A tale of 3 turbo setups
 

All tuned on the same dyno. Temperatures were different so SAE correction was turned on. All are close to 200 kPa. All are 1.8L with some overbore.

https://www.miataturbo.net/attachmen...ine=1316808255

3 cars (blue, green, gray), and 3 things are plotted:

- Torque (highest trio of plots)
- MAP (middle trio of plots)
- BMEP/MAP aka torque per MAP, (lowest trio of plots)

The 3 cars are:

Blue: GT3071, Rebello 99 head, FM turbo mani, 3" FM exhaust, AEM, 9.0 pistons, 91 octane

Green: GT2560, FM 99 head, FM turbo, 2.5" FM exhaust, Hydra, 9.0 pistons, 94 octane

Gray: GT2560, Rebello VVT head with VICS mani, ETD shorty tubey mani, BEGI custom divorced downpipe, 2.5" custom exhaust with metal cat, AEM, exhaust cam gear advanced 5* (crank), 8.6 pistons, 93 octane

All of them hit MBT at ~15 psi.

BMEP per MAP is an indication of V.E. and proper timing.

Looking at BMEP/MAP traces
Some observations:
- the VVT car (gray), has the broadest torque spread in the midrange
- the green car loses VE at the top (no I don't think it's the 2560 running out of breath, compare to gray car with same turbo)

The below 4000 RPM ignition timing tuning may not have been well optimized in the blue and green cars. The gray car was, although it wasn't spooling as well as it should. (see MAP curves, it's a lot later than the green car, and only a bit earlier than the blue 3071 car!)

The VE of the 3071 (blue) at high RPM, and the VVT/2560 car (gray) is very similar, and may be a testament to either VVT or the ETD short tubey turbo mani vs. the FM log.

The way the green car's VE (looking at BMEP/map) spread in the midrange, seems to suggest that its cam timing may be off, strongly biasing low RPM and spool over topend!

gray please. I'll pm you for my mailing address.

Isn't one 1.6, one 1.8, and one 1.9?

No, all 1.8 with some overbore.

Here is the green car (2560, 99 head), comparing 12 psi dyno tuning (light green traces) and 15 psi (dark green traces) after dyno tuning:

https://www.miataturbo.net/attachmen...ine=1316811301

Note how the BMEP/MAP plots suggest that the turbo isn't choking at the topend, when going from 12 to 15 psi. There is little loss of efficiency at all RPMs. The weak topend appears to be something else (cam timing?)

Here's the blue car (3071) going from 15 (light blue traces) to 18 psi (dark blue traces), both after dyno tuning.

https://www.miataturbo.net/attachmen...ine=1316811414

There is slight loss of efficiency going to 18 psi from 5500-7000.

Some commentary, comparing gray and green cars:

https://www.miataturbo.net/attachmen...ine=1316812875

The shelf-up in BMEP appears to be due to wastegate opening (sudden drop in TIP).
In the green car, there appears to be a hole in the timing. This hole is very visible in the BMEP curve, and not easy to see in the torque curve, (see plots in post 6), because it's happening during spoolup.

Pulled it apart to check the cam timing this morning:

https://www.miataturbo.net/attachmen...ine=1316890297

At TDC:

https://www.miataturbo.net/attachmen...ine=1316890297

It looks the same as other photos I find online...

--Ian

Turns out I had a loose wastegate during my dyno tuning, bummer. VVT should be spooling up at least as quick as Ian's car (green traces)

You know, I wouldn't mind it if a thread on BMEP/MAP was stickied with these dynos and Jason's explanations as examples. The way that it clearly shows little dips and peaks in efficiency (that would lead to further examination and tuning) is really cool. I've seen Jason talk about this before, but seeing it in pictures . . . .

Prop to you, my friend.

I don't have a recent dyno plot but that looks like it spools about as fast as my big laggy thing. I bet it screams down low with the WG fixed.

The extra 800 RPM of spool makes a HUGE difference in around-town driving.
Amazing how much I lost from < 1mm of slack in the wastegate rod.

Ya the hole in the ignition timing in post #8 would is not visible in the torque curve. It's only visible in the Torque/MAP curve bec it's in the spoolup region.

I don't know for the life of me why dyno software doesn't plot torque/MAP automatically with every pull.

It would also help tremendously when boost is changing when trying timing changes, so you know if the improvement in torque was due to the timing change or due to the increase in boost.

Jason, have you seen Virtual Dyno?

https://www.miataturbo.net/ecus-tuning-54/virtual-dyno-59374/

I've been playing with it, seems to work pretty well. I'm in contact with the developer. Maybe I can convince him to add a Torque/MAP trace. Would be really simple in that type of program.

Back in 1998 I wrote some virtual dyno routines for Mathcad and MATLAB.

I also gave the smoothing and calculation algorithms to Glenn for DataLogLab.

I've been around... :)

In AEMLog (free), you can do smoothing and plotting of RPM_acceleration divided by MAP, vs. RPM.

Someone on the AEM forums also wrote some macros for AEMLog that takes car weight, gearing and aero drag to do the same.

AEMLog can accept text datalogs as input.

Virtual Dyno is exactly this. It's free and the author is now a member here.

Yes, torque divided by MAP in there would be very useful. Torque divided by MAP divided by displacement would allow comparisons across different engines.

I hope it picks up in popularity so that dyno manufs will get off their collective asses and add it.