of the blocks - you have an op-amp ckt to measure battery voltage, an op-amp ckt as a "level shifter" for the PIM input, a voltage regulator for the board - I assume it is 5V, the DW/ISP chip is a clock generator(?), your AVR micro, and the output driver to convert the 5V output to 12V - very nice!!
The op-amp is a unity gain buffer, i.e. signal in = signal out. It's used so that no extra load from the ADC is placed on the PIM signal. I have a Gen 3 so the PIM signal is crucial for ECU operation.
The voltage regulator is indeed for 5v. The DW/ISP is just a programming header to debug/re-program the AVR micro. The output driver is an N-channel MOSFET which controls the low side of the gauge. The unregulated 12v goes directly to the other pin of the gauge, so the AVR reads the battery voltage through a divider to compensate for changes in voltage. The AVR uses PWM to drive the MOSFET.
The AVR software is pretty simple, it just reads the PIM signal and battery voltages, averages them and calculates the desired PWM on/off ratio to move the gauge needle.
So did you have the board fab'd or did you make it yourself? If you have any extra boards I'd be interested in trying it out (of course with appropriate compensation), if not I can look at what it takes to fab a board - can't imagine it is cheap for 1 unit though.
I had the bare PCBs made and then populated the components myself. It's a few years ago now but I don't throw anything away, so I should have some spare PCBs and components somewhere.
PM me and we can sort something out (y)
 

·
Registered
Joined
·
1,118 Posts
Discussion Starter · #17 ·
I happened to have an NA voltage gauge (from having replaced voltage for an Omori gauge in a previous MR2) and was doing some testing with it. My thought was that all the various Toyota/Denso gauges (temp/gas/voltage/turbo) use the same components so I thought I'd use it for testing out gauge behavior. Here is the data for posterity (feel free to skip)

First observation - the voltage gauge has a lot of hysteresis ( going from 0 to lets say the 3rd dash marking requires 10mA, but going from lets say the 6th marking down to the same 3rd dash marking requires 8.1 mA) I think this is why it is hard to get the gauge to read accurately. Assuming the turbo gauge works the same way, I can see why it is inconsistent. I think there may be ways to deal with this by applying an over current for some number of milliseconds (to get the needle to start moving) before settling on the desired value for, maybe, 5x of the overcurrent time. Much experimentation would be needed to check this out.

Anyhow, for posterity I got some data on the NA voltage gauge:

Serial inductor and resistor of the gauge winding with values:
inductor: 7.0mH Resistor: 72 Ohm
In addition, there is a 357 Ohm resistor in serial at the negative voltage side of the gauge. So total resistance for the gauge + resistor is 429 Ohm.

I tested current required to move the needle on the gauge alone (without the 357 Ohm resistor), since the needle movement should be proportional to current through the gauge. Because of the hysteresis, I measured the current after setting the gauge to some higher current value and then reducing current until I reached the desired marking. Current is in mA, voltage in Volts.

76233
76234
 
1 - 17 of 17 Posts
Top