Honda / Acura K20a K24a Engine Forum banner

781 - 798 of 798 Posts

127 Posts
Wow great progress! So close to finishing this journey. Very excited to see the outcome. Good luck for the final few steps

Sent from my G8141 using Tapatalk

5,718 Posts
Discussion Starter #790
Happy New Year to you guys!

How did it go?
Well, the week was cracy. It began 600 km away from home, where I recognized I forgot my ECU. Called my wife to sent it after. Said, done. DHL Express. The next day I worked on the engine harness and sensors to replace what was just for improvisation. Get bored the afternoon, watching the track infor available. Called DHL Express several times, but none of them known my package.

The next day we worked on a Drag engine project, this was the day the package should arrive till 12 o'clock. DHL Express Service said it was likely lost in the normal package shipment mode and it won't be here before Thursday (it was Thuesday). So we were calculating the best turbocharger design for this 8 s beast. Said, done.

On Wednesday the tracking status changed when I saw it 3 o'clock in the night. I was at the workshop at 7 in the morning (1 h travel time!) and sat on a seat at the street to catch the DHL van. I sat there for 10 h as there is no postal identification of the building, just google maps knows it. I placed a A1 sized cartonage on the building drive in to route the DHL guy for safety once I am not on there. Was a long cold day (1 °C to 7 °C). Thanks to DHL and my over hesitated travel start, it was a very long day.

On Thursday the tracking status changed again, I called DHL Express to ask where my package is. She said, it should come before 12 o'clock, as I booked the "before 12 o'clock" option. I said to here, "...yes, I booked that, but with the idea to have it 12 o'clock two days before today!". Most of the day I was again on the street, observing for DHL. It was 10 min. before the delivery is officially ended, around 18 o'clock, a black Van cam in and stopped at the building. I didn't recognize it as a DHL service car, as there was now identification like DHL. I got a phone call: "Äääähhh, sorry, you may be off the building already, but here is your package!" I said to him ", I am here, just come to you now!". Wuuuuuuuuuuuuuuuuhuuuuuuuuuuuuu, may ECU came 4 days late, cost me the sending, 2 dyno days delay and the recognition never ship with DHL again if it is urgent!

So with 2 days delay we trailered my car to the dyno, which was around 80 km away. It was around 14 o'clock on Friday. We pulled the car of the trailer, towed the car down to rollers of the MAHA dyno roll and started with the dynamic idle tuning. I recognized a pulled off coolant hose. One hour later we mounted it back. This was the most regride and unrule hose position ever. I layed on the dyno role with my head, my nose almost touched the gearbox. I was able to see the hose or able to touch and move it with my fingertips, but not both at one time. Finally we managed it, my back of my hand still knows it even today, I still can feel the edge of the fuel tank bracket. One our later a harness issue came up again. I called 911...or was it the harness specialist?! It was urgent and important, anyway, he still told me a story, so I went back doing my thing: testing the wires, wire by wire. But everything was fine. I tested it again from battery side and found a not visable broken fuse. Changed that, but the issue was still there. I begged the test stand guy to shake the harness while I am measuring, there we got it, it was a chassis harness connector. Around 20:00 I went to my host.

On Saturday I fixed the wires via old school style connectors, you know what I mean :D. Finally the ECU was again online, car starts and idles fine. We were ready to rumble. Brake in oil was already in, brake on the dyno was started. Zero VTC, lambda tuning while brake in the engine at 1500, 2000, 3000 and 4000 rpm. At 13 o'clock we had to calm down for noise emission reasons. I had to decide to stay for the weekend tune it on Monday and for sure I wouldn't be able to finish it or to travel back home and come back for the 2nd week of January. As this was the 2nd week from home and off my family I decided for the 2nd option. So tuning will happen on 7.-9. of January.

Cracy week, but also a bit dissapointing, as after the brake in the compression was still low, I saw values from 12.20-12.45 bar, as well as the leakdown was around 14 %. I will investigate this, as I expect a leakdown of < 5 % and a compression of around 16 bar. In any case I will finish the tune and maybe rebuild the engine if there is something faulty with the crank assembly (ring gap to big or clearance or piston to liner to big). We will see.

Take of with the car


On the dyno with the 102 RON Ultimate juce :D

After brake in, nicely colored runners
IMG-20191221-WA0004 (2).jpeg

The wiring issue

The old school style (improvisation)

After brake in compression test result





135 Posts
That's a pretty bad run of luck there, hope you have a better start to this year!

Did you work out where the air was escaping during the leak down test? Would be nice if it was just something like tight valve clearances,

5,718 Posts
Discussion Starter #792
That's a pretty bad run of luck there, hope you have a better start to this year!
Not too bad M111, I've learned a lot beside the theoretical aspects of the systems around engine development. That has a great value for me, even I've learned it the hard way. But all the failures could be repaired. Sure there are some open questions and further recognitions out of the data analysis, which has to be clarified.

This project is a massive challenge for me. From programming complex mathematical solver structures, studying the NA engine system, over creating 3D models of parts and negotiating missions/orders with fabricators up to handle 3rd party failures and my own ones. In addition there are innovative systems here in use, which where never tested before. All of that in an all in one shot approach, instead of doing a simple step by step development, is a more massive and more long-lasting challenge than anything I did before. E.g. idle tuning was complex. The parameter setup was different to a normal K-series, so the MAP-ignition timing-idle speed-relation, here 28 kPa/30°/2000 rpm, was cracy distorted compared to the typical idle needs of an K-series engine. A lot of controller relaxing adjustment was necessary while the EWP issues came up: 1. harness wrong connected, EWP PWM inversion needed, 2. venting of the coolant circuit with a not proper controlable EWP did take more time till it was properly controlable. Idle tuning is one of the basis steps and sometimes the most challenging part, especially if the engine is new, the engine setup is new and different and no predecessor tune is available. It was good to have the Infinity in my boat, the proper pre-setup'd protection moduls of it safed my engines ass and mine too for several times. It was like having 4 eyes on the incoming data :D. Last but not least I've got so much help from the Honda community of Berlin, this was amazing. I never experienced fully strange people to me helped me out with giving me a host, supply me with parts, supply of a word when I was down, support with the mechanical issues. I love those guys, I appreciate their help and hope I have a chance to help them once too!

Concluding this, there was more on the plus side as there was on the minus side, M111.

Did you work out where the air was escaping during the leak down test? Would be nice if it was just something like tight valve clearances,
I heard gaseous flow from blow by and intake, but only a slight. I have to say the local air pressure supply was only at 5.5 bar, the leakdown-tester-valve is calibrated for 7 bar. So the test result isn't of best quality. It was more a differential measurement as an absolut measurement value.

One of the guys may supply me a flexible endoscope to see over the injector bore the status of the valve guides (cold pressed in with 0.04-0.05 mm). Further I will test with another compression tester and I will give the engine brake in a further time window of 30 min before I check the valve lash, swap in the right engine oil. Of course we will check the liner honing after brake in.

I hope I can focus on the above mentioned approach and the tune of the engine instead of coping issues. Even if the engine isn't at best compression, I will finish the tune to have it already when after the engine rebuild. I optimized the InfintiyTuner pages for fastest and best overview tuning, as every day cost me 800 € and I likely need 16 h to have the sculpture of the calibration right on spot for street fine tuning.

A last point, which is still undecided, what injectors to choose for the 2nd stage. The 30° half angle two port spray pattern RDX, or the 5° half angle superior matched ID725? The decision describes the EOI-strategy to be tested and time is money.

RDX spray pattern:

ID725 spray pattern:

5,718 Posts
Discussion Starter #796
Any news?
That was a amizing week. On Monday afternoon we went back to Berlin and were allowed to stay at my friend Casi again. The man helped me so much, I hope I can make it up to you! My cart was still being tested in Finowfurt, where it stayed safely over the holidays and was still clamped.

1st day, Tuesday:
I arrived and started the engine, warming it up. It took about 1030 a.m. until the driver came, we divided the tasks, he was responsible for the test bench in the car, as well as load and speed, and I sat on the outside of the park bench and tuned the ECU. In the car it is too tight with the 17 '' laptop, so I got a park bench and put it down next to the cart to sit on. We worked our way through the delivery rate map of the ND camshaft from 1200-4000 1 / min in steps of 500 or 1000, depending on the gradient. That For each operating point from speed and load, set the optimal delivery level via the intake camshaft phase control (VTC) and get the measured lambda and target lambda to the point:

Challenge 1: the engine rang violently at full load. I had already turned the ignition back to 2 ° KW v.OT and almost dropped from belief, either that was my DAMPFHAMMER technology on the exhaust side or we are chasing after a wisp.
Challenge 2: the MAP oscillated at full load with up to 12 mbar and was therefore not easy to catch / coordinate.
Challenge 3: at 4000 rpm, the lambda probe began to make lean jumps of 0.4-0.8 s in length
Challenge 4: the parameter "NewVE", which suggests the probably correct degree of delivery due to the measurement, did not work properly and I was unable to clearly analyze the cause of 1st, 2nd and 3rd

So the day ended with a sack full of challenges. I left the test bench after I let the oil run out. Took the key to security with and from Casi, which meant just under an hour's drive through the north-east-south tangent around Berlin. Actually, I am so spoiled by the landscape of the Oberallgäu, but in this Honda nest in Berlin I always feel very well. The east, especially Brandenburg and Berlin, has something that I miss here in Stuttgart. The Honda community here is extremely helpful and well connected, the people are clear and have their own opinion and picture of the world and have all stayed on the ground. I miss that a little here in Stuttgart, it seems to me that people are more likely to play a role in what they would like to be ... but that is just what it meant.

When he got to Casi he asked as it was, I showed him the oil filter and said "... let's have a look!". We worked with a screwdriver and hammer and on the vice on the filter material. In terms of method, it was a mixture of can opener tactics and simply chiseled through the sheet in the direction of the layer, ie "cutting". It was exciting what would come out of it. The engine was flushed with flushing oil for 3 hours, got new oil and a new filter and what was now on the filter paper after running in ?! We really wanted to see it. The paper is firmly clamped at the top and bottom with Päpp / glue and is a very slippery story. Casis chose a cutting-roughing technique to cleanly separate it. Rolled out on two blue paper towels, we looked at the filter paper from left to right. The further we progressed, the more we found. In total I guess it was 15-20 spahn snippets. The spahn size was in the range 0.3-0.4 mm, the shape could be assigned to a drilling process. We didn't do a material test (magnet, heat, ...). We rather tried to make things nice to ourselves, it won't be bad.

2nd day, Wednesday:
Although I was working on the data until shortly after midnight, I could not clearly see how 1.-4. depended on why the day started with the error analysis. Shortly before, I got the two oil filters I ordered yesterday from the 2UR-GSE engine (Lexus IS-F, 5-liter V, because in the morning the oil change was also due. ATU and parts, that's a short story. Ask the parts master if When he got the oil filter from the IS-F, he asked back "Key number?" Me: "I didn't. Lexus IS-F ..." He again "... I need the key number." So I called Martin what a key number the cart has, since I got the first filter from him. He: "No idea. The ATU man should check it very quickly in the system, that would be a laugh if he wasn't inside!". The ATU'ler stubbornly "... key number?" Martin heard this on the phone and dropped a triad on the phone so the customer and ATU'ler overheard "Wat isn dat for ne Fatzke from Nullchecker ... § $ / &% § $ & / ((... juice store ...! "§ $% & / ()!". I love this direct way. At least without another request from me the ATU'ler walked to the computer and got and had already found the filter when I calmed Martin down and thankfully hung up. "Is there tomorrow at 9 o'clock, costs express surcharge." Well, now!

Back to the test bench, oil was now gone! So screwed in the oil filter, mounted the oil screw and tipped the good juice from Amiland into the box. The additional liter then disappeared up to 3/4 of the maximum display in the pipes and coolers. The used oil still had a purple tinge and was extremely tough (single-grade oil), which wasn't bad at all. The engine was warm around 10 a.m. and we went to the knock points again. Had to recognize at the point, the Ravenol intake oil was unfortunately not the reason. I had the slightest hope that a high volatility, i.e. a high proportion of volatile components in the oil, could be an additional knock source. So that could be from the list of suspects. Next, I took on external components. Knocked the engine and looked for loose components around the engine. I hardly believed my eyes, now that I knew where the screw under the oil pan belonged. The heat shield for the passenger axle was loose. I called Martin what the story was, he said he couldn't get the screw behind the manifold and he didn't want to invest the hour of fumbling. Aha, understand. The right screw was set quickly (fifteen minutes of fumbling), the left was actually a monster task and especially not available. We let it come to a test, I clamped a 350 mm long screwdriver between the sheet metal and the block between the primary exhaust pipe # 3 and # 4 so that no vibrating contact could occur ... of course at 4000 rpm and full load .. The knock was gone for the current ZZP setting. Jiiiiiiipiiiiiiiiiiiiieeeeeeee. We did the same test without the screwdriver, which was already getting slightly tempered, and see no change. Power strike. At 4000 rpm I was able to drive a ZZP of -33 ° KW in full load at 40 ° VTC (stop on the TODA VTC wheel). Knock free. Yeeeeeaaaahhh.

We took advantage of the hour and increased the tuning range to 5500 / min. Unfortunately I seemed to be happy too early. The lambda lean jumps were becoming more and more, so much more that the cart could no longer be measured reasonably under full load. I looked at the lambda sensor at full load and saw for the first time that the KAT and post-CAT came slightly brown-red. Isn't it going to be a heat protection function of the lambda sensor? Time to send Master off for lunch so I could look at the dates. For a yogurt, I first looked at the protection modules for cooling water, lambda, oil pressure and temperature. It quickly became clear that I was setting the limit curve for the minimum oil pressure too high. Because at 4000 rpm there was just 3.4 bar oil pressure / 105 ° C and the limit pressure wanted to see 3.5 bar, the module had to strike and took the fuel completely away until the pressure condition was met again. The pressure loss across the oil-water heat exchanger should not be neglected ... again a little wiser! The drop came only from 102 ° C oil, i.e. you had to drive into the working point and hold until the oil temperature rose so high, then it hovered around the 3.5 bar for a while because the protection module currently has a delay of 2000 ms. That means short and light shortfalls do not lead to the taking away of fuel, it has to be there for the fuel scissors to start. After all, the challenges 1 and 3 were mastered.

We continued tuning at 5000rpm in the LS camshaft. Now I felt really sick. The knock was back, this time I could clearly assign it to the engine with the long screwdriver on the ear and on the block. At 5000 / min / 40 ° VTC and full load, the engine was just knock-free at -15 °. I was really shocked. Especially because I "turn the burn" out of the knock, i.e. with a later VTC angle it was gone as if there was no knocking. This is normally a clear sign that the tuning of the exhaust system does not fit here. Five years of work to determine that? Should it really be that way. Should I have gone sooo badly wrong? The date was so clear that I was simply sick of vomiting. Unfortunately there was no time to surrender as there were two more challenges:

Challenge 5: Injector # 4 or # 3 seem to be leaking
Challenge 6: the LL-MAP is getting higher, initially we were at idle (LL) at 26 kPa, now at 36 kPa

It was already late in the day and we broke off for the neighbors who thanked us. Man, man, man, I wavered between frustration, disappointment and anger. Again small things stopped my march to the goal, again I had to analyze what happened and didn't. My best wife of all gave a boost to my motivation. Once at Casi he told me that he wanted to come tomorrow for the morning. I was really happy, otherwise I might have done one or two of the frustration. That evening I sat on the dates until 1:30 a.m. to find out where the knock could still come from. Unfortunately, I did not find the cause, but was more and more convinced that this was another parameter that, like the oil pressure limit curve, I did not have my sights on. The reason was, even if the exhaust system started to "plug", the transition would not be so blatant to be able to drive at 4500 / min -30 ° ZZP and at only 500 / min more VTC by 20 ° late by -20 ° ZZP ? There is something rotten!

That evening we discussed the measures against the removal. 3, 5 and 6. That was a good thing because I was looking ahead again and wanted to steer things in the right direction.

3rd day, Thursday:
My first action was a stop at Netto, Ausfdg. 3 needed the net solution badly. Glitzi, the cleaning sponge! When I arrived at the test bench, I announced that the construction sites had to be repaired before I could continue. "It will probably continue in two hours!" ... I said to the host on site. We removed all of the nozzles to understand what is sealing and what is not. We found that the Skunk2 carbon fiber fuel rail, unlike the OEM, did not depress the pen head, i.e. the small lever on the injector, at least only 1 out of 7! This means that the sliding sleeve at the bottom of the RDX nozzles is not moved in the direction of the O-ring, which means that no additional sealing pressure is built up. That there are 3 scenarios:

the nozzles do not have to be pressed down like it is done in series production at Honda. The ID725 nozzles have no sliding sleeve, and they are also much easier to use, so the already tight-fitting OEM nozzles. The nozzles must be pressed down by this small bending arm so that they seal tightly at the plenum. To do this, however, the distribution strip would have to be lowered, which would, however, cause the bending arm to be bent less because it would become shorter. The noble juice just sweated out when it came to these miserable singular work games without fuel, because the box shakes so that all the looser things are guaranteed to be loose (see heat plate, throttle valve, wiring harness, ...)

We chose 4th, i.e. Push the nozzles down by hand and observe whether they work their way up along the axis of the nozzle. This can be measured quite simply at the transition from the plenum to the distance to the injector body plastic in blue. So we threw out the ID725 and installed 8 RDX. Conclusion: Skunk2 does not know how the OEM hold-down system works. Nice mess, because I installed 2 of the expensive manifolds.

While I was dismantling the nozzles, Casi unscrewed the throttle valve and checked the seal (and inserted the glitter wool as specified), which led to no complaint. So he also dismantled and we let the engine operate again in the LL while Casi sprayed brake cleaner onto the DK flange. The engine answered yes to his indirect question and went out. Nice shit, the flap is still leaking. It was already 11 o'clock, our nozzle insights did not help, the flap had to be lowered again. We found that the upper left thread was already damaged. That was a big disappointment, as the thread was attached to a € 2000 prototype ASB. Where there is planing, there are shavings. That must have been the last time you screw it back on, or it may have happened before that. M8x40 steel screw on M8x10 flange aluminum thread is like the prince on the pea. Casi offered to provide me with his Helicoil tool kit. Hammer, that would be a clean solution. Challenge 6 is almost solved. We agreed to keep going and let the engine warm up again. This time I started the engine from the passenger door and touched the AEM plug and play adapter harness, which led to the engine shutdown. Me: "Hääääää ... what should that shit ... eat ?!" We tested the X-times, wiggled the wiring harness on the cabin side, but the engine shutdown could always be provoked 100-200 mm before the main connector. Together we condemned the junk from Amiland to a lifelong ban on our engines and praised DIY and JDM goods.

Casi should have been at home for 2 hours to cover the canopy he was building with a buddy. He canceled the appointment and stayed. Hammer! He conjured up a few Wago clips from his car, which we used to repair the 16-hour-old AEM adapter wiring harness. Do you want to know why the cable broke? Better not .... The sarcasm aside, they actually had 5 strands clamped on one strand, without adhesive shrink tubing on a 5 mm long wire sleeve. Can be done, does not have to be...! After the temporary repair, I started the engine, but it did not come from the starting speed (cold about 195 / min, warm about 270 / min). I wiggled off all the vital plugs: fuel valves, ignition coils, but without a positive result. We went to measure their passages, they were all there. It was strange. Before we went. What remained was to take the wiring harness home and measure it. We discussed the solutions for # 6 and then drove back without a minute of test bench mode. At 12 noon I had examined and measured and documented all 80 cables. Wiring harness was ok again, it seemed the problem was "downstream" in the electron path ... maybe. Chassis harness ?!

4th day, Friday:
The day started with the installation of the AEM adapter wiring harness. I then removed all nozzle plugs again and put them back on. Ignite the ignition to # 4 against ground with a spare coil and candle. Everything there. We had charged the battery overnight as a precaution. 9:30 a.m., ignition 1, 2, start, engine running. Fine, why not right away! Today's goal was to run through the critical points again in order to test them again with different settings.

Challenge 7: the gears no longer went straight away. Sometimes I had to slow down so as not to miss the enema. He said it came from the gearbox ... well, I suspected that too! After about 40 minutes of knock examination, it smoked horribly from the engine corner. I shouted "STTOOOOOOOOPPPP", the engine was immediately off. Root cause:

Challenge 8: Axle boot on the passenger side tore at the ventilation hole, you know the rest...the test bench driver said, "No problem, we drive further, that makes no sense", in trust in God and without a bag we went straight to:

Challenge 9: gearbox ventilation waved the flags with hello and a loud hello, that came close to a pressure cooker with an open pressure relief valve. Probably oil overheating ... probably because of the axle boot. I was devastated and broke off. Called the lorry that was on the here two hours later. From, heme, need more to do. You can do more with the gearbox. Now the damage balance has to be drawn. It went back to the workshop, with drizzle, to match my mood. I wasn't even angry or frustrated anymore, I was just appatic.

The aim now is to get the cart ready for travel, to bring it home and to strap the engine onto the engine test bench.

5,718 Posts
Discussion Starter #797
The DAMPFHAMMER technology is verified. Below you see the not finished tune (ign. timing was at -12°, VTC not at 40°) but the VE was already at 130 %@5000 rpm.


The engine health isn't bad. Just my compression measurement sucks, shows a mean value of 12.8 bar only....


But the leakdown values are in an acceptable range, not at aimed spot of 1 %, but acceptable...


The helicoil inserts in the TB flange...


5,718 Posts
Discussion Starter #798
I had a lot of stops on the dyno, the last before the gearbox broke...


...all over the grease troplets.


The after only 16 operation hours broken AEM qrimp fitting...

101358 can't qrimp 5 to 1 lead without any further safety guard like a glueing shrink tube!


Next stop caused by the non fitting Skunk2 fuel line. The injector must be pressed down by the fuel line at the lever arm tip. Instead of that the finger of this arm is outside the injector bore surface.


This is needed to push the sliding sleeve, which presses the O-Ring out to increase the sealing force.


The ID725 even had not the sliding sleeve, as it is fixed or to tight fixed. I took them out of the 2nd stage to replace it with the RDX. I am really sick of that low quality stuff, which is not well designed or not well functioning. I was 70 h on the bench and just 12 of them the engine was tuned. The rest I was coping the low quality stuff issues. Surprisingly always US aftermarket stuff:
  1. Skunk2 TB MAP oscillation because of a too big ID of the communication channel to the MAP sensor - issue solved (2 h, 49 cent)
  2. Skunk2 outer fuel rail diameter at the injector bore (OD) too small - issue not solved (1 h)
  3. AEM plug and play adapter harness broken wire - issue temporarely solved (4 h, further to solve it needed)
  4. ID725 injectors axial position not defined properly - issue not solved (0.5 h)
There were other issues like the handling of the Infinity during tuning, which I were responsible of. And there were a lot other issues related to the harness, which was the main issue. I learned alot of that. Don't buy US aftermarket stuff, don't relay to other guys and do as much as possible yourself.

The gearbox issue we clarified to 60 %, there was only 50 ml oil left. I was calling this morning the guy who was one of the guys reassembling it. He ment he didn't reassemble the gearbox. Anyway, it is likely the gearbox wasn't filled with enough oil and it heated up, so the axle boot on the right side, the one who exploded, had likely no proper prierced vent hole and produced a overpressure, which crack the vent hole to an bigger one. The gearbox housing itself was heated at about 150 °C or more. When I put on my finger on it, I couldn't stand the heat and pull my finger of immediately.
781 - 798 of 798 Posts