Topic: Twincharged Solstice = Turbocharger + Supercharger

[Steelmesh]

Wearing glasses definitely helps me weld. But machine can also make a difference for sure.

my old one failed a year or two ago, and couldn't be repaired. They supplied me with aused, but newer model of the same, and it just works better all round.
And I'm not a great welder by any means, so even that was a surprise. I even wonder if mine had a fault since new, it was around 12 years old or so when it died, and I could never really use the foot pedal at full power or it'd trip the thermal overload/duty thingy.

New one, not a problem, it'll take full pedal.

Looks like it's a welding machine problem and not the welder.  Thinking how I want to take my  out on the old machine, for now I'm busy with the car

OLD WELDER, prepped to all heck:


NEW WELDER, No prep worst case!!!:

[Steelmesh]

New welder Lincoln name and competitive with the 200A low name import brands with AC TIG, hopefully this one can be serviced unlike the disposable harbor fright unit.  I don't regret the HF unit, it got me seat time versus excusing myself from welding because I have to have a $6k dynasty otherwise it's not worth it to me I'm too good for anything less, I refuse to weld reeeeee. I'm sure some of you have had a friend like that, oh yes, Snob is the word.




New Fuel System, 8AN feed and 6AN return all PTFE.  I have to eat my earlier words, just couldn't do $600-700 in name brand hose and ends...so I went with all Evil next day delivered ~$400.  Random picture I had, that's the old filter with hose tails. PTFE much smaller diameter, quite nice re-routing everything.




This was me a couple weeks ago (it's all a blur since July 4th weekend): 

Why the welder going down hurt really bad, like the sky is falling.

This is after I was filling up the Coolant system, I got a new vacuum filler to see how that fairs with the Solstice. I noticed all the silicone coolant hoses collapsing.

Do you see where I am going with this?!?!?  I messed up big big big time on the Bypass Design. Notice anything different:

[Steelmesh]

I think I'm put about 70-80 hrs into the car this year. 

On top of the operations problems I've been battling, old cheap junk tools I bought 8 years ago expiring....at the heart I think it's stuff like this that adds up and things I forgot adding many hours to the project.

The Bypass system takes priority, need to relocate the dipstick....again.






Borderline Sheit because I used my old welder and it wasn't fun, so it left some boogers inside, I think I found an angle it will not get jammed up at.

[Steelmesh]

Thanks to whoever suggested the H-bridge earlier for throttle control.  The broken straw in the camel stack was the lack of TPS feedback and closed loop control overall.

And now I understand why people don't share any details how to make this work with a microcontroller, Electronic Throttle Control.  I'd estimate 12 hours of engineering work into the project start to finish.  Arduino's aren't high performance, I tried to get my STM32 working and bricked it.  I think an STM32 would perform significantly better, however suffers from cost and inconvenience dealing with 3.3v logic. 

Finally it's behaving consistently, not freaking out and all:
https://www.youtube.com/shorts/Hr8yiz8wbCQ

Went to bed laid down, however for some reason looked at the H-bridge datasheet on the phone, noticed I didn't wire and command the reverse direction properly (so it was using the spring to return then), so I got out of bed and added that to the circuit and coded it in, I think it got smoother.

Durability cycle, see the thermals and try to glitch it out: https://www.youtube.com/shorts/fRQDEhdDbOA


Philosophical thoughts: There are people with F-U money, and people with F-U skills.  People sometimes feel entitled to other peoples money because they have a lot.  Don't you think people sometimes also feel entitled to other peoples skills because they have a lot of skill?  For example, a 20 year experience locksmith opens your door in 3.2 seconds and charges $100.  You know people get upset for paying $100 for 3.2 seconds of work.  The point is figuring out my response to people asking for the code and wiring diagram.

[Sol Asylum]

It's the engineering triangle: fast, high quality or low cost pick two because you can't have all three!

[stevieturbo]

That's a nice looking bead roller.

The pipes look very close to the brake reservoir, no risk of contact with engine movement ? Engines can move a lot.

Good, and bad to hear about the welder. Good the new one works well....bad the old one was the reason behind really shit welds.

[Steelmesh]

That's a nice looking bead roller.

The pipes look very close to the brake reservoir, no risk of contact with engine movement ? Engines can move a lot.

Good, and bad to hear about the welder. Good the new one works well....bad the old one was the reason behind really shit welds.

Thanks for pointing that again.  It is about 6-7mm clearance, I'm running poly shackle bushing engine mounts and a stiffer trans mount. What kept me on this path is considering the engine is spinning clockwise viewed from the front, therefore when it reacts, it will react in the benefit of the reservoir anti-clockwise.  I hear engine braking is pretty bad for the Ecotec timing chains, maybe as long as I don't load the engine in reverse it will be fine.  Will certainly use a telltale like a thin layer of grease and see if it transfers.

[Steelmesh]

It's the engineering triangle: fast, high quality or low cost pick two because you can't have all three!

That's right!  Like here, in the early stages of programming, the trade off the time vs benefit can be higher, I just did 2.5 hours some how and this magic popped out, so much better control algorithm it's worth it, now I got to translate this into a project box and wiring harness. 

Running the durability glitch cycle on it again with the new software update:
https://www.youtube.com/shorts/dD9-Xn3dNaw

[Steelmesh]

Throttle box version 1.  My bench testing involved using an Arduino 0-5V PWM signal.  The ECUmaster Black has a PWM output as well, however it sinks to ground.  Therefore used a 1k ohm resistor to a 5V source to pull-up this signal line, then the ECU can pull the 5V to ground to make the 5V PWM signal.  The logic is flipped, so in the ECU 100% duty will be throttle closed and less than that will be opening, it works for now!



I saw the notch coming from the beginning, chop for science:



The car alternator hasn't charged the battery since the new ecu.  I tried the ol' 5V pullup trick above so the ECU can control it, and not sure what happened I may have wired it wrong initially didn't work, this was when I installed the ECU.  I swapped out the alternator when I rebuilt the car this year so maybe it was an alternator issue previously.

For GM, you can just buy a PWM generator.  Since I'm setup to develop microcontroller solutions from the throttle box, I put in a few hours and built a closed loop PWM alternator controller, 3k/1k voltage divider so it reads car voltage directly and then adjusts PWM accordingly, my guess was right if you want more voltage then add more PWM duty %.  Great to see 14.2V on the line finally.

I think this is a $5 rebranded PWM generator, they're charging $90: https://brand-x-electrical.myshopify.com/products/voltage-control-knob-pwm-control-based-on-internally-regulated-alternators-pcm-controlled-05-gm-vehicles

I guess since it went together so easy and the IP is in the closed loop control, I can also charge $100 for $10 in cost.  Closed loop control is the benefit there so it maintains a voltage unlike a knob adjusted "open loop" pwm generator.  Heck, get a PCB designed and assembled for me and make a bunch, however this seems like sell 2-3 a year on ebay type of product.

[Steelmesh]

This past weekend I took it for a test drive was feeling good.  I did a small hit, looked in the rear view mirror and saw the thickest cloud of white smoke covering both lands, that's not good.  Immediately pulled it over and looked all around it the engine and didn't notice anything off, fluids are good.  I checked out the log and didn't see anything alarming, a little rich maybe like 11:1 AFR.  I then limped it back home and it was putting out a constant amount of smoke.  It does have a bluish tint and smells like burning synthetic. 

Started doing research, 1) turbo seal, 2) clogged return, 3) excessive blow by. 

Today was then root cause day.  Start from the worst case I broke the internals, so pulled all the plugs and they looked great.  Next steps I thought, check the breather system and after that pull an O2 sensor to see what's up before taking apart the hotside.  I pull the hose off the venturi valve (uses exhaust flow to create a vacuum) and it started draining a lot of oil, okay so this is the source of the oil.  Why is there so much dang oil in the breather hose.  I did redesign the breather system this year.  I welded a second hose barb to the valve cover when I built the first race engine.  The new setup has the OEM nipple with a one-way valve to a filter element, this is for fresh air to enter the crankcase and the added nipple goes do the custom air-oil separator then to the venturi in the exhaust.  Chapter 2 at some point I'm adding an electric air injection pump to actively pull from the crankcase. 

Realized, another thing that could cause this is being over filled with oil.  I had tape on the windshield and steering wheel that said oil from months ago.  I certainly drained some oil because I redid the whole oil cooler loop. 

She had 10 quarts of oil in it guys, I totally messed up here and didn't read the dip stick right, I'm the guy who fills his cars near the top of the dipstick that's my excuse, it sure was to the top of the dipstick. 
 

====================

Another thing I did was rebuild all the brakes put Hawk HP Plus on up front (SSBC BBC, uses 1997+ front corvette pads), cleaned and relubed the slide pins.  I picked up some new OEM brake hoses all around for a good value, figured it's better than than 19 year old hoses and continues the 2025 theme of replacing every single hose on this car.  Flushed some new prestone fluid through it as I did the corners, then did a full bleed with ATE200 brake fluid.

For Hawk pads, the trick seems to be to purchase new old stock from eBay to save money.  I got HP Plus fronts for my Cruze <$100, and then front and rear HP Plus Sols pads for <$100 each. And come to find, it's been so long I forgot I already put HP Plus pads in the rear. 

August 24th is the next SCCA event, I would really like to make it to this one since I missed the last two events.  Lots of road blocks and land slides this year, miscalculations and broken tools, scope & feature creep. Often problems come in 3's, I agree for me it's 33.  Don't know if it's a test, something out there trying to frustrate me to stop, my mindset at this time is still, "come at me bro" and "let's goooo!"

[Steelmesh]

Runs a lot less smokey with the proper oil level.  I got it out last weekend.  First did some drives with the bypass fully open just to verify operation, curious how the big turbo feels alone.  With the bypass activated, the log shows a dip in boost when boosting up, likely when the bypass opens however I'm not logging the bypass position.  I think that is a need to know thing.  It overboosted when I pushed it so I stopped completely and want to figure out how to add another analog input when all the analog inputs of the EMU Black are taken. 

The EMU Black ECU allows analog inputs and switches through CAN. With that, I assembled this last night (ordered parts the day before), it is an analog to can box (these are $600-800 off the shelf or imported).  Since I'm going through the trouble I'll add some extra channels on top of the bypass TPS, including fuel pressure, crankcase pressure, coolant pressure, oil temp as some ideas to fill it up with stuff, moar data!!



Showing a 50Hz sign wave, 70 Hz starts to deteriorate, not very good a 100 Hz; 



========================================================================================


Note about the electronic throttle body and I'm leaving some efficiency on the table:
The LUV throttle body (like most) has a rest position that is slightly open, therefore I need to update my throttle controller to actively force the throttle to actual 0% during SC operation or possibly crack the TB open and manually adjust the rest position to 0% if it has adjuster screws (like some). 

[Steelmesh]

With the big rad mod I did it blocks the CAI from getting up front.  My understanding is some classes and road legality means I have to run with a headlight.  I guess I could find a headlight and chop it up. However, I've had this naca duct since pre-cvidious and finally got around to it.  Will need to fab and airbox to block all the rad heat and take full advantage.  For rain, the plan is to make a cover for it and I've been avoiding rain with this car for a long time now anyway.



not attached yet, mock up:

[Steelmesh]

Still working on the transition from SC to TC.  I'd say now with empirical data that the turbine is a bit too big because it doesn't hit the boost target until about 5,500 rpm on the turbo "only".  That would be great for drag racing and a high redline.  I'm going to try out an 0.82 A/R turbine housing next to bring down the boost threshold.  Having the SC is pretty awesome as expected, playing with and without the SC working.

Bypass open during WOT (so riding the turbo here):


Exciting run here, this is with the bypass activated, when the bypass opens you can see the boost drop off, before that you see it hit overboost and then I pedaled it, trying to collect data here:

[stevieturbo]

Are you just opening it in one go, or progressively ?

What diameter is the blade ? In reality, you wouldn't really need much diameter. Even mine with a standard 54mm blade ( turbo only ), can make plenty of power, and I deliberately kept it small to ensure good throttle resolution.

You might find it needs a quite slow and progressive opening, and maybe not fully open for the best transition.

When mine was compounding, I could easily make 2 bar boost at around 2000rpm. Now turbo only, I've strong boost from around 3500, but it is making boot even from around 2000rpm. It never feels lacking.

[Steelmesh]

Are you just opening it in one go, or progressively ?

What diameter is the blade ? In reality, you wouldn't really need much diameter. Even mine with a standard 54mm blade ( turbo only ), can make plenty of power, and I deliberately kept it small to ensure good throttle resolution.

You might find it needs a quite slow and progressive opening, and maybe not fully open for the best transition.

When mine was compounding, I could easily make 2 bar boost at around 2000rpm. Now turbo only, I've strong boost from around 3500, but it is making boot even from around 2000rpm. It never feels lacking.

Agreed it should be a transition.  Execution of this is under development. See CAN analog #6 in the graph that is the throttle blade position kind of goofy it comes through as 0-5V which would represent 0-100% TPS.  From contacting ECU master last week, they only have hardcoded inputs I can apply CAN analog signals to like coolant pressure, crankcase pressure, fuel pressure, backpressure, and a few more common race engine inputs. I'll take any data I can log, so it is what it is.

There is some throttle control challenge, like at the very opening it unstable, as a hack fix I just force it directly to ~15% when it comes off 0% position. That is not ideal considering flow is not linear with % throttle blade opening, most of the flow increase happens at low throttle positions as you probably know.  The version 1 throttle controller I made with the arduino is not an ideal solution, however it was quick to get going, need to try out a faster processor (STM 32) to get a higher frequency.  I was already setting up an STM32 during the build after proving it out with the arduino nano, the STM32 was giving me trouble (using Arduino IDE) and somehow I bricked it, so jumped back to the nano to build out the throttle box and see what happens and make progress.

It's a brain twister for me wrapping my head around the ECU-map starting out here it's all new and unknown, the ECU has a table setup as RPM vs. Manifold pressure to output a PWM signal (that the throttle box reads).  In my initial bypass testing on the street I was leaning completely on the wastegate for the lower RPM boost control, however I should be cracking open the bypass to help out with the compound boost at low RPMs that can be quite excessive with certain conditions, transients.

Development is on a temporary hold because I'm planning on racing it this coming Sunday there is an autocross event. I did a proper shake down of the car; took it to Woodward Dream Cruise for the annual kappa meetup and cruise, the car made it there and back home without anything concerning, was being nice to it however tried to pretend for a second here and there I'm making autocross turns and accelerating out and it's doing it.  With that, I want to achieve my goal this year of putting the car on the track I'm determined and the season is almost over, I'm the epitome of scope/feature creep and taking the car all apart, I see it a little better now so why I cut myself off from tinkering until then. Also, hard to legally replicate an autocross duty cycle so the best way I can think to really see what's happening is to do it. I'm also kind of exhausted, this has been everyday and full-time weekends since the beginning of July with a few vacation days where I worked full-time on the car.

Probably good to sit back for a minute and rethink about it, take care of other life things I skipped out on, and take in comments like yours to inspire some ideas about the control strategy for this setup.

[stevieturbo]

I would really consider just using a wastegate for the bypass, then use a typical 3 port solenoid to offer some control around it.
If the blower tops at say 15psi, then set the control from the turbo side, so when that hits 15psi it starts to bypass, and it should be fairly progressive.

Although in theory yes, a DBW blade should allow you the best control, seems it's more of an ecu limitation at the minute preventing good fine control ?

I would maybe just try a very small opening scaled to boost from the turbo, maybe only opening to around 20-30% or so ?

[Steelmesh]

I would really consider just using a wastegate for the bypass, then use a typical 3 port solenoid to offer some control around it.
If the blower tops at say 15psi, then set the control from the turbo side, so when that hits 15psi it starts to bypass, and it should be fairly progressive.

Although in theory yes, a DBW blade should allow you the best control, seems it's more of an ecu limitation at the minute preventing good fine control ?

I would maybe just try a very small opening scaled to boost from the turbo, maybe only opening to around 20-30% or so ?

Self-regulating mechanical systems are awesome, it's a good idea to K.I.S.S. the bypass system.  Maybe for development having a little feedback loop to see what is going on is beneficial (wastegates don't normally come with position sensing).

I scratch built a throttle blade controller using a microcontroller that controls an H-bridge.  I have experience developing a gyro stabilized light bar, had the same problem of instability and the solution was to go to a faster processor to clean it up, so the plan is to look at that before scrapping the idea.   

The ECU is not putting out any current / not driving the throttle blade even though it could; for now it's a PWM output ground side switched signal.  I have a 5v DCDC through a 10k resistor that pulls up the signal line, and the ECU PWM pulls that to 0v when it pulses, creating the 5V PWM signal input as one hack solution to talk to the microcontroller and command a throttle blade position.

Currently the boost controller is looking at intake manifold pressure, so for example if that is set to 25 psi, the wastegate has no idea what the turbo outlet pressure is it's only looking at manifold.  Let's say the PR is 1.5 across the SC, therefore compounding only no bypass, turbo out is 10 psi and SC out is 25 psi, being controlled by the wastegate that is holding the turbo at 10 psi.   The next step I'm thinking is to match the Bypass Throttle cracking at around 25 psi manifold pressure, so when it cracks and manifold pressure drops, then the wastegate should take up the slack and close a little to maintain 25 psi.

[Steelmesh]

9/21/2025:

*Norms front bumper cover removed because of cones and my tow dolly wants to rip it off the car when I back off.

Those are the new wheels and tires (sponsored by the person who rear ended me last year).  FAST wheels again I went to Canada to pick them up, 18x9.5 because they're lighter and mostly because 18" performance tires much more affordable than 19" tires.  265/35R18 Kumho V730

It's all a blur, I better make a post before I forget what happened. I did make it to the autocross last month, which exposed the lack of optimization in the tune.  The SC was being bypassed too early, hysteresis from the bypass opening that lowers MAP and a lowered MAP re-closes the bypass and so it oscillates, finally the ignition timing was way too aggressively retarded.  I ran the car on stock springs, stock bars however with konis and also ran on 275 wide old tires, only did spun around once and 1 sideways slide into a cone.  In the end, I made it to the track, I made all the runs, powertrain didn't fail, my 2025 goal achieved!!! I am human, that is not enough, now for redemption.

That afternoon, I unloaded the car at home and started looking at it, added timing and really woke up the supercharger, acceleration G forces are back.  Then that week I redesigned the throttle controller, got an STM32 up and running.  Rolled back the code and created a new control strategy.  Before I ripped apart the old throttle box and built a new one, I flashed the old throttle controller with the new code and with that it works acceptably good across the range of opening from 0% and more.

Following on with this is implementing a hysteresis management strategy, the ECU command signal output is limited to MAP vs. RPM table which is the foundation in my opinion for the bypass strategy.  My bright idea is to do something complex again, which is build a CAN bus Rx reader (the analog input box I built is a CAN Tx transmitter).   Had another CAN controller module and plenty of arduino's, so I just made a dedicated CAN Rx box this time.  It's reading CAN from the ECU including RPM, MAP, TPS so it can better know the demand.  Out of the Rx box is a digital on/off out signal to the throttle controller, this signal will force the bypass valve to stay open.   Imagine something like [ if RPM>4000 and MAP>250 and TPS > 90 then command open ], then if any of those drop below those numbers [minus deadband], it commands off.  I'm not seeing or feeling oscillations anymore.

The car really needed an alignment (I was trying to get one last year, then got rear ended, messing up the car, taking out fuel pumps, etc).  The major challenge this year however was the Energy Suspension bushings that I installed pre-covid.   They were so stiff, after lifting the car and setting it down it didn't compress, I had to drive it around and it would settle.  The job involves removing each control arm and pressing out the spacer and re-lubing.  This was a massive job, however effective and what needed to be done.  Since I was already in there, I installed the ZZP coilovers (dang ZZP black friday sales, the price was right).   I'm gaining XP in proper suspension setup, so with that this was also a major job. Due to time constraints, I did not take on assembling the hub stands, as I'm trying to make it to the next autox race on 9/21/25.   Suspension strategy to start is to bias towards understeer, which I interpreted as: thicker front bar, lower rear ride height, rear camber, rear tow in, front conservative camber.

Suspension work, on the ground, more tempering of my will to get it done:


The dreaded lower control spacers, the aluminum washer locators things here that I made years ago are holding up great, however lost another metal one, so broke out the lathe again and spun one out:




The ZZP install when pretty good, one thing that is odd are the 10k rear springs seem to be a little too light or they're on the edge where my rear mount battery and the 1 gallon surge tank is enough to over compress the spring.  What I mean is, with the rear springs adjusted all the way off the threads pretty much (maximum preload), the suspension still compresses too much (75-80% of the total travel is compressed down).  The front 12k with all the added weight is in the range of 5-10mm of preload.   I also got ZZP bump stops and cut those in half, they're a little more progressive than than the tiny ones that came on the coilovers. 

Step 1, pull the coilover springs to find the minimum ride height to confirm I'm not going to blow a tire through the wheel well.  I still raised the car up from this spot, it was a baseline move recorded for future reference if I want to drop the car.  For now, I raised the car because I don't yet understand the suspension geometry and how lowering the Kappa impacts roll center.

The tire's shouldn't rub at full compression with a little buffer added for additional bump stop compression under shock load:




Almost forgot about the JY sourced thickboy front swaybar, they still sell ES bushing kits with bracket:


Thanks to my friend for letting me take over the rack on a busy Saturday:


Rear with a touch of toe in:


We made a custom profile for the car, red color means nothing as I just made up some numbers to get it going, and Caster was all messed up and this was the compromise as the shop was getting busy

[stevieturbo]

Would a control strategy that uses turbo pressure vs rpm maybe be simpler ? as opposed to intake pressure ?

That way unless there is good positive pressure available from the turbo, the blower is always the main ?

It would be complicated to get everything smooth and seamless all the time.

[Steelmesh]

Would a control strategy that uses turbo pressure vs rpm maybe be simpler ? as opposed to intake pressure ?

That way unless there is good positive pressure available from the turbo, the blower is always the main ?

It would be complicated to get everything smooth and seamless all the time.

Thanks for looking at this. Yes this is getting complicated, the optimization. I think that is my take on twinchargers now, they're easy to slap together and get running, however complex to optimize performance and drivability.

With the no-bypass series setup, the turbo wastegate was also referencing intake manifold pressure.  For example series system, the boost controller is set to 26 psi and then if the intake manifold is 26 psi, the wastegate would be limiting the turbo compressor outlet pressure to 10 psi (with an arbitrary P/R of 1.6 through the SC).   Same tactic with the bypass system, the wastegate is still referencing intake manifold and would act similar to the series setup until the bypass opens, then it should act like a turbo only system where turbo outlet pressure will be ~1:1 with manifold pressure.

With all that built up in my mind, I think using turbo-out pressure vs. rpm would require an additional function to prevent overboost, covering most bypass-valve states and operating conditions;  why I'm walking the IM path, where the boost controller referencing the IM would be able to crack the wastegate in some weird operating condition if IM pressure gets too high.  Maybe a fail safe if the bypass valve fails, because the bypass valve I'm using is a fail-danger (when it fail, it goes 100% supercharger mode).

I still appreciate the value of a mechanical system valve you suggested to control the bypass, however what I'm observing with the electronic controller is the bypass is open at 2 pressures, it opens at <60 kpa (idle, coast, engine braking) and opens at >250 kpa (~SC boost limit).  Maybe need two mechanical valves, small one that opens with vacuum for idle, and a larger one that opens with boost.

[stevieturbo]

The simplicity of my first setup where I just blew through the blower is hard to beat in that regard. No real issues there at all. The only downside as such was it created lots of boost everywhere. And boost per power...just wasnt where I wanted it.

It would easily make 20+ psi even at 2000rpm, response, torque etc was superb. But even at the top end making say 25-28psi or so....whilst it went ok, it was not as fast as it should be for that boost.

But it was simple and smooth everywhere. But ultimately the blower was definitely a restriction, and the turbo never really shone

Now with just the 7163, for normal driving it is already making usable boost from 2000rpm, but is on strong at 3500rpm. Which for me really is superb. And it's just night and day faster than with the twin setup

But I am glad I tried both, albeit in the very simply format I did.

Obviously you have the benefit of bypassing the blower, but with more complexity to get it all right.