1.2L TDI Honda Insight

[ninedee_golf_tdi]

Nuts, he made this less than 60 miles from my house and I never knew about it

 

[Scott_DeWitt]

I predict real world mpg's of 88-96 depending on driving cycle and style.

Although an improvement, that's not much more than the lupo's 84mpg...

http://www.usatoday.com/money/consumer/autos/mareview/mauto497.htm

 

[ryanp]

i've never seen a 1.2tdi for sale in the UK! the 1.4 is not much worse from a conversion point of view as it has a proper gearbox!

Ry

 

[nicklockard]

Although an improvement, that's not much more than the lupo's 84mpg...

http://www.usatoday.com/money/consumer/autos/mareview/mauto497.htm

Agreed. The 3 cylinder PD engine used in this project is about 2X too powerful and won't operate in its best BSFC range much of the time.

For comparison, see the Loremo page:

http://evolution.loremo.com/content/view/98/141/lang,en/

The conversion is excellent though. Good luck and great work JP.

 

[jpstaub]

All,

After 4 grueling days on the road, the Insight 1G made it across the country with no problem more serious than a minor coolant leak. The ultimate data point: a 92mpg average makes it a bastard of a time for the driver's ass. Also, waiting for the 10 gallon tank to run dry strains the bladder to its limit.

The electric motor side of the car still isn't working properly. There wasn't enough time to sort it out with a 31JUL show date in Washington State (one draw back of working for Uncle Sam). On the bright side, we did manage to convince the Honda ECU that an engine is turning under the bonnet. So, we're on the right track.

Additionally, there was a MAF trouble code that we couldn't solve before getting on the road. Correcting that problem might yield better efficiency. Time will tell.

Is 100mpg probable? I think so given: 1. A 92mpg starting point, 2. A fully operational electric side, 3. Correction of the MAF problem, 4. Improved ECU mapping. This of course presumes that the vehicle is driven in an efficient manner (no full throttle clutch dropping babe wooing drag races down Main ST).

I'll have some more YouTube stuff up within the next few days (I hope). First order of business is to recover from the journey.

Best wishes,
Jake

 

[vwmikel]

Jake, what is the MAF problem that you're having now? That ECU pretty much has the stock 1.2L maps in it but I'm sure there is room for improvement.

 

[Ski in NC]

Maybe the chip guys can confirm this: The ecu fuel maps in EU and NA diesels are set to meet emissions, with some hit on efficiency. My South Africa buds claim better efficiency with their tdi's, probably because those ecu's are mapped with no consideration for emissions. Or they are good BS'ers, which of course is true.

So the question is: Could third world ecu's be plugged into EU and NA tdi's? Any benefit there?

That may help in this project. I know emissions also must be met, but since duty cycle is now different for the motor, the emission tweaks in the EU ecu may not really help in that regard also..

 

[VeeDubTDI]

Glad you made it to WA without any big problems. 92 MPG is an awesome starting point considering that your ECU is stock and your electric system isn't working yet. I think that 100 MPG is an attainable target. Keep us posted!

 

[Scott_DeWitt]

Jeff,

have you thought of using the ECM from the 1.2 A2? I don't think the A2 doesn't have that funky auto-manual transmission the Lupo has.....

 

[vwmikel]

Jeff,

have you thought of using the ECM from the 1.2 A2? I don't think the A2 doesn't have that funky auto-manual transmission the Lupo has.....

The ECU in the car is actually for a 1.4L, but it has 1.2L map info.

 

[jpstaub]

MAF trouble

Jake, what is the MAF problem that you're having now? That ECU pretty much has the stock 1.2L maps in it but I'm sure there is room for improvement.

Mike,

I'll pull the code again tomorrow to relay exactly what the computer is saying. Based on my recollection, the computer isn't happy about the MAF value.

I'm also having a bit of difficulty with the brake pedal position sensor (F and F47). The ECU isn't happy with F. I can bring up an advanced measuring block with F, F47, and F36 which is the clutch pedal position sensor (1/1/0 without brake pedal pressed). F47 and F36 move with brake pedal actuation (1/0 to 0/1 with pedal actuation). I get nothing out of F (constant 1) and a trouble code stating that the value of F is incorrect. Is it possible to tell what positions F, F47, and F36 should hold in the ECU 121 pin connector from the ECU program? I think we've got the wiring right (based on the wiring diagram anyway) and the brake switch is obviously working based on observation. The engine seems to respond fully to the throttle pedal (1.2L SSP suggests that a bad brake pedal switch will limit engine power). Perhaps the advanced measuring block is incorrect (I'm using VAG-COM)?

The MAP value also falls outside the given range of values shown for the sensor by VAG-COM. Is this a result of overlaying 1.2L maps onto a 1.4L computer so that VAG-COM displays a normal range of values based on the 1.4L engine which may not match typical 1.2L values? Lots of of small details I've been thinking about. We're way off the beaten path so it's learn as you go.

Mark Malone contacted me about dialing in the ECU program (http://www.topgear.com/us/features/more/we-build-a-70-mpg-car/). I told him what we did to make the car run. He's expressed interest in working with you to see if improvements could be made and how they might be implemented given our base program.

Best wishes,
Jake

 

[vwmikel]

Mike,

I'll pull the code again tomorrow to relay exactly what the computer is saying. Based on my recollection, the computer isn't happy about the MAF value.

I'm also having a bit of difficulty with the brake pedal position sensor (F and F47). The ECU isn't happy with F. I can bring up an advanced measuring block with F, F47, and F36 which is the clutch pedal position sensor (1/1/0 without brake pedal pressed). F47 and F36 move with brake pedal actuation (1/0 to 0/1 with pedal actuation). I get nothing out of F (constant 1) and a trouble code stating that the value of F is incorrect. Is it possible to tell what positions F, F47, and F36 should hold in the ECU 121 pin connector from the ECU program? I think we've got the wiring right (based on the wiring diagram anyway) and the brake switch is obviously working based on observation. The engine seems to respond fully to the throttle pedal (1.2L SSP suggests that a bad brake pedal switch will limit engine power). Perhaps the advanced measuring block is incorrect (I'm using VAG-COM)?

The MAP value also falls outside the given range of values shown for the sensor by VAG-COM. Is this a result of overlaying 1.2L maps onto a 1.4L computer so that VAG-COM displays a normal range of values based on the 1.4L engine which may not match typical 1.2L values? Lots of of small details I've been thinking about. We're way off the beaten path so it's learn as you go.

Mark Malone contacted me about dialing in the ECU program (http://www.topgear.com/us/features/more/we-build-a-70-mpg-car/). I told him what we did to make the car run. He's expressed interest in working with you to see if improvements could be made and how they might be implemented given our base program.

Best wishes,
Jake

Perhaps we could meet up sometime and try to get some of these issues ironed out. Are you using the cruise control with your setup? If Mark wants to contact me about it, that's fine. I basically just set it up to run like the 1.2L to get things off the ground, but there is definitely more work to be done and I'm sure some improvements to be made.

 

[e*clipse]

Regarding the brake pedal error - this may be the same as the 68 pin ecu on a 1Z - The circuits sound very similar and the error is also similar.

When I installed the 1Z engine into my Toyota truck, I used the Toyota pedals and switches and ran into this error.

The clutch switch (F36) is a straightforward on/off switch that should be grounded when the clutch is off and open circuit when the clutch pedal is depressed.

The brake switch (F and F47) is actually two switches in one device.
F should be pulled to ground when the brakes are off, and 12V when the brakes are on.

F47 should be grounded when the brake is off and open when the brake is on.

You will get an error code (measuring block 6 on the 68pin ECU) if F, F36, and F47 don't read '0' (gnd) when the pedals are at rest.

Good luck!

 

[jpstaub]

Regarding the brake pedal error - this may be the same as the 68 pin ecu on a 1Z - The circuits sound very similar and the error is also similar.

When I installed the 1Z engine into my Toyota truck, I used the Toyota pedals and switches and ran into this error.

The clutch switch (F36) is a straightforward on/off switch that should be grounded when the clutch is off and open circuit when the clutch pedal is depressed.

The brake switch (F and F47) is actually two switches in one device.
F should be pulled to ground when the brakes are off, and 12V when the brakes are on.

F47 should be grounded when the brake is off and open when the brake is on.

You will get an error code (measuring block 6 on the 68pin ECU) if F, F36, and F47 don't read '0' (gnd) when the pedals are at rest.

Good luck!

I follow what you're saying except for brake switch F. Based on the wiring diagram I have it appears that when the brakes are off F is open (0V) and when the brakes are on F is closed (12V). I'm having trouble seeing a way to pull F to ground when the switch is open then provide 12V when the switch is closed.

I'll rewire my translator box and see if I can't at least get all 0's and make the trouble code go away to establish a baseline. Thanks for the input.

Best wishes,
Jake

 

[e*clipse]

Hello Jake,

I'm having trouble seeing a way to pull F to ground when the switch is open then provide 12V when the switch is closed.

One way to do this is to solder a resistor between the output and ground. This is done commonly with embedded computers and is called a "pull down resistor." When the switch is open circuit, the output will be pulled down to ground potential by the resistor. When the switch is closed, the voltage drop through the resistor will be equal to the input voltage.

The resistor can have a fairly high value ( 1000 ohms > 2000 ohms ) so the current through it will be minimal.

Are you using the Insight's switches? Does it use a double switch like the VW and Toyota?

- E*clipse

 

[jpstaub]

One way to do this is to solder a resistor between the output and ground. This is done commonly with embedded computers and is called a "pull down resistor." When the switch is open circuit, the output will be pulled down to ground potential by the resistor. When the switch is closed, the voltage drop through the resistor will be equal to the input voltage.

The resistor can have a fairly high value ( 1000 ohms > 2000 ohms ) so the current through it will be minimal.

Are you using the Insight's switches? Does it use a double switch like the VW and Toyota?

- E*clipse

E*clipse,

I'm using the Insight brake pedal switch. In the manual transmission version of the car it's a single switch. It's a normally open switch (brake pedal released--open, brake pedal pressed--closed). The switch controls a 12V input. With the pedal released the computer sees 0V and with the pedal pressed the computer sees 12V.

From the 1.2L/1.4L ECU diagram it appears F operates the same way. It appears to be a normally open switch (brake pedal released--open 0V, brake pedal pressed--closed 12V). So, my strategy was to parallel off the Honda switch (A32) to the F input (T121/32).

For F47 (T121/65) I'm using a PNP transistor controlled by the Honda computer input (A32) to switch F47 to ground as depicted on the 1.2L/1.4L ECU diagram. T121/65 is tied to the transistor emitter, A32 is tied to the transistor base, and ground is tied to the transistor collector. When A32 goes high (brake pedal pressed) F47 encounters an "open" circuit. When A32 goes low (brake pedal released) F47 is pulled to ground.

Something isn't working quite right. I'm getting no change in F with brake pedal movement in the F/F47/F36 advanced measurement block. F36 changes from 0 to 1 with brake pedal movement. I'm getting DTC P0571 Brake Switch (F): implausible signal.
F F47 F36
1 1 0 Brake pedal released
1 1 1 Brake pedal pressed

My strategy to control the clutch signal isn't working either (NPN transistor) since nothing happens when I actuate the clutch pedal but we'll table that one for now. The most confusing bit is F36 responding to brake pedal actuation. Suffice it to say that more trouble shooting is required. I believe our wiring is good but I'll be checking that again.

Best wishes,
Jake

 

[jpstaub]

Perhaps we could meet up sometime and try to get some of these issues ironed out. Are you using the cruise control with your setup? If Mark wants to contact me about it, that's fine. I basically just set it up to run like the 1.2L to get things off the ground, but there is definitely more work to be done and I'm sure some improvements to be made.

Mike,

I'm sure a face-to-face meeting would be helpful plus it would give you the opportunity to check out/drive the car around. I'm free this weekend if you've got some time.

As far as trouble codes the last time I pulled them I got the following:
P0571
P1671 Definitely open since I didn't do any wiring for a pre-heat indicator
P1503 No load signal wired for
P0101
P1556.

Best wishes,
Jake

 

[vwmikel]

Mike,

I'm sure a face-to-face meeting would be helpful plus it would give you the opportunity to check out/drive the car around. I'm free this weekend if you've got some time.

As far as trouble codes the last time I pulled them I got the following:
P0571
P1671 Definitely open since I didn't do any wiring for a pre-heat indicator
P1503 No load signal wired for
P0101
P1556.

Best wishes,
Jake

Jake, lets see what we can work out for this weekend. If you could, shoot me a PM and we'll try to figure something out.

 

[e*clipse]

Hi Jake,

Regarding the brake pedal, it may be easier to obtain a Toyota or VW switch. They are set up w/ one circuit that closes and another that opens at the same time. Apparently it's critical to get the timing right, or you get yet another error.

OTOH, if you're messing around w/ circuits to do the same thing, perhaps an N- or P- channel mosfet would work? They are voltage operated (rather than current) and it's easy and inexpensive to find one that will turn on @ about 3-5V.

It sounds like some of the circuits are messed up. I can't get into it right now, but:
F47: brake off - ground brake on - open
F: brakes off - ground brakes on 12V

Good luck!

** edit for clarity & spelling **

 

[e*clipse]

Hi Jake,

I had a little more time to think about it this evening.

From the 1.2L/1.4L ECU diagram it appears F operates the same way. It appears to be a normally open switch (brake pedal released--open 0V, brake pedal pressed--closed 12V). So, my strategy was to parallel off the Honda switch (A32) to the F input (T121/32).

Something isn't working quite right. I'm getting no change in F with brake pedal movement in the F/F47/F36 advanced measurement block. F36 changes from 0 to 1 with brake pedal movement. I'm getting DTC P0571 Brake Switch (F): implausible signal.
F F47 F36
1 1 0 Brake pedal released
1 1 1 Brake pedal pressed

Dumb question: you did wire to the "output" of the F switch? If wired correctly, that should work. However, it's going to need a pull down resistor, because you need the F switch to change from GND > 12V. If it is just an open circuit, the computer may not detect the change. I've seen this happen w/ my PIC projects and my conversion.

Regarding F47, I haven't convinced myself that the transistor is an open circuit, like an open switch. Since F47 requires a change from ground to open circuit, the computer may (i'm not sure) not see the "off" transistor the same as an "off" switch. Also, if F is misbehaving, the same behavior would be reflected in F47 with this circuit.

Regarding F36, you may have started some ground (or worse) loop between the Honda ECU and the VW ECU with the addition of the transistor. You could be injecting current into the VW ECU from the Honda ECU through the transistor base.
It may sound a bit kludgie, but a signal relay could pull off what you're trying. The coils on small ones pull can pull only 20mA.

The relay could allow isolation, because the coil could be connected from
A32 to the Honda ECU ground, or parallel to the Honda switch. The switch part of the relay would connect pin 121/65 to the VW ECU ground. In fact a 2p/2t relay would allow you to get all the on/gnd/off relationships right for the VW ECU and ensure isolation between the ECU's. You may want to wire in a diode acrosss the coil to ensure the inductive kick doesn't hurt the Honda ECU.

Good luck!

 

[jpstaub]

Pedal position switches

e*clipse,

Thanks for the suggestions. All very good points that will provide directions to explore. I like the idea of using MOSFETs or relays. After double checking the wiring and seeing if I can force F, F47, and F36 to behave properly, I'll see if I can't implement a MOSFET or relay solution.

Best wishes,
Jake

 

[jpstaub]

Overal gearing adjustment

I was soo going to do this a few years ago! glad to see someone else has. Can you give us more info on the technological hurdles... like how you regeared the tranny for the diesel and what you did about the ecu for the tranny/ car vs the ecu for the diesel? how did you get everything to play nice!? PS upgrade your batteries 150mpg should not phase you with a good driver and some programming on the system!

As a result of a World of Speed requirement for H rated tires I started looking in to what could be done with overall gearing through tire size. It turned out that I can achieve nearly the same overall gearing with a 175/70R14 tire. Presently I'm chasing down a set of Michelin Energy XH1's. They're supposed to be the berries in terms of low rolling resistance. I'm tempted to go even further with tire size to take advantage of the Insight's superior aerodynamic characteristics. If the 175/70R14's work as expected going bigger might be a thing to try. With the 175/70R14's I'm hoping for a 5-7mpg improvement on the highway as a result of a significant reduction in engine RPM at cruise.

Fly it like a rental,
Jake

 

[06SpiceRedTDI]

There is a big get together in Portland, OR on August 22nd (TDI West Fest). Looks like about a 4 hour drive from Anacortes You should drop down and show your rig off. Sounds like the ultimate car, Honda reliability with A VW Diesel, you can't beat that.

http://forums.tdiclub.com/showthread.php?t=248544

 

[bakerinva]

Little late on this, but this is AWESOME!! Thanks for sharing and great work!!
Like Rotarykid said, proves just how well light duty diesels would do if we were allowed to have them here in the US...hopefully soon...

Thanks again and great thread!

 

[jpstaub]

Missed opportunity

There is a big get together in Portland, OR on August 22nd (TDI West Fest). Looks like about a 4 hour drive from Anacortes You should drop down and show your rig off. Sounds like the ultimate car, Honda reliability with A VW Diesel, you can't beat that.

http://forums.tdiclub.com/showthread.php?t=248544

I'd love to hit this event but will unfortunately be away on business. Please keep forwarding events. Showing interested people the car would be a blast.

We're going to make Bonneville this year (World of Speed 16-19SEP) and run wide open on the salt to see what the car will do. If anyone is in the area be sure to stop by and say hello.

Best wishes,
Jake

 

[tasdrouille]

For best highway fuel economy, you want the tallest tire that will fit in the wheel well, the skinniest one that can take the load, which will fit on the smallest steel wheel you can get for that size. Don't forget to shave the tires, the less material there is to bend as a tire rolls, the less the resistance. For disc brakes you should install springs to help pull the pads apart as much as possible.

 

[nicklockard]

For best highway fuel economy, you want the tallest tire that will fit in the wheel well, the skinniest one that can take the load, which will fit on the smallest steel wheel you can get for that size. Don't forget to shave the tires, the less material there is to bend as a tire rolls, the less the resistance. For disc brakes you should install springs to help pull the pads apart as much as possible.

Why steel, tas?

 

[twob4s]

I ended up buying a Lupo 3L in Munich and had it imported by a contact after getting the Federal DOT and EPA to buy off. Before you ask, I took the thing to the crusher last Friday to keep myself out of jail. Exporting the car or destroying it was part of the deal to import it. Ridiculous really but that's the way it works.

Having the whole car was quite helpful. I was able to verify many important steps in the build which saved a lot of time and effort. Additionally, I was able to pull a bunch of other parts (fuel cooler, vacuum canister, exhaust system, etc.) that were used during the build. In the long run I'm certain that although costly, working from an entire car saved me a pile of money in the end.

What is the requirement as far as destroying the vehicle? Based on what you are saying, it seems you just have to destroy the shell and you can pull all the parts off that you wish?

 

[jollyGreenGiant]

jpstaub, you've become a celebrity with this one. I've seen many postings on this and several made it to my igoogle home page ( slashdot covered it for example ).

Again, amazing work and best of luck getting all the reamining little details worked out.

 

[tasdrouille]

Why steel, tas?

Because there is something called ventilation resistance, and steel wheels usually are best at that because of the simple way they are designed with the least area of openings. Isolating both sides of the wheel as much as possible.

You want the smallest gap possible between the disc and the wheel.

http://ecomodder.com/forum/attachment.php?attachmentid=1841&d=1222521673

http://ecomodder.com/forum/attachment.php?attachmentid=1842&d=1222521673