VW 9A - 2.0 16V MS3/3X

General talk about successes. See older information on MSRUNS forum.
Post Reply
walterclark1
Experienced MS/Extra'er
Posts: 292
Joined: Tue Jan 01, 2013 3:37 pm
Location: Dover, MA., USA

VW 9A - 2.0 16V MS3/3X

Post by walterclark1 »

This has been a project for the winter layover between hillclimb seasons here in New England. Since I am getting pretty close to having all the various functions working properly, have a dyno session scheduled in 2 days and now have confidence that it WILL actually work for my race engine, I thought I would start a "success" thread. This is my 1986 VW GTI hillclimb car that races in P3 of the NEHA series.

The engine was re-built over the winter of 2011-2012, after several race seasons on its stock block with way over 100k miles of road use in a 1990 Passat. The head had been reworked for improved flow during 2010-2011 winter. The head features porting, including ports matched to a 50mm intake and a TT 4-1 Swain coated race header, with the bore smoothed and enlarged to the valve seats and TT 268/276 cams. The fresh in 2012 bottom end uses 11:1 Wiseco pistons with Swain ceramic coating, the block is decked to optimize squish, and all moving parts are is fully balanced. The ignition and fuel have been managed by a Bosch KE-Jet from a 1.8 16V, with fuel enrichment at WOT driven by an Innovate LM-1 wideband feeding the CTS input on the Bosch ECU. Performance (Lambda and ignition final advance for max WOT power) was optimized for 100 octane fuel on a dyno in 2012. On that dyno it delivered just under 200 crank HP. About the second event of this past race season I came into a quantity of VP 109E fuel for a good price and began using it. It is highly oxygenated fuel so the amount of fuel being delivered at WOT by my enrichment increased automatically to maintain the required Lambda, but once it was clear I was running on just the 109 rather than a 109/100 mix, I found the Lambda was increasing (leaning out) at high RPM and the fuel delivery was maxed out from the fuel distributor. Several years ago I told myself that if and when I ran into a problem with the KE-Jet system that was not easily corrected I would switch to Megaqsuirt. Well, this was it. Sure, there are ways to get more fuel thru the CIS fuel distributor - but they are all hacks of sorts, involving questionable used parts from other engine types, and lets face it, despite Myth Busters proving it can be done...a polished turd is still a turd.

I decided to take this opportunity to do a few other things the car needed as well. The wiring harness was a mishmash of 27 year old stock wiring, various owner additions, hacks and deletions. The dashboard was ergonomically impossible as a race dash - everything was too far away to be reached while strapped in, contained a lot us useless stuff (the instrument cluster itself included) had no place for things I wanted and needed, and with my lower further back seating - the steering wheel blocked the instruments. The rules of the class I run permit removal of almost everything that is not directly associated with making the car go, turn and stop, except the requirement for a working horn (as an audible distress signal). And a lot of racers do remove everything - but based on the past 10 years of experience I wanted to retain brake lights, wipers and a defroster. So this was a good time to strip out and rebuild the entire electrical system, a new lightweight dashboard and modify the instrument cluster so the useless among the idiot lights became valuable to me.

Once I decided what all I needed and where everything would go, I drew up wiring diagrams for the power distribution, and the sense and control signals for the Megasquirt, and designed the dashboard.

A LOT of hours later this is what I have:

Maybe the most time spent on one component was with the trigger wheel and sensor (theDIYAutotune 36-1 wheel and flange mount sensor). (Sorry its so dark - almost everything is black)

Image

I spent a huge amount of time getting the trigger wheel to run true - bolted to the harmonic balancer - and chasing what turned out to be a very insensitive hall sensor. The flange mount Hamlin from DIYAT I am using now sits a comfortable distance from the toothed wheel and is rock solid.

The engine bay cleaned up real nice:

Image

I reused the stock Bosch/VW fuel pump/filter/pressure accumulator and added a 3 bar regulator to a BBM fuel rail to feed the 380cc high impedance twin spray pattern injectors. I am looking forward to altering the fuel pressure control from a shunt regulator to a PWM controlled closed end system when MS3 supports dynamic control of such a configuration. I modified the original (ported and polished) throttle body to use a VW TPS. I reused the distributor to function as a cam position sensor for sequential injection and installed a dual coil for semi-wasted spark. I added oil pressure, fuel pressure, EGT and IAT sensors. In addition I have a magnetic reed switch on the clutch pedal along with a dash mounted toggle switch and latching relay for Launch activation and control input.

The dashboard with new homes for the MS3 and LM-1:

This was taken before I was 100% done connecting the new wiring harness to everything so a few wires are still dangling.

Image

The dash is flocked to cut windshield reflections and while the new dash is larger and moves the control panels closer to me it is actually 3 pounds lighter than what came out. It retains the defroster and floor heater as racing up New England mountains can sometimes mean dealing with cold and damp conditions. 2011 Mt Washington was a great example...t-shirts at the bottom, waterproof parkas at the top. I have a RAM mount in the upper right corner of the control panel part of the dash for options like a GPS (actually handy at Mt Washington) or a future Android tablet (running Shadow Dash), some toggle switches for Launch control, Data Logging, Table switch, and TBD, plus 2 permanent digital gauges. The Instrument cluster center section which is dominated by warning LEDs was modified to give me a shift light, Lambda, and EGT warnings, and Launch active in addition to the original oil pressure and alternator warning lights. The LEDS have been swapped out for much brighter and appropriately colored versions as the originals were dim (27 year old specimens) and/or just the wrong color choices for the information, caution and warning indicators I chose.
Last edited by walterclark1 on Tue Apr 02, 2013 5:24 pm, edited 2 times in total.
A2 VW GTI 9A - MS3/3X
walterclark1
Experienced MS/Extra'er
Posts: 292
Joined: Tue Jan 01, 2013 3:37 pm
Location: Dover, MA., USA

The knock sensor is a great tool.

Post by walterclark1 »

As part of my "tuning" regimen I have been slowly enabling features like the launch control and the knock sensor module. This week, in anticipation of the first full throttle pulls on a dyno I started in on the knock sensor module. I am using a single Bosch knock sensor, but the module and the engine block permit a second so I plan to add that soon now that I know what more a second one will allow me to do (pinpoint individual cylinder knock).

I basically set the module settings using a calculated bandpass frequency for my 83mm pistons and started gathering an analyzing data logs of Knock In and Knock Retard under low loads (neutral) to figure out sensitivity and knock thresholds.

One thing I noticed pretty quickly when looking at the logs with MegalogViewer was that within the bandpass I selected (no matter what frequency I selected) I was seeing pretty substantial spikes. These were on the order of over 50% when the noise outside the bandpass region was maybe 10%. I am running a 98 octane blend of fuel and they remained pretty much independent of timing and Lambda. I removed the new stock heat range NGK Iridium plugs for a read and found all 4 of them identical but indicating they were running hot (VERY white ceramic insulator - no deposits of any kind). So they were probably causing at least some "knock" like symptoms. I had purchased the next two colder ranges of these plugs since I didnt know exactly where the plug need would fall, and I installed the next range colder last evening. The first data log in neutral looked very promising - under 10% within the bandpass. I will check it again today, do some partial load tests in the driveway and inspect the plugs.

An added note. I have a PC-scope tool and had that connected to the knock sensor input to the MS3 box. While the FFT (spectrum analysis) clearly showed an RPM dependent frequency peak (around 900Hz-1.0KHz at my idle), it did not show any peak around the calculated engine knock detonation frequency of 6.9KHz.

This was a great learning experience for me. The knock sensor module and MegalogViewer can help identify and/or confirm a too hot plug problem quickly if you know what to look for and none of this was obvious on my PC-scope tool.

Update: I may have misled myself with the use of the PC Scope. I noticed that yesterday when I repeated the "test" of the engine noise as seen by the MS3 knock module and displayed in MegalogViewer, the spikes were up to just under 50% in the bandpass region. Looking for a difference in the previous evening led me to realize that I had the PC-scope connected up AFTER I swapped in the new plugs, but not before. I reconnected the PC-scope and the Knock In levels went down as previously reported. I went looking for an explanation. It turns out the Bosch knock sensor has an effective output impedance of something greater than 1 MOhm and the mic input to the external sound card I use has an impedance of 20 KOhms. So it seems the sound card is loading down the knock sensor's output. Unfortunately the sound card uses a single chip solution that includes the mic signal input resistors within the chip so it is not something I can easily address on the card itself. When I have a chance I will construct a little buffer with a >10 MOhm input impedance and feed that to the Line in of the sound card to eliminate that loading effect.
A2 VW GTI 9A - MS3/3X
walterclark1
Experienced MS/Extra'er
Posts: 292
Joined: Tue Jan 01, 2013 3:37 pm
Location: Dover, MA., USA

Re: VW 9A - 2.0 16V MS3/3X

Post by walterclark1 »

I finished the high impedance buffer so now I can connect the knock sensor to my PC scope or just thru the PC sound card to listen to it while it is connected to the Megasquirt without loading it down. Marc At EFI Express showed me what to listen for and how to differentiate knock from drive train noise. Very helpful.

Speaking of EFI Express, I spend Thursday there tuning on the steady state dyno he uses. I now have the ignition and VE tables adjusted from idle and no load to 7200 RPM and full load. Tuned for most torque at all loads loads and good no knock performance. We ended up turning off AE during the runs as it was messing with enrichment even when the TPS and TPSdot were steady - not sure whats up with that yet. The final verification pull yielded 147 lb-ft torque from 5800-6400 RPM and 175 HP @ 6500 RPM at the wheels. My torque is above 130 from 3000 to 6800 RPM. Stock the engine was rated at 133 lb-ft @ 4400 RPM and 134 HP @ 5800 RPM.

After the session I was going over my logs and noticed something interesting. Above 6000 RPM and about 85% load I see the fuel pressure spiking downward as much as 10 pounds and back from the around 40 PSI the regulator had been calling for at lower RPM and full load to. These spikes are very short duration, maybe .5 seconds but nearly continuous from 6000-7000 PRM. We didnt look at or notice this on Thursday but it appears we compensated for it as the VE settings we established under steady state testing step up from 130 around this area to 150 within it. I plan to leave it as a step at least until I correct the pressure issue as the final "pull" shows nothing obviously wrong in the power curve or Lambdas. I began looking for the cause of the spikes and decided first to just swap the pumps and fuel filter. I swapped out the in-tank pump, it is not too old but it is one of those second tier replacement products that work OK in normal applications but may not be up to the OEM or premium part so I put in a VDO. And I replaced the inline filter between the main pump and the fuel rail. I noticed the main pump I am using is an earlier 60mm internal Bosch pump - number ending 012 and with the specified VW part number. I dont have any specs for this one but I do have them for the later 033 which I also have as a spare, and on the chance the 012 isnt up to delivering the 1200 cc/min above 3 bar I appear to be needing, I swapped in the 033 too. VW used the 012, 013 then the 033 as the same part over the years and Bosch seems to have dropped those others some time ago and I cannot find specs on them, unlike the 033 which should be more than adequate for the fuel I am consuming at peak power.

I must have tweaked the plastic VW pump housing when removing the banjo bolts from the old filter. Unlike the new filter, the old one has no features to help me keep it from turning as I tried to break free the bolts so it put some pressure on the mounting which is part of the main pump plastic enclosure. The housing began leaking at a seam next to the filter bracket on the housing as soon as I lit off the pumps to purge the system. This particular housing has been NLA from VW since maybe 2009 and ws never produced by aftermarket parts suppliers as far as I know. So I spent today checking (for leaks) and then removing one from my parts car, stripping all the rusty bits off and cleaning the gunk out of the inside. I am sick of being up to my wrists in and smelling gasoline so I wont pull the one off the car and swap parts between them until maybe tomorrow. Update: I assembled the filter, pump and accumulator on the salvaged pump housing and installed it...no leaks. I will not have an opportunity to do full power to the redline pulls until the first hillclimb in May so wont be able to log and monitor the fuel pressure to see if the pumps and filter improved it or not until then.
A2 VW GTI 9A - MS3/3X
audios
Helpful MS/Extra'er
Posts: 95
Joined: Sun Aug 08, 2004 9:19 am
Location: Alexandria NH
Contact:

Re: VW 9A - 2.0 16V MS3/3X

Post by audios »

Looks great Walter, glad Marc was able to get you some HP's..... VERY impressive #'s from a NA 16v!!
ryster
MS/Extra Newbie
Posts: 1
Joined: Mon Mar 31, 2014 4:48 pm

Re: VW 9A - 2.0 16V MS3/3X

Post by ryster »

looks great!

Would you mind taking a moment and elaborate on your choice of MS input sensors?

What did you use for your CLT and IAT?

Where and how did you mount them?

I assume the other sensors you listed are for gauges only?

If you didn't use the two suggested/generic GM sensors (DIYAT), what sort of calibration process did you follow?

thanks!
Post Reply