[scalbert]

Obviously this will cause some controversy as many cannot leave that out of topics. But here it is.

This is the first step, next will be road tests with portable data acquisition to log various parameters for a comparison to see if any differences in "feel" can be quantified. At this point I am only testing the PROcede Rev I with the 12-15 Map loaded and JB3 1.2F. Below is mainly the cold hard data. There will be some statements as I feel the data gathering warrants it but it will be based on a professional opinion. If your opinion differs, please respond but do so in a logical manner.

As we all know very well, there is a vast difference in the hardware design. The PROcede is a manufactured board using the normal filtering functions and surface mount technology. The JB3 is through-hole design with a minimalist approach as the amount of components on the board could be counted with just the fingers on your hands. Both companies claim their product design is superior and make valid arguments as such. In fact, the simplicity of the JB3 is very interesting that they were able to accomplish what they did and I can absolutely take my hat off to this. However, the lack of filtering and suppression components worries me. In system design you normally do not add additional items for the 99% of the time. It is that unusual 1% that normally requires the extra. But it may end up being a moot issue as time will tell.

As a side note, BMS has also done a complimentary job of servicing their customers with trade in. It is almost like purchasing a support contract, or at least that is my interpretation. It is somewhat unprecedented in this market and is truly impressive.

On the bench we simulated all signals each unit receives and used the same values for each unit. Meaning, the same boost levels, RPM, etc. were used. To do this mapping we had to determine the actual function of each pin. For the PROcede I already had this and referred to my Cheat Sheet so this was easy and publically available. On the JB3, we had to dig some and actually made a full schematic on it which I will not post unless Terry gives the OK.

PROcede Signals

B1S1 O2 Sensor Biasing - Analog Output
B2S1 O2 Sensor Biasing - Analog Output
Speed Sensor - Digital Input and Digital Output
Throttle Pedal - Analog Input, Just Tapped, no Analog Output used
Crank Position Sensor - Digital Input and Digital Output
TMAP - Analog Input and Analog Output
Intake Air Temperature – Analog Input, Just Tapped, no Analog Output used
Boost Control Solenoids - Digital Input and Individual Digital Outputs
Fuel Pressure - Analog Input and Analog Output

JB3 Signals

B1S1 O2 Sensor Biasing - PWM Output
B2S1 O2 Sensor Biasing - PWM Output
Throttle Pedal - Analog Input, Just Tapped, no Analog Output used
Crank Position Sensor - Digital Input, jumpered at terminals
TMAP - Analog Input and PWM Output
Intake Air Temperature – Analog Input, Just Tapped, no Analog Output used
Boost Control Solenoids - Single Digital Output
Fuel Pressure - Analog Input and PWM Output

As you can see, both reference most of the same signals excluding the Speed Sensor which the JB3 does not touch. Is this an advantage to the PROcede, for me, no. I would never push it past 155 but others may want to. The first fundamental difference between the two units is the method in which analog signals are controlled. Both use a traditional analog input A/D. For those who are not familiar, A/D is Analog to Digital conversion. Basically it takes an analog signal, mostly 0 – 5 VDC in this case and converts it to an integer value represented in A/D counts. For instance as with the JB3, it uses a 10 bit A/D. 10 Bit is 1024 steps. If the voltage measurement range is 0 – 10 VDC as with many devices, the A/D function chops the voltage up into these steps and returning a value between 0 and 1023 with 0 being 0 VDC and 1023 being 10 VDC. This integer can then be converted to engineering units using the math we learn early on; the linear equation (Y = MX + B). As mentioned, both perform this in the same manner. The difference comes in the way each retransmits the signal. For the PROcede it uses a D/A converter which is basically opposite of the A/D; give it an integer and it outputs the corresponding voltage. This is the same way the other piggybacks perform this operation. In fact, it is the way practically all control systems handle analog control. For instance, the control system I professionally work on, such as 10,000 gallon chemical reactors, use D/A when analog outputs are needed.

The JB3 takes a different approach by using a PWM signal. What is a PWM? Pulse Width Modulated; in other words, is a series of pulses measured as a duty cycle signal. A PWM or Duty Cycle signal is typically measured in 0 – 100%. That is percent is the amount of time the signal is high; basically on. When using a RMS measurement on a PWM signal it appears as a solid analog signal. For instance, give it a 50% duty cycle and you will measure about 2.7 VDC when using a standard volt meter. BMS must have figured out that the DME does some filtering, most likely RMS, of the signals and could use a PWM output. My only exception to this, as it works fine in this case, is lack of filtering. A PWM signal can contain noise and sometimes harmonics at certain frequency/duty cycle ranges. A single resistor and capacitor would have been preferred as currently the PWM signal is attached right to the terminals. But it may very well not be needed and so far appears as such.

As an example, in the below images you can see the actual signal difference. Both seem to work just as effectively, just go about it differently.

PROcede Analog Input and Output



JB3 Analog Input and PWM Output



The next difference in the way objectives are handled is in the Boost Control Solenoid. The PROcede has isolated Inputs and Outputs for this. In other words, the PROcede has direct control of the solenoids and if its output failed, you would have no boost at all. The benefit to this is that they can do whatever they want with boost. The drawback is that they are at the mercy of the hardware working and would need to put more scrutiny on the control software; it would be easier to get out of control. But if done correctly, there is more capability to be had.

The JB3 sinks the signal through a transistor. Meaning that the DME is still directly connected to the solenoids but the JB3 has the option of adding duty cycle. The benefit to this is that in the event of a JB3 component failure, the DME can still run stock boost levels. The negative is that it limits overall control of the solenoids which is debatable as to if it is needed.

The last major area of difference, or at least it appears to be, is in the timing control. The PROcede, once again, uses isolated inputs and outputs. It reads the crank position sensor signal and retransmits a slightly modified signal. The DME uses the CPS to know where the engine is at in its rotation. If you offset the phase (the delay the timing some), you are effectively offsetting the ignition timing. The difficulty in this is that you cannot have abrupt changes in this signal as the DME will detect it as an implausible signal and set a code and/or misfires can occur.

As for the JB3, it does not appear the JBS offsets the crank position sensor. We ran numerous tests and could never induce this timing offset. It was previously mentioned that the In and Out pins are jumpered together so it makes since why the since look the same. I then thought maybe some IAT control was done, making the DME think it is warmer than it really is, but this is also just tapped and looks normal. I am not going to say the JB3 is not controlling timing. But it does not appear to be easily measureable and if not, how are you sure it really is? There is no question on how the PROcede is doing what it is doing so I didn’t feel there was any harm in illustrating it as most of the intelligence is in the software. I assume the JB3’s intelligence is in the software as well. If so, why not discuss the mechanics of how it is doing this function just to clear the air? Obviously this will be up to Terry to decide on but I welcome it.

For the testing, I had two CPS simulated signals which were identical but optically isolated so that any control influence on one would not affect the other. I used the PROcede Reader software to set up the simulation with the following data:



I ran high boost and hot IAT reading to truly push the timing back. As you can see in the screen shot the PROcede was pulling back timing by an indicated -4.75 degrees. When we calculated the timing offset we saw on the scope we came up with -4.8 degrees as shown below. The top signal is the isolated source; the bottom trace is the PROcede output. If you look at the leading rising edges and compare them in the horizontal (time) you can see the PROcede delaying the signal.



The below is the same sets of signals with no changes to them fed into the JB3. As you can see, there is no timing offset of the CPS signal.



This does not mean the JB3 is not controlling ignition time in some other manner. But it does show there is no CPS timing offset occurring. I am open to suggestions on how it is controlling timing. But at this point it is all on faith.

The next installment will be doing some in car data acquisition. I’ll put together the test application tonight and wire up a patch harness. I will log IAT, RPM, Throttle Pedal Position, Throttle Plate Position, Fuel Pressure and Boost. Any other suggestions? I’ll capture some Tab Delimited files and post them for everyone’s analysis as well as offer mine. What we will be looking at is stability and response.

That is it for now, let the whining begin.

[SfValley335i]

[sflgator]

No whining...just a pat on the back for your time, effort, and technical knowledge to do these comparison tests. Although I am not an electrical engineer and have no idea what you're talking about (most of what you're testing anyway ), I do appreciate what you're doing. So, keep up the great work, Scalbert!

[ar design]

Nerd alert!!!

Just kidding, nice work!!

[DrDomer]

[bubbletea 4 me]

from me too even though i've no idea what you were talking about..lol. ive JB3 but i can appreciate tests/facts as well as learning the pros and cons of each tune.

[scalbert]

I see a Moderator changed the title; good idea!!

[Vudoo4u2]

nice man

good to see it all coming together

look forward to your further testing

[scalbert]

Originally Posted by shiv@vishnu

Maybe Terry can explain.

I'm prodding and hoping he will throw me a bone to help make my work easier.

[RiXst3r]

I love technical stuff

[mike281]

I noticed you said you were using the older Procede Rev1 and the newer JB3 1.2. Just wondering if not using both of the newest products (or both of the older) would make a difference in what you're testing? In other words is there any big differences with the Rev1 and Rev2 that would change anything you have done or will be doing?

Great work so far! Very interesting.

[scalbert]

Originally Posted by mike281

I noticed you said you were using the older Procede Rev1 and the newer JB3 1.2. Just wondering if not using both of the newest products (or both of the older) would make a difference in what you're testing? In other words is there any big differences with the Rev1 and Rev2 that would change anything you have done or will be doing?

Great work so far! Very interesting.

The only difference between Rev I and Rev II would not affect anything done. The same maps would load up.

As for JB3 1.2, Terry indicated he would send me a new chip.

[Dustin Bramell]

Rev I. and Rev II. both use the same firmware and maps. Regardless of the hardware version, the results of this test would remain the same.

Edit: Scalbert beat me to the punch.

Originally Posted by mike281

I noticed you said you were using the older Procede Rev1 and the newer JB3 1.2. Just wondering if not using both of the newest products (or both of the older) would make a difference in what you're testing? In other words is there any big differences with the Rev1 and Rev2 that would change anything you have done or will be doing?

Great work so far! Very interesting.

[OpenFlash]

Originally Posted by RiXst3r

I love technical stuff

+1 For those who aren't sure what scalbert is talking about right now, it would really benefit you to ask questions. The data he is providing explains the fundamentals of engine control. While the data may be intimidating, he concepts are really quite easy and glossing over them would be shame as they will relate to any add-on tuning system used on any car type. It's critical knowledge to have when it comes to distinguishing marketing hype/misinformation from reality. And that just benefits you in the long run.

Shiv

[Juicer]

Originally Posted by scalbert

I'm prodding and hoping he will throw me a bone to help make my work easier.

Maybe Terry will also reveal how he gets the JB3 to produce more power on the dyno than the V3 with the same or less boost.

[jpsimon]

Great work scalbert! I think I can speak for everyone and give a big "thank you" for this work you've been doing.

In other news, having only "faith" that a tune is controlling timing isn't good enough in my book. It'll be interesting to see if Terry can explain how he goes about accomplishing it with the Juice Box

[scalbert]

Originally Posted by Juicer

Maybe Terry will also reveal how he gets the JB3 to produce more power on the dyno than the V3 with the same or less boost.

Perhaps many things will be revealed in the upcoming weeks. I would not lay down any bets in certainty at this point either way.

[OpenFlash]

Originally Posted by Juicer

Maybe Terry will also reveal how he gets the JB3 to produce more power on the dyno than the V3 with the same or less boost.

I really haven't seen any boost logs from JB3s. Probably because they can't log boost. So I'm not sure where you are getting this info from. But assuming your statement is correct, the only way to get more power when holding things like boost/fuel equal is to run less or no timing retard under boost.

And then ask yourself why do we run more timing retard? Running less or none is just a key stroke away. The only maps that run little or no timing retard are race gas maps. Do you know why?

Shiv

[normtrum]

Wow. I read it twice and I'm still scratching my head.

I appreciate the effort your making scalbert. I look forward to the next set of testing, even though it's more then a bit over my head.

[RiXst3r]

This is from the jb3 launch details... its what was provided by Terry when we first started taking orders for the jb3 back in august. It might help people understand what is what, and what i/o is used for which things.

The Features:

Microcontroller programming opens up a new level of features that the JB3 can provide. Here is a partial list:

1) CPS offset (AKA "timing control"), which allows a little more boost on pump gas than we can with the JB2 variants. In addition boost will be mapped by RPM which is now fed in from the Crank Position Sensor.

2) Air intake temperature based boost decay. As temperatures rise due to ambient conditions and extended loads, the ECU actually raises boost. Which is good enough for the factory tune, but with the JB2, BMS been limiting boost to account for the most extreme conditions.
With IAT based boost decay the JB3 will auto-tune itself on the fly for the current conditions. High-temperature (>90 degrees) performance will be similar, but when temperatures are cooler we will be able to up the power considerably. Those of you with intercoolers will really benefit from this change.

3) Throttle based boost input. The JB3 will adjust the boost targets as a function of throttle input for smoother partial throttle performance.

4) Fuel pressure remapping. Part of the 29.2/v81 compliance package, by remapping the fuel pressure we can sneak more fuel past the ECU. Since the ECU is measuring fuel consumption now to determine if you have a tune this is a critical benefit for diagnostic invisibility.

5) "Direct solenoid control". The JBX series relies on vacuum line configurations to trick the ECU in to a higher solenoid duty cycle, which actually works well. But the JB3 handles the I/O of the solenoids directly, eliminating the need for any changes to the vacuum lines. BMS has developed a method to adjust the solenoid duty cycles without having to replicate the internal boost control system. One that I might add BMW has spent millions of dollars developing to perfection. The result is a smooth non-oscillating boost curve, without spikes... and an ECU that is very happy.

6) Multi-position persistent (e.g. stays where you set it) map switch. To allow you to go between stock, 91, 93, and the race gas tune on the fly with the turn of a dial. No more turning the car off first, and no more issues with leaving the race map on 24x7! Wireless options are not planned for now, but there are other options in the works for easy map switching from in the car.

7) Lean-cruise mode, for those of you looking for maximum fuel economy on your private roads. Ideally selectable via the map switch. BMS is also toying around with the idea of nerfing boost on this map to 2-3 psi, so your 335 feels like a 328. Perfect for when your "buddy" needs to borrow it.

8) Plug and play harness. JB3 harness is color keyed and completely plug and play for super easy install. Power and ground come through the harness directly so there are no external taps or grounds, and no extracting of power wires.... Just the 4 ECU sub connectors.

[Juicer]

Originally Posted by shiv@vishnu

I really haven't seen any boost logs from JB3s. Probably because they can't log boost. So I'm not sure where you are getting this info from. But assuming your statement is correct, the only way to get more power when holding things like boost/fuel equal is to run less or no timing retard under boost.

And then ask yourself why do we run more timing retard? Running less or none is just a key stroke away. The only maps that run little or no timing retard are race gas maps. Do you know why?

Shiv

No. Why?

[Adamk24]

This is awesome, although some of this kind of files in that 'uh ok, cool thanks I guess... ' part of my brain it is interesting that the JB3 doesn't have a clear method of timing control. I have a suspicion that is is controlling it somehow but I have no idea how it would be if it does no have the taps directly for the system. It seems to me the only other methods of even affecting timing control would be through modifying sensor signals to the ecu for other parameters such as the IAT as you suggested. However I would think that doing so would have knock on effects on other areas besides ignition timing. Some explanations on how they make this work would be nice