MechE

Hello everyone,

I am having some idle and boost problems, and it seems the computer may be at fault. I purchased this car a few weeks ago from a fellow AutoX'er and he told me that the car simultaneously started idling poorly and the electronic boost control stopped working back in October.

Here are the problems/symptoms I am aware of:
-ECU boost target of 12psi is not met, and the computer seems to be doing nothing to manage boost - 100% mechanical right now.
-The car dies 2 seconds after the first start when the engine is cold. The car will usually stay alive and idle on the second start.
-When the car is shut down hot, and restarted immediately or 15 minutes later, it will not idle smooth; revs oscillate from 400-1,500rpm and settle down after 20 seconds to ~800rpm. Behaves much like an low-dampened control system oscillation that overshoots and then smooths out. This can also happen without shutting the car down on a hot day.
-Idle also oscillates severely when the AC is on.
-Compression isn't perfect: Cyl #1,2,&4 are about 180psi, Cyl #3 is 160psi. (dry) Leak down test is on the list to determine where the leak is in Cyl #3.

Since the lack of electronic boost control seemed to be the easiest problem to tackle, a good portion of this weekend was spent thoroughly inspecting the electrical and mechanical components around this. Here's the checklist:

• Wastegate can be controlled manually and the car was able to create 12psi boost around 4k rpm. Wastegate: check. Turbo: check.
• Boost Control Solenoid was tested and functions properly.
• Wiring running to a toggle switch in the cab (to switch between mechanical and electronic boost control) was checked for continuity and checked out okay. The toggle switch properly closes and interrupts the circuit.
• The light green w/yellow stripe signal wire, from the TEN port of the Diagnostics Connector running all the way to Pin 2H on the 26-pin Engine Control Module Connector was checked for continuity and is good.
• Then the connection between the TEN port and A5 in the hydra connector and that also checked out perfectly.
• All wiring was doubled checked again fully assembled and the wiring appears to be perfect.
• Monitored the TEN port while driving and no signal is coming from the ECU to trigger the BCS.
• Data-logged while driving and the ECU says it is putting out a "wastegate PID (%)" as boost builds, but no signal is actually being sent out through this signal wire.

Please let me know if anything was missed or should be looked into further, but this seems thorough and really points to the ECU being the problem. Based on what the previous owner has told me, it seems a driver board in the ECU has failed, which may be tied to something else in the idle circuit to cause poor idle conditions? Is this a likely scenario? I'm hoping I missed something and I don't have to buy a new ECU... :(

I have attached some screenshots of datalogs (HotStartIdleOscillation.PNG, 3rd&4thPull.PNG, 3rdPull-12Hz.PNG) and a short .xls log at ~12Hz of a 3rd gear pull with a little bit of 4th gear too. Somehow I am no longer able to log faster than 1-2Hz... I have no idea why I could get 12 a couple weeks ago, and no matter what I do now I get 1-2Hz... :/

Thank you!

fredb

iTrader: (1)

The hydra could be at fault but I'd start with a mechanical boost control and a fresh base map from Jeremy at FM. The EBC on the hydra has never worked properly ( and is unlikely to ever) and the rest of what you describe just sounds lik bad tuning. Fred B

codrus

When you say that "no signal was detected", what were you using to look for it? You know that the Hydra switches ground (and not power) on the solenoid, right? You've confirmed that the Hydra software is configured to have EBC enabled in all of the right places? At a minimum, it probably needs a boost target set, to have open or closed loop EBC enabled, some base settings for the PID/etc, and also some configuration for outputs/triggers. Are you using an oscilloscope to look at it, or just a DVM?

At the open house, Jeremy was saying that he's gotten open loop EBC to do something semi-reasonable with Hydra 2.7, but yeah, with 2.5 it's pretty much useless. Before I switched to the MS3 I used a GReddy standalone boost controller to do EBC, which worked very well. It's probably worth figuring out why the EBC isn't even trying to do anything (if for no other reason than that it will teach you about the Hydra), but in the long run I don't recommend actually trying to use it. A manual boost controller or a standalone EBC is a better choice.

As for the others -- 2.5 does not have great idle control because the only controls the ECU has for that are the fuel and the IAC. Other, newer ECUs (including Hydra 2.7) will adjust timing to keep idle in the desired range, and this works a lot better.

If it's dying 2 seconds after startup on cold start, then your start fuel is wrong. There are three or four different settings for fuel on startup, and they stack on top of each other. Have you read Jeremy's Hydra document? I forget what all the settings are, but they're all listed in there.

The warm restart problem is probably also fuel related (or possibly PID settings). Getting this right requires a lot of fiddling, but it's never going to idle like a factory car. Idle oscillation with the AC on means you need to retune the PID settings.

Bad compression isn't something the ECU can fix.

--Ian

MechE

Yes, I am aware the Hydra is sending a ground signal through the light-green/yellow wire and the other end of the solenoid is on 12V+. I used an automotive DVM to see if there was a signal and when nothing showed up, I hooked up a light in place of the EBC to see if it would illuminate. With a duty cycle of 50% in the logs, I would expect a light to behave as if it were on 6 volts. Nothing on either. Because both you and Jeremy @ FM asked if I checked it with a oscilloscope (I could if I could hit boost parked in my garage!), I hooked up a 10MHz data logger and went for a drive after work. There was no signal, just noisy floating voltage in the +/- 0.6V range (data logger in place of the EBC).

I just ran through the instructions for enabling the EBC and checked one of the maps I loaded into the car after my data logging. EBC was not activated, but everything else was setup. It is too late right now to check to see if the map that I have been running and testing is also not activated. I'll load that map back into the ECU and check it out. It's unfortunate that I cannot open various maps on the PC without being connected to the car. It seems I can only see what is in the car, and then go offline from there... I would really just like to see how that map was setup.

Thanks for questioning all the steps. I knew I was bound to miss something. I'm hoping it is as simple as activating.

Thank you for the info. The previous owner provided a new-in-package FM manual boost controller that I could use, but he claimed the EBC worked great. We shall see I suppose. It's an easy thing to swap out and test.

I looked up the GReddy standalone BC you have and I'm a little shocked by the price. How complicated can that device really be? It controls one thing. I'm sure its worth the price, I just need to learn more about it...

With both you and Fred saying it sounds like a bad tune, I decided to run a different map provided by the previous owner. Your comments reminded me that he said the map in the car now was tweaked to run better after whatever happened in October. I loaded his last confirmed "good" map from 2012 and the hot start issue is gone and the idle when the AC is on is now revving high instead of low. Haven't had a chance for a cold start yet, but clearly it was a bad tune.

I'm thinking I may just need to start from scratch now... Recalibrate everything and work on building a good tune.

Haha, I could have described that list a little better... All known problems with the car.


Thanks for the detailed reply! I'm really hoping the EBC just wasn't activated and there is nothing wrong with the ECU. Unfortunately, I won't have time to find out until Wednesday night.

Oh, and one side note: I did do a datalog on the 2012 map and despite the EBC not being activated, the "wastegate PID %" values show up in the log just like the previous logs. This supports the fact that it could have been disabled all along...

Cheers,
Steve

fooger03

iTrader: (2)

Forget about the EBC until the rest of your problems are sorted. EBC is it's own separate problem. If the previous owner said that the *hydra* EBC worked great, he was either lying or stupid; either way, he was wrong. There are ways of *tricking*it into working, but it's open loop, and at that point it's really less of an "electronic boost control" and more of an "electronic boost suggestion".

I've built some methods using 3D PWM tables in order to drive actual control, but I never derived what the EBC equations look like in the 2.5. In the 2.7 I can define and control through P and I equations, but I have not identified a viable method of enforcing a D value.

The rest of your problems should be easy enough to fix with a base map from Jeremy and instructions for initial setup. Good luck!

Once you get a good running setup, you can delve into the EBC settings if you are dedicated. I have documented my own research here: http://www.mazda-speed.com/forum2/in...topic=27981.15 My posts start at #26, but you can read canyonfive's posts up to that point to see how he was tricking it into working.

fredb

iTrader: (1)

If you don't have a lot of expierence tuning or have someone near by who does , the best thing you can do is lean heavily on Jeremy at Fm . Even if that means paying the $130 for the life time phone support. It'll easily save you that in grief. Any way you slice it you've got a steep learning curve ahead. If your hydra is a 2.5 , well .....that's not going to make it easier. I'm on a 2.7 and it works pretty well ( EBC excluded) but I'd have a tough time recommending spending the money to upgrade. Hydra's business model pisses me off almost as badly as Apple's . Fred B

codrus

I've done a bunch of troubleshooting using a small, portable oscilloscope (one of the Rigols) powered off a 12v inverter. It needs a second person in the car to run it, but it's worked pretty well. The data logger should do it in this case, though.

Yeah, I think it was $400 when I bought it over a decade ago, I can only imagine what they want for it now. Speaking of which, now that I'm done converting to MS3, I really should go post that in the classifieds...

Yeah, I would be inclined to get a new base map from Jeremy, figure out the difference between that and the one that's in the car, and port over only the differences that you *know* are right. It's annoying that Hydra uses a binary map format so you can't read it directly. I started reverse engineering the file format once, hacking up a perl script to decode it, but never got around to finishing it, so it only displays a few things. TS for the MS3 uses an XML format, which is much easier to read.

--Ian

MechE

Thanks for the link. I had no idea the EBC could be so terrible. I have no problems moving to a functional independent unit... My reason to dive into the EBC now is that I thought it was a problem with the computer, which has now been proven to be a mistake - the ECU is fine. Yay!

Now that I know my ECU is fine, I can invest the time to start a better map and deal with boost control later.

I've talked to Jeremy a couple times this week so far. I have been holding off on the base map and the life time support until I determined the ECU was okay. Those things would have came free with the $900 ECU replacement/upgrade. Thanks for your opinion on the value of upgrading to 2.7 - I haven't seen a good reason to do the upgrade to it.

The plan for this car is to see more autoX and track time (and the time driving to these events) than daily driving. If 2.5 can be tuned well enough for those events and idle isn't perfect on the street, I'm happy to live with that.

Regarding the tuning, my father specialized in tuning cars, so I think he will be able to help out quite a bit. I will certainly rely on forums and Jeremy at FM to help out as well.

I have a Rigol as well, but no inverter. I've thought about buying an inverter on more than one occasion, but I know some equipment requires a full-sign-wave inverter to prevent damage, which is a different price tier and not worth it for the few times I would have a use for it, IMO. At least I have a data logger.

There seems to be quite the range for EBC's... I'll have to do some more research on them and determine if I should buy a nice one or just settle for a manual controller. From what I've read so far, manual controls boost as well as the cheaper EBC's.

I really wish maps could be viewed simultaneously and not have to be connected to the car. It would make comparing them so much easier.

MechE

too many posts... please delete

codrus

I drove my car for 7 years using Hydra 2.1 and 2.5 (very similar in functionality), so yeah, it can certainly be made to work. The problems can be worked around and are of the "annoying" kind, rather than the "can't run it" kind. Idle quality is the worst, but you can alleviate a lot of it by opening up the throttle screw to raise the mechanical idle to around 1000 RPM. I never did get it to idle right with the AC on, though.

--Ian

MechE

Okay guys, it appears the EBC was never activated in any of the maps that came with the car... Mystery solved. Everything was setup except settings/output/PWM3 was turned off. I ran all the "confirmed good maps" and picked the best one, and proceeded to turn on EBC. Basically I have a map from 2012, 2013 and 2014. 2013 seems to run the best having very little idle oscillation after a hot start and handles the AC pretty well (1200-1400 rpm rapid oscillation) at idle.

Each map had slightly different EBC settings, so it appears that the previous owner did play with it. Anyways, I went for a drive and very quickly learned the EBC was now activated. I also learned it was not going to do anything about over-boosting either... I very quickly jumped from 8psi to 16.5psi in less than a second and within 150rpm. Ouch. :( Screenshot of log attached.

After that I immediately turned the EBC off and installed the manual boost controller that came with the car: Dawes Devices/3 Bar Racing (might be a MKI - looks slightly different)
I spent some time adjusting it and set it to a conservative ~11.5psi for now. Having a log of the EBC and this MBC, there is clearly some boost missing in the lower RPMs with the MBC. I can only assume a nice independent EBC would give me all the boost available up to my target of 12, and maintain 12 well, right?

I have attached three log screenshots of third gear: 2013 map with no boost control, 2013 map with EBC on (over-boost), and the 2013 map with the manual boost controller.

For those interested in seeing more, I have also attached a log of consecutive 1st, 2nd, and 3rd gears for each the 2013 map with no boost control and the 2013 map with manual boost controller.

I think the next step from here will be to contact Jeremy tomorrow and start talking about tuning. Thanks everyone!

fooger03

iTrader: (2)

I tuned my idle and everything else per FMs instructions when I installed my unit as a 2.5. Idle has always been solid since, including A/C on/off cycles, and full-electrical loads. The only woe is the occasional heat soaked IAT sensor.