My 4L80e suffers from the same thing. I can command it to shift at 3800 and it still bounces off the limiter before going into second. It's annoying as hell and is directly related to power like GMCtrk stated with his. At 5psi, it shifts normal but at 12psi, it just doesn't want to shift into second at all. 1-2 shift doesn't happen until I let off the throttle and get back into it. 2-3 shift is perfect every time. A local guy rebuilt my 4L80e but it was back when I wasn't making any power so I'm sure it was more of a stock rebuild.

OP, stupid question I know but I'm new to the turbo world. How do you plot RPM and VE % on the compressor map like that? How do you calculate compressor flow?

 

It's not a trans issue, it's a PCM issue. I've logged the shifts and the trans shifts immediately when the pcm commands. I know Parish had this issue years ago.

 

Well wtf, yo.

 

Im not the OP but Ill shed some light.

Volumetric flow of the engine is easy



Trouble is that VE is a tricky number. OP assumed 85% well...VE is a strong function of RPM and is different for every cam and engine size, and is also a strong function of boost pressure. The higher the boost the better the VE. Engine VE will peak at peak torque, always, this is cam-dependent. Induction VE will peak at peak boost, always. Engine VE decreases on either side of peak torque.

So really drawing the engine as a straight line across the compressor map is very rough, it should really be kind of wavy depending on where the induction and engine VE combine to form the overall engine VE.

Compressor flow is given by the map itself.

Also notice the speed lines on the map...for a compressor to maintain constant boost through the whole range is HAS TO speed up. If you are dropping boost, you are either choking the turbo or have a wastegate issue.

 

Did you change both shift rpm and speed or just one of them? Just the WOT tables or the part throttle tables that match the WOT tables too?

 

3456? Yeah I know you made that **** up, Richard.

 

Using compressor maps how I've plotted is just a loose guess. It's close, but atomic is right about ve changing throughout the power band. The idea is thst you want all plotted points on the map. Too far right or left indicates that the compressor is too big or small. I like to put redline in the far right of the map. You reside there for very little time and considering most ls ' can't hold peak torque that high, it's pretty safe. The points on the map are air requirements that the engine needs at a given rpm, ve and psi.

 

12^3 (to convert cubic inch to cubic foot) * 2 (4 stroke engine) = 3456

 

turbobygarrett.com has a bunch of tech info on compressor maps. If you're interested I'd considering taking a peek.

Here's a link to how to plot an engine's required airflow on a compressor map.

Plotting Points on Maps | Turbobygarrett

The other thing to take into account is that compressor maps aren't giving you a whole picture. The maps themselves are based on specific conditions, therefore not being accurate for all configurations. Again, these are used as a baseline for picking the right size.

You'll also notice mfg's that don't have them. This is because of a couple reasons. First being they are too small and don't have the r&d group to create the maps. The second reason has more to do with current technology. Before advance CNC technologies could be purchased for a reasonable price, mfg's had to put a considerable amount of research in both the compressor and turbine wheels to avoid having defects after a production run. You'll notice that there are a lot more maps for older compressors. The idea is that they wanted to get it right before they ran 100k compressors in a production run.

In today's mfg, the access to cheap CNC allows mfg's to tweak profiles and spit out a new compressor in a matter of hours. that's why everything is billet. Keep in mind that billet is not a better because it's material and machining, but better because they will tweak the original profile to get the best performance.

A prime example of how this can go the other way is the batmowheel from bullseye. From what I understand and someone can chime in, the batmowheel is more a gimmick and in fact flows less than the stock cast counter part directly from borg warner.

You'll also notice that companies like precision only put a HP rating on their wheels and do not supply maps. I would assume because they are tweaking their profiles actively, the cost associated with producing maps isn't worth it. I believe they rate their turbos based on calculations but I also believe they are relaying heavily on real world application.

One thing to note is that compressor wheels haven't changed all that much over the last 30 years. If you're trying to size a turbo from precision you can compare an older t series garrett map but take into account that the precision unit will flow more. It will be a little misleading, but will get you in the right spot.

 

Boosted, have you run across any/many groups using CFD analysis? I do a lot of analysis work (FEA, CFD, thermal, multi-physics, etc.), and I'd imagine it's a tall order to try to do a CFD analysis with a spinning turbine (unless you're NASA and can burn the taxpayer money like it's going out of style....lol...I've done it with some impellers, and it was a bit of a challenge for the software). Just curious, since the analysis tools have gotten much more powerful and accurate over the past 20 years, I wonder if these complex analysis has worked it's way down to the small companies. I know I could probably set up and successfully run these analysis, but it might still be at a point where emperical testing coupled with quick-turn CNC is more efficient (in the development world...which I live in....it's often a compromise).

Oh, my big prob is spec'ing a cam. Looks like I just need to talk to a few of the cam companies, as their specification templates on their websites seem geared more towards N/A vehicles and not boosted ones. Who have you had the best experience with?

Last, you gonna put that thing on a dyno or the track? Me being a numbers guy wants to be fed some numbers, even though that wasn't your stated objective early on.