When you click on links to various merchants on this site and make a purchase, this can result in this site earning a commission. Affiliate programs and affiliations include, but are not limited to, the eBay Partner Network.
No real changes for the car over the last year performance wise. I do plan on installing the
new turbo shortly.
I did paint the car during the summer. (in the garage)
I removed the trim and easily accessible panels.
This broke the painting out into much smaller jobs, which also allowed me to get better at painting as the job progressed.
Now I wished that I had removed the soft top, but anyway..
After I painted, I sanded down to the belt line, and then buffed it out. I should have buffed it with a more aggressive compound, but I was
concerned about burning through. As it stands, I need to buff it again to get it where I want it.
Perhaps overkill, but I did build a largish chamber to pass the compressed air through. I loaded this with
desiccant to ensure that there was no water in the compressed air. This worked really well.
What did not work well, was the plastic I used for masking. The paint did not stick well to it, and some of the primer
came off when I was spraying the top coat. That made a mess of things and was one of the things that made this take
far longer than I was planning. At least it is now done.
When I changed from a 4.1 to 3.63 diff, I knew I was going to take a hit on acceleration at autocross.
Typically in Canada, the courses are rather small, and the speeds much lower than in the US.
Regardless, I tend to not worry about shifting, and just leave it in 2nd gear.
With this in mind, I was thinking I would see about a 12% loss in acceleration due to gearing.
Actually, it is much worse than that.
Suppose I normally come out of a corner with my 4.1 differential at 3000rpm, from the below virtual dyno I would have 250ft/lb of power available. (blue dot)
Now if I come out of that same corner, without shifting with the 3.63, I am at 2650rpm, which means I have 220 ft/lb less the 12% equals ~ 200ft/lb equivalent. (green dot)
That is a huge reduction. There looks to be more shifting in my future.
The line in blue on the above graph, shows the amount of power I would need to negate the effect of the new diff.
I do plan on improving the bottom end power, but I doubt I can make up for that much of a difference.
The challenge is on.
Great work on the paint Oreo looks really fresh and clean.
Final drive is an interesting one. I guess the payoff of the longer ratio is if you found yourself on the limiter in 2nd without the time to change into 3rd, then you gain an increase speed in those areas.
But I would hazard a guess that you spend more of your time accelerating out of the corners on Autosolo where the torque disadvantage is in action.
Could you not go back to the 4.1 but rev it higher?
I use a 3.6 atm, but I tend to do trackdays - higher speeds / bigger tracks / longer corners that I don't want to change gears mid exit.
^ Thanks Tchaps.
I don't plan on changing back to a 4.1, even with the performance hit. There are other advantages to the 3.63.
I got a surprise when I looked closely at my Garrett T04E turbo today
No contact with the compressor housing, but the bearing is toast. Some of the internal parts got to a high enough temp
to blue the steel. Having a bitch of a time getting the exhaust housing off, so I can get this thing completely apart too.
So I did mount and do some testing on the new turbo. The old turbo was a T04E-50 trim (size wise like a GT3276). This is a sleeve bearing turbo with a .63A/R T3.
I also got and installed a new Turbosmart external waste gate. My old Turbosmart wastegate worked fine, but I couldn't get any springs, which meant I was using a 6lb spring to control 28psi boost pressure.
Not ideal, but I managed to get it to work.
So initial thoughts on the GTX2863G2 .63/T3.
First surprisingly I did not need to change any of the in boost VE settings in my MS2. Also the settings for the boost control work as is.
This is based on limited testing, but the logs I have got, all look fine. Surprisingly, I do need to pull some fuel at idle, as I am running about 15% rich. I find that a little strange.
In turbo only mode, I am only getting about 9psi at 3500rpm, compared with 6psi with the previous turbo. Response is much quicker, with full boost coming on over 1000rpm eariler in lower gears.
Comparing this to T25/.63AR setups, spool appears to be significantly delayed with the T3/.63AR setup. I am considering getting a .48 but need to do some other tests first.
Once you put the supercharger on, these small differences equate to even smaller improvements. I need to reconsider my options as this is not doing it for me.
So I got lazy and ended up buying a vest from Compcooler. The shirt I had was pretty tight on me and was difficult to put on and off. Building a new shirt/vest was going to be a chore, so I hit the easy button.
It shipped for free from China and was attractively priced at $149..
Model COMP-MLCV-801 in black. Cool Vest
Build quality looks good, it's comfortable and seems to have the same number of tubes as competing units. I have not tested it yet.
I modified my heatsink by adding another surface to attach more peltier devices to. This gives me the physical space to mount 24 devices.
I machined slots in aluminum plate to mate to the heatsink fins, and then assembled them with heatsink compound and a bit of glue.
The fins were a bit of a friction fit into the plate and there seems to be a good interference fit.
The heatsink was surplus so had some threaded holes, which were going to be under the Peltier devices. I made some threaded aluminum rod
and filled these holes. (then machined the surface flat)
I machined channels into 2 plates (1 for each bank of peltiers), for the water, and glued some fittings into them. I then machined the back sides of these plates, with areas to mount the
Peltiers, with slots for insulating foam around in between the banks.
I then mounted the peltiers to this plate using thermal paste and a bit of glue on the wires. I used a scale to ensure a consistant amount of thermal paste was applied to each device.
This is what it looks like with 1/2 the peltiers mounted.
In typical fashion, this has ended up being far more work than I envisioned. It better damn well work!
Anyway, I am going to test it with just the 12 Peltiers, to confirm what I have will work.
So in early February, I installed the new turbo (a 2867) and wastegate. It performed like crap. It spooled only slightly better than the T04E-50 trim.
After checking for an exhaust restriction and any obvious boost leaks, I decided to do some overdue maintenance.
I found 2 things that were probably impacting my spool. I had some intake valves with little to no clearance, and I had a leaking BOV.
After fixing things, I am much happier. In turbo only mode, 200ft/lb is reached at about 35-3600rpm.
In twin charged mode, 200ft/lb comes right at 2000rpm and 250 comes by 2600rpm. Can't get enough traction right now to see when 300 arrives.
In a fifth gear pull, the supercharger now gets bypassed at about 37-3800 rpm
Once I have things completed, I am going to need to get this over to a dyno.
Here is how the boost curve now looks.
I do have some other parts to still try (like another turbo manifold) but at least things are looking reasonable now.
So in early February, I installed the new turbo (a 2867) and wastegate. It performed like crap. It spooled only slightly better than the T04E-50 trim.
After checking for an exhaust restriction and any obvious boost leaks, I decided to do some overdue maintenance.
I found 2 things that were probably impacting my spool. I had some intake valves with little to no clearance, and I had a leaking BOV.
After fixing things, I am much happier. In turbo only mode, 200ft/lb is reached at about 35-3600rpm.
In twin charged mode, 200ft/lb comes right at 2000rpm and 250 comes by 2600rpm. Can't get enough traction right now to see when 300 arrives.
In a fifth gear pull, the supercharger now gets bypassed at about 37-3800 rpm
Once I have things completed, I am going to need to get this over to a dyno.
Here is how the boost curve now looks.
I do have some other parts to still try (like another turbo manifold) but at least things are looking reasonable now.
I thought of this the other day.. and it is really tricky to measure efficiency/power based on what boost you get where.. .. boost is a good way of measuring flow resistance, but actually not power.. because if the turbo exhaust housing is flowing less.. it will result in a higher boost, but lower power output, since the amount of air going through the "air pump"/engine will be less.. (therefor less fuel at the right afr and thus less energy/power developed).. The idea you had to take it to a dyno, might be very interesting actually.. just to compare numbers per rpm and boost (unless you also change ign-timing of course)..
I thought of this the other day.. and it is really tricky to measure efficiency/power based on what boost you get where.. .. boost is a good way of measuring flow resistance, but actually not power.. because if the turbo exhaust housing is flowing less.. it will result in a higher boost, but lower power output, since the amount of air going through the "air pump"/engine will be less.. (therefor less fuel at the right afr and thus less energy/power developed).. The idea you had to take it to a dyno, might be very interesting actually.. just to compare numbers per rpm and boost (unless you also change ign-timing of course)..
You are correct here, boost pressure does not necessarily reflect wheel HP.
I am using the data logs generated by Tuner Studio and importing them into Virtual dyno to generate torque and HP curves.
So I am pretty confident in the gains/numbers I mentioned above.
The problem with this, is that to get good traction, you need to use a high gear, which means you will be going fast
if you are calculating high rpm power.
In Ontario, if you go more than 40km/hr above the limit on a minor road (80km limit), or more than 50km/hr on a major highway (100km limit), they
classify it as "Street Racing".
So you don't want to get caught speeding around here. Hence the desire for using a dyno.
On another note, it seems that I need to use fairly high boost levels to make significant power.
In part that is probably due to not having E85 and the supercharger being a parasitic drag at high RPM's.
It still seems a little out of line to me, but I can't put my finger on any further reasons for it.
You are correct here, boost pressure does not necessarily reflect wheel HP.
I am using the data logs generated by Tuner Studio and importing them into Virtual dyno to generate torque and HP curves.
So I am pretty confident in the gains/numbers I mentioned above.
The problem with this, is that to get good traction, you need to use a high gear, which means you will be going fast
if you are calculating high rpm power.
In Ontario, if you go more than 40km/hr above the limit on a minor road (80km limit), or more than 50km/hr on a major highway (100km limit), they
classify it as "Street Racing".
So you don't want to get caught speeding around here. Hence the desire for using a dyno.
On another note, it seems that I need to use fairly high boost levels to make significant power.
In part that is probably due to not having E85 and the supercharger being a parasitic drag at high RPM's.
It still seems a little out of line to me, but I can't put my finger on any further reasons for it.
Oh wow!.. that is discouraging.. better not loose license or car then!.. We have similar rules here.. as a rule of thumb the limit for loosing your license many months (don't remember exactly) is +30km/h and above.. and depending on how much above they might send you to court..
Virtual dyno seems good if the conditions are right .. alot of people use it.. i have trust issues though.. (in general.. haha!).. so i would feel much better about measuring in a more controlled manner. Still, i also need to keep an eye on not getting to high torque in my setup.. you don't seem to have that problem!
When it comes to SC parasitic.. it should be much less when it no longer generates boost ? .. i think we discussed that before at some point ?.. although it surely still is some loss.
How does the IAT look at those higher boost levels?.. (are you also measuring heat/pressure in exhaust manifold?)
It's really nice to reread your thread.. it's such an impressive build you have!
Thanks, wow 30km over and loss of license. I thought our laws were ridiculous.
Regarding SC parasitic losses. I could dyno the car with the supercharger belt connected but with the SC output pipe disconnected, and then repeat with the belt
disconnected. The difference in the power for these 2 dyno runs would be roughly how much power it is using. I would guess a 15ft/lb loss at high rpm.
The supercharger is still generating about 1 psi of boost due to restrictions in the piping but that should not make much of a difference.
The temperatures after the intercooler are quite similar to a turbocharged car, however air going into the turbo is 150-200F at lower RPM's.
If I do a track event, I remove the supercharger belt, otherwise the intercooler gets heat soaked pretty quickly. IIRC, at one point I was seeing intake temps up
near 200F when trying to run twin charged at the track. This is one of the reasons I wanted to get a quicker spooling, smaller turbo.
I have not measured the temperature of air exiting the supercharger when the turbo bypass is open (high RPM). It should be cooler.
I am not measuring anything on the exhaust side of things. Ignorance is bliss.
The weather is poor here right now, but hopefully next week I will be testing an electric diverter valve.
The problem is that the BOV I am using does not react to the the smaller boost spikes caused by getting off the throttle at low rpm, when in boost. At low RPM it takes
a while for manifold vacuum to build after the throttle plates close. This delays the BOV and in some cases where the boost spike is only say 6psi the BOV will not even open.
Just blipping the throttle and having the RPM get up to ~2500rpm, can result in 15psi spikes. These boost spikes cause the supercharger to make weird noises and probably is not good for the supercharger either.
The plan is to have an Arduino monitoring boost and throttle position, opening the valve when the throttle is closed, and boost is greater than a couple of psi.
Thanks, wow 30km over and loss of license. I thought our laws were ridiculous.
Regarding SC parasitic losses. I could dyno the car with the supercharger belt connected but with the SC output pipe disconnected, and then repeat with the belt
disconnected. The difference in the power for these 2 dyno runs would be roughly how much power it is using. I would guess a 15ft/lb loss at high rpm.
The supercharger is still generating about 1 psi of boost due to restrictions in the piping but that should not make much of a difference.
The temperatures after the intercooler are quite similar to a turbocharged car, however air going into the turbo is 150-200F at lower RPM's.
If I do a track event, I remove the supercharger belt, otherwise the intercooler gets heat soaked pretty quickly. IIRC, at one point I was seeing intake temps up
near 200F when trying to run twin charged at the track. This is one of the reasons I wanted to get a quicker spooling, smaller turbo.
I have not measured the temperature of air exiting the supercharger when the turbo bypass is open (high RPM). It should be cooler.
I am not measuring anything on the exhaust side of things. Ignorance is bliss.
The weather is poor here right now, but hopefully next week I will be testing an electric diverter valve.
The problem is that the BOV I am using does not react to the the smaller boost spikes caused by getting off the throttle at low rpm, when in boost. At low RPM it takes
a while for manifold vacuum to build after the throttle plates close. This delays the BOV and in some cases where the boost spike is only say 6psi the BOV will not even open.
Just blipping the throttle and having the RPM get up to ~2500rpm, can result in 15psi spikes. These boost spikes cause the supercharger to make weird noises and probably is not good for the supercharger either.
The plan is to have an Arduino monitoring boost and throttle position, opening the valve when the throttle is closed, and boost is greater than a couple of psi.
I think that is reasonable temps still.. since you are not cooling the iar between the sc and turbo..
I think that bov-issue might come from two things.. combination of throttle position (before supercharger?), and bov then also far from where throttle body normally would be placed.. (like on a turbocharged car)..
other thing would be to put a really light spring in the bov.. (as long as you feed it with manifold pressure).. i think i did that in my setup in the beginning also.. because the spring gets alot of support from the boost as long as the throttlebody is open.. and the spring should open on vacuum (it does for me).. what are your thoughts that?
^yes, the main culprit is the electronic throttle body at the inlet of the supercharger.
It takes about 75-85mS to go from fully open to fully closed. Also, the electronics adds another ~60mS, which I could probably
get down to 5mS if I replaced the MS1 with an Arduino. Since I am putting in an Arduino anyway to control the
diverter valve, I will see if this can easily be done. (I am thinking it can)
My BOV is set up so that if atmospheric pressure is on the valve side, and normal engine vacuum is on the diaphragm side,
the valve will be closed. If the engine is decelerating, pulling additional vacuum, then it may actually open the valve, IDK.
I think it would be preferential to keep this valve closed, unless there was definitely positive pressure on the valve side. Unfortunately
my BOV does not have any adjustments on it to do that. I could always try shimming the spring a bit I suppose, but this is not high on the
priority list.
^yes, the main culprit is the electronic throttle body at the inlet of the supercharger.
It takes about 75-85mS to go from fully open to fully closed. Also, the electronics adds another ~60mS, which I could probably
get down to 5mS if I replaced the MS1 with an Arduino. Since I am putting in an Arduino anyway to control the
diverter valve, I will see if this can easily be done. (I am thinking it can)
My BOV is set up so that if atmospheric pressure is on the valve side, and normal engine vacuum is on the diaphragm side,
the valve will be closed. If the engine is decelerating, pulling additional vacuum, then it may actually open the valve, IDK.
I think it would be preferential to keep this valve closed, unless there was definitely positive pressure on the valve side. Unfortunately
my BOV does not have any adjustments on it to do that. I could always try shimming the spring a bit I suppose, but this is not high on the
priority list.
Ahh.. i see.. how about having a mechanical tb.. how would that change things?..
I don't remember if i changed spring in my bov or if i cut the spring.. i do remember checking specs on bov for sc vs. bov for turbo and the spring rate was different according to some of the manufacturers.. as in much lighter on the sc-purposed bov..
well it will be interesting to see how this develops and how you solve it !
^ While a 2 mechanical throttle bodies will generally work, I am unsure if I could get it to work in my set up.
Perhaps if I put a smaller throttle body on the inlet of the supercharger, or perhaps reduced the drive ratio of the supercharger.
Below in Blue, is a graph of my stock throttle body opening percentage against the programmed 80mm ETB throttle opening. See at 50%
stock throttle opening, I am at slightly less that 20% opening on the 80mm throttle. When playing with this, I ran the curve
in red and found that with these settings, the supercharger was generating on average 2psi (with some peaks up to 3.5psi) when
running at part throttle with vacuum in the intake manifold.
That really reduces the drivability. So for the current setup, the DBW throttle is the best solution.
I did do a few things.
1. I mounted a new fuse box and relays in the trunk, and ran some thicker wires to the fuel pump and MS1 computer and ETB's.
2. I mounted a diverter valve, directly below the supercharger connected to a small Arduino board.
3. I mounted the Arduino board in the box with the MS2 and MS1.
Results:
At 3000rpm the throttle response from 50kpa to 10psi is .2 seconds measured with megasquirt, so it's probably a bit slower than that.
At 2000rpm the throttle response from 40kpa to 10psi is .467 seconds measured in the same manner.
The 3000rpm is much better, in part due to the better spooling turbocharger.
The 2000rpm spool is only about .08 seconds faster. Spool up to full boost at 2000rpm is pretty slow.
I moved the fuel pump relay and added some relays to a small metal plate, and moved the wires away from the base of the trunk. Much cleaner with 10 more fused outlets.
Here is the Arduino board (TLH corner), which helps improve throttle response and is supposed to activate the bypass valve. I really need to do a new PCB for this stuff once I lock it down.
After doing a bunch of testing on the bench, the diverter valve still does not operate the way it is supposed to. I hate software!!!!
Here are what the diverter valves look like. The one on the left is a newer design, but is not fully sealed, so would not work here.
With such a small valve diameter of .65", I wonder how much of a difference this will actually make. I guess we will see.
The aluminum bit was supposed to make mounting easy, however it turns out that the valve on the right (from a BMW) is slightly different than
VW one on the left. oh well.
Love the bench testing rig Oreo, so much better to realise you've still got an issue there than in the car, I'm sure you'll get there.
Apologies if this has already been covered - I did have a scan back - what are these diverter valves? Are they the components that transition the car between supercharger and turbo? Are they a PWM or H bridge control?
- what are these diverter valves? Are they the components that transition the car between supercharger and turbo? Are they a PWM or H bridge control?
They are basically a electrically operated BOV. I have a standard BOV, but at lower rpm/boost levels where the supercharger is doing the heavy lifting, it won't activate or it activates late.
You basically just apply 12v to get the diverter valve to activate. Right now the Arduino is just monitoring boost at the outlet of the supercharger, and if it is above 2psi and the throttle is less than 5% open
then it triggers the diverter valve for 250mS.
Did you get the cold plate design installed? How much voltage/amps did the 12 peltier draw? Awesome build with lots of cool engineering. I remember back in the early 80's Ferrari F1 was running the super charger/turbo system when I was at the Long Beach GP. I think FaI outlawed it soon after.