Bosch Throttle Body (Holder) Woes

My 360 started going into limp mode a while back. It was subtle at first, with it starting to lose power after 4,000 RPM's and then less subtle when it would barely go above idle. Limp mode protects the car when some vital part has malfunctioned.

I know a master Ferrari tech that is nice enough to answer questions for me about the car. He lives more than five hundred miles away and it's like telemedicine. The codes indicated a problem in the drive-by-wire circuit, likely culprit: One or more TB's or the accelerator potentiometer (AP), which is the foot pedal sensor. I had a good code reader that could monitor the AP voltages (2) and they appeared to be within specification. That left the TB's, one or both.

After a long discussion with the tech, his answer was to replace the throttle bodies (TB). The Ferrari 360 has two TB's. One for each bank of four cylinders. To the Ferrari mentality, replacement could only be with two new TB's: "Ferrari likes replacing things in pairs." I totally understand this mentality and the reasoning behind it. The TB's are an important part of the car and a major component in the powertrain. Replacing one would leave the engine with one strong, new unit and one old, weak TB. They had to be in balance to work effectively.

TB's on most cars are reasonably priced, typically below $200 for even a Porsche. Ferrari's, being special, have an added "Ferrari Tax" that increases the price of spares. In this case, the list price for a 360 TB is $1750. And I should replace two. That's $3500 for just the parts. Then of course, while I'm at it, might as well replace the Accelerator Potentiometer (AP) for another $400. In for a Penny, in for a Pound.

Naturally, the first thing I did was ignore the advice of an expert. I bought a used TB from eBay, supposedly from a car that was running when it crashed. 

From discussions with the tech, we determined it might possibly be the left unit since it seemed to have a longer lag time returning from full to idle using the "finger" test. To do this test, the MAF (Mass Air Flow) sensor housing is removed, the TB connector removed and the throttle plate moved to full by pushing it with a finger. Letting it go and looking for some magical indication is how the test worked. It seemed like the left unit had a longer lag. No science was used in this test.

Ferrari 360 Throttle Body: Bosch 0 280 750 038 

Removing and installing a TB is a level 1 on the difficulty scale. The parts are above the engine, held in place by a single ear crimp clamp and four hex screws. 

Getting the electrical connector off is a bit of a trick. It requires a squeeze and jiggle action to get it off with a struggle. An easier method I found is to pull back the rubber grommet from the connector, then slip a thin screwdriver to the side, between the TB and below the connector and pry up carefully while squeezing the tabs. It would be really bad to break the connector so care is required. The connector on the left side can be accessed best once the TB is free from the intake. The right one is easier to access and can be removed while attached to the intake.

Here's a picture of the connector for reference.

The other tricky bit is getting the engine vent off without breaking the clamp. The clamps are called "single ear crimp clamps" and can only be tightened once.

A special crimping plier is used to crush the little spacer at the top, drawing the clamp tight. They are extremely good at holding a seal. Problem is they are use-once only. Generally, the way to remove them is to break them by lifting the little tab (on the left side in this picture). That ruins the clamp.

I found that if an awl is carefully worked into the crushed zone, the clamp can be loosened enough to remove it without breaking. To retighten, just use the crimp clamp again. This should only be done on lines that are not critical. The engine vent leading into the throttle body has no pressure, carries no fluid and if it fell off, wouldn't damage anything.

I installed the eBay unit on the left side and took it for a test. All seemed good. The car ran fine for 100 miles. After that, it was back to limp mode and the trouble codes returned. The fault could have been that I had purchased a bad unit from eBay, both my TB's were bad or I guessed wrong and the right one had the problem. Too many variables, too many unknowns. An unsolvable equation.

Up to this point, a TB was a magical bit that did something amazing to make the car go fast. I knew of them and what they did but that's all. How exactly they function and fail was a complete mystery. Since my tech essentially said there's no way I could test them, I had to assume they were malfunctioning.

Before shelling out $4k for the repair, I thought it would be a good idea to see if I could understand what a TB does, and I mean exactly what. How can it be tested? Can they be fixed? Are there any sources of spare parts? 

Back to the Basics

The internet is a wonderful place to do research. I found all sorts of interesting general information on Drive-by-Wire and lots of theory. 

Before: Cable Driven Throttles

A cable driven throttle powered by a human foot does only one thing: Open the throttle. A throttle controls the amount of air entering the engine. More air (+ fuel) equals more power. This is fine for a 1960's era car with a carburetor but a modern car that has to meet emission and MPG standards needs a much finer control system than a foot powered cable actuated throttle. 

Now: "Drive-by-Wire"

Cars switched from cable driven throttles to electronic ones for essentially two reasons: Performance and efficiency. 

When the driver presses the pedal, they want more go, more power. Opening up the throttle valve to let more air in does this but in a rather crude and inflexible way. Allowing a computer (the ECU) to do the thinking and set the air/fuel ratio to a much finer degree would allow the engines to be more efficient, higher powered, and cleaner too.

The electronic throttle control system, or Drive-by-Wire, allows the engine to understand what power level the operator wants and then decides how much fuel and air the engine needs to deliver it with peak efficiency.

Drive-by-Wire System Components

Three essential components make up any drive-by-wire system.

Accelerator Potentiometer (AP)

Throttle Body (TB)

Engine Control Unit (ECU)

The AP

The AP has two potentiometers (pots) that work in such a way that the lower is half the volts of the higher. So if the high pot reads 2v, the low will read 1v. The more pedal, the higher the volts. This relationship holds regardless if the pedal is pressed or released. 

The ECU monitors this relationship as a double check that the pedal is actually moving and it's not a stray voltage or malfunction of the pot. When the two don't align within a given tolerance, an ECU trouble code is generated and the operator may see a "Check Engine Light" (CEL) displayed.

The failure modes for the AP are limited to failure of either the linkage, the connector or either of the sensor pots. Since this component is in the cabin, it experiences a benign environment. 

Most good quality OBDII scanners can read the the voltage of both pots and the position of the pedal. This makes it quite easy to check on the functioning of this part.

The TB

The TB, which controls the air flow into the engine, has several key parts:

Throttle valve - No different than the foot driven models: a butterfly valve. Very robust. A possible failure mode would have the valve encrusted with gunk and crud, which would cause turbulence in the airstream and restrict flow. Obvious answer is to clean it.

Actuating motor - A high performance carbon brushed DC PWM motor. A lot of power is passed through the motor every time the pedal moves. The motor doesn't spin more than a dozen times in either direction and spends most operating time holding a setting. In a Ferrari, the motor gets a pretty good workout as many drivers (me included) use it as a binary device. Either on full, or off. The motor fights the torque of the spring to open and the spring closes the throttle when torque (electricity) is removed. It takes far less power to hold the throttle at a certain level than move it higher. The motor can weaken with age and rough handling. The insulation on the windings can fail, causing shorts in the motor and increasing the power required. The carbon brushes can wear out and when they are nearing that point, become less effective, which adds resistance into the circuit and consequently, increasing lag. The bearings can wear, the lubrication can dry, and carbon from the brushes can clog everything up. These failure modes may show up as excessive throttle lag and of course, trouble codes from the ECU.

Reduction gears - Allow using a lower torque motor and fine positioning control. A three gear train reduces motor RPM's by about a factor of ten. The gears are under a lot of load and stress from the normal engine compartment environment. If they start to fail, the backlash in the system increases and throttle positioning errors will creep in. Eventually, the gears can strip out.

Sensor pots - Two potentiometers are used to assess and confirm the position of the throttle valve. These sensor pots work differently than the AP in that one pot goes from high to low, the other from low to high. When the throttle is closed, the high will read a low voltage while the low will read a high voltage. Mid-throttle will show both pots with the same relative voltage while full throttle will show the high at a high voltage and the low at a low voltage. The real numbers appear to be at idle: High - 0.75v, Low - 4.25v. They add up to 5v regardless of position, another sanity check for the system. A typical failure for the pots include dead spots, an entire pot failing or the total voltage from the two pots beyond limits.

Return Spring - Pretty much the same as the foot powered spring except this spring has to be precisely calibrated to function properly. The failure mode for this part is breakage or becoming weak. Spring technology is well known and these springs probably are good for decades of service and unlikely to fail.

Engine Control Unit (ECU)

The ECU ties together the input from the AP with everything else it knows about the operating condition of the engine and decides how the throttle should be set. It's a lot more complex than push the pedal and the throttle opens. If the ECU fails, it's likely to show up as car that won't start or even crank in many cases.

Marvel of Engineering

The TB is amazing for what it does. It uses a pretty solid and basic method of motor control that will last many years and endure all kinds of temperatures, vibrations, salt, water, oil, pretty much anything. When they fail, they fail gracefully so if a gear strips out, the throttle will close from the spring. If the motor dies, it closes. If the pots fail, the ECU cuts power and the throttle closes. Essentially, it won't cause the engine to speed up. The engine will continue to run at idle too, so if the car has to limp home, it can do so at a walk.

For most cars, simply discarding a failed TB and replacing with a new one is rather inexpensive. It's an easy part to change out. With a Ferrari, the part is literally ten times the cost of even a Porsche equivalent. 

Can a Throttle Body be Rebuilt?

Since the cost of a new TB set on my car is about $3500, perhaps my old units could be rebuilt? Unfortunately, a search of Google returned nothing useful. TB rebuild kits for a Porsche include the TB. Why rebuild something so inexpensive? 

After looking at dozens of photos and specifications of other similar Bosch 0 280 750 series TB's, I came to the conclusion that the only notable difference between a TB for a Cayenne (474) and a Ferrari 360 (038) is a vent near the throttle valve stem plug. The bodies of the TB's are otherwise identical. Same diameter throat. Same mounting points. I couldn't see a difference. The 474 is one tenth the price of an 038. I ordered two from Amazon for $165. I love Amazon's return policy. If they were not usable, I could send them back, no problem. Upon physically examining the 474 and comparing it to an 038, I couldn't find anything that was different except the addition of the engine vent on the 360 model. That was the only difference.

Qualifying Spares 

Are the internal components of the 0 280 750 series the same? If they are, then a new TB of similar size should yield a full rebuild kit that includes every moving component. A simple swap of parts (it's never simple) would allow rebuilding the body into a like new unit. 

In order to do this, I had to qualify the parts and the final rebuild with a series of tests that would characterize the unit. The tests would have to check the pots, gears and motor to see if they react similarly to the existing 038's.

The Problem

The difficulty in determining if a TB has malfunctioned or another TB would be a source of spares is that I would need to test each unit to find out the specifications. Bosch doesn't appear to publish any sort of details about their Motronic system. For instance, what's the pot voltage? What's the motor circuit voltage? What PWM frequency does the motor circuit use? 

The Answer

After pondering and studying the internet on such topics, I came up with some reasonable assumptions I could make to start the process. An Arduino would make an ideal platform to develop a test since they are cheap, reliable and I have a dozen of them in a box, along with all the stuff needed such as oscilloscope, multimeter, breadboards, boxes of every possible electrical and electronic parts and a place to work.

The hardware part of this project was pretty easy and made with inexpensive parts. It's the software that's complicated.

The Software

Microprocessors are a wonderful invention. The entire computer and all sorts of switches are incorporated into one tiny chip: ATMEGA328P. It comes as either a prototype board with lots of available stuff built right into it, or as a raw chip. The raw chip needs a few bits to function, like a clock and a stable source of 5v power, but that's it. An entire computer with more processing power than the lunar lander on one tiny chip.

My impressions in the past of such tiny computers was that they were nothing but toys and lacked the reliability to do anything other than silly stuff. They weren't serious pieces of kit.

My opinion changed after I read an article in the trade journal "Nuts & Volts" that discussed how useful they were in real world settings. Another article later had a review of a kit for a board (known as a shield in Arduino speak) that had a bunch of hardware on it such as switches, a speaker, potentiometers and LED's. The Arduino board plugs in below it, on standard pins. The review was positive and it sounded like a good way to investigate the world of Arduino.

My first Arduino project, a shield with lots of switches, buttons, pots, LED's, and a speaker. I soldered everything except the surface mount features. 

It doesn't take a lot of skill to work with modern electronics. Whatever you want to make has already been done a dozen times and many of the folks who did them wrote an article, posted in a forum or made a YouTube video on how to do it. I've been able to do whatever I needed so far with hardly any math. The real power of a microprocessor is the software.

Using the Arduino language and common add ons, I wrote a software system that can run a TB effectively. The software can find the limits of the TB valve, find the minimum duty cycle, set the valve at any percentage open, detect if the gears are stripped or if there are dead positions in the pots. It can determine the lag time from idle to full throttle and also check the pots to be sure they are in spec. It can check the power usage and display the results of the tests on a small LCD screen. Additionally, all the data is logged in 5 microsecond intervals for later analysis.

Throttle Body Health Check (TBHC)

When the TBHC first starts, it prepares the components for testing. That includes calibrating the amp meter, checking the full range of the pots and initializing the SD card for logging.

Gear Test

If the gears are malfunctioning or stripped out, the initialization test will fail and the TB cannot be tested further. It has failed mechanically and the electrical part of the TB is unavailable.

Lag Test

During the initialization of the TBHC, the throttle is run to the limit by applying full power and checking for when the pots stop changing, which is when the valve is fully open. These limits are used for the pot one maximum and pot two minimum (pot two goes from high volts to low volts). Once these limits are found, the range of the pots can easily be calculated. The lag is a byproduct and is the time from full power command to limit reached. The return from full to zero is also logged, including lag time.

Pot Test 

Everytime the throttle valve moves, the pots are checked against each other. To be sure that every possible spot on the potentiometer is within specification, the throttle has to move slowly. In this test, the throttle moves in two percent jumps. The two pots are compared by adding them together and checking the variance from the ideal, which is 5v. I used a 50mV limit to trigger a log entry. I believe the ECU allows up to 200mV.

Amp Test

During every move the milliamps required are monitored and recorded. The highest amperage is saved and displayed on the LCD screen. The log records the time, throttle position (pot one), and the average amps being drawn from the last log entry (about 5 milliseconds).

Stress Test

To verify the condition of the unit, a stress test consisting of running the throttle from zero to full and back to zero a number of times in rapid succession. The pots and amperage consumption is monitored and recorded.

Minimum Duty Cycle Test

This test determines the minimum amount of force required to move the throttle plate from idle to full. Starting out at zero duty cycle, the cycle is gradually increased until the throttle moves all the way to the limit. The motor in the throttle body uses PWM or Pulse Width Modulation to control the amount of power the motor receives. A PWM duty cycle is the percentage of the wave that's in "high" voltage (as opposed to low) per frequency cycle.

Example of a PWM signal.

As indicated in the above image, 100% duty cycle means the voltage is high at the entire time. 0% has the voltage low. The Minimum Duty Cycle is a good test for the TB health in that it shows exactly how much power is required to overcome friction and electrical issues in the motor. During this test, the amperage is recorded and the Pot Test is done as well.

Here's a video showing the tester in action.

Log Entries

A microSD card is used to record a csv like file. It's not pure csv since it has header and trailer records for each test. To analyze the data, the specific chunk of csv proper data can easily be imported into a spreadsheet for graphing and other analysis.

Example of the data.

For a quick check on comparing the two pots, add the tps1Low to tps2High, which is 1023.  tps2Low plus tps1High equals 1027. The ideal value would be 1024. The voltage in the pots ranges from 0v to 5v and the sensor high value is 1024. So to determine the millivolts at any setting, take the number and multiply it by 4.88. This proves that the pots are working, at least at the ends of travel within a tight range.

Here's what the tester hardware looks like.

It's been an evolution so a lot of messy wiring results. The next step would be to clean it up, ditch the Arduino board for a bare ATMEGA328P and logically place the components on one breadboard. After that, a soldered prototype and if I really get ambitious, a custom circuit board. 

Hardware Components

Arduino Uno

LCD screen

Motor controller 

microSD card reader/writer

Current sensor

 

The real work in this project is the software. Lucky for me, I'm a software engineer. No sense boring anyone with those details.

The three interesting results were the MDC, Minimum Duty Cycle, the maximum current draw and the lag from idle to full throttle.

TB #1, left unit from my car: MDC 37, milliamps 2450, lag 280 milliseconds

TB #2, right unit from my car: MDC 61, milliamps 2350, lag 320 milliseconds

TB #3, eBay unit, first purchased: MDC 35, milliamps 2560, lag 310 milliseconds

TB #4, eBay unit, second one: MDC 40, milliamps 2540, lag 300 milliseconds

TB #5, Porsche unit model 474: MDC 27, milliamps 3040, lag 240 milliseconds

TB #6, Porsche unit model 474: MDC 26, milliamps 3110, lag 245 milliseconds

The two new ones matched up really well. The right side (#2) from my car looked the worse and had a nearly 2x factor in the MDC from the other units. That one must be the giving me the codes. 

I installed the left TB back into the left intake and an eBay unit (#3) in the right as they seemed to be a matched set. I've had conflicting reports as to the proper method to reset the ECU on the car so that it goes through a relearning cycle. The ECU has to understand the limits on the pots in order to set the throttle at an exact position.

My tech buddy said I needed to perform a "throttle actuation" (goes by many names) in order to reset the values in the ECU. Others have said it's simply a matter of driving the car as it relearns all the time. I have an Autel MS906 which can do the reset.

I drove the car and it's running very nicely. After ten miles, no codes. The proof will be if I can go past 150 miles without throwing any trouble codes.

Rebuilding a TB

One of the main points in this project is to determine if a TB can be rebuilt using donor parts from a different TB model. I have the two 474 units from Amazon that are virtually identical to the 038 unit on the 360.

The TB's come apart easily. The clips can easily be removed if they are pried from the curved side. Two small screwdrivers opposed pops them right off. Keep your finger on them or you'll be searching the other side of the room for the clip. Take care when removing the plastic face. Pull straight up to keep the motor contacts from being bent. 

The inside of both units was clean and without any sort of lubricant. No grease or oil of any kind is inside the case. The motor itself appears to ride on a bronze bearing that's probably self lubricating. Not a lot of spinning going on in this motor so a simple bronze bearing is fine. I saw no evidence of degraded gears. They looked like new. 

The pots are mounted in the plastic face. I could see no easy way to remove the part from the face. Five tiny wires carefully soldered connect the pots to the face. The consequence is that the new plastic face should be used if the pots are going to be included in the rebuilt. That means the wrong part number will appear on the outside of the TB. If that's a bother, then perhaps a way can be devised to remove and reinstall the pots. 

The pickups for the pots are tiny flexible wires on a spring arm. They rub against the pot which appears to be coated with something. The old unit showed no sign of degradation so I opted to leave the old pickup and gears in place. The pot pickups are mounted on the throttle valve pivot. In order to change the gear or the pickups, the throttle plate would have to be removed and the pivot would have to be pressed out of the housing. I decided to leave that to a future rebuild. I'm sure these parts can be removed, just not today.

The pots on the old unit had obvious wear. The tracks were shiny and a visible groove existed where the pickup wires travelled.

Worn potentiometers on the old throttle body

The central gear just pops out. The motor armature is held in place by two torx 25 screws. Before you remove the magnets, mark them. Switching their polarity will cause the motor to either not run at all, or run in reverse. I put the magnets in upside down in one of the rebuilt units, even though I MARKED THEM. Not only mark them, but pay attention to their orientation when reinstalling. 

Here's what they look like broken down:

Bosch 0 280 750 038 Throttle Body parts

Bosch 0 280 750 474 Throttle Body parts

I installed all the old parts into the new donor body first to see if they actually fit. The reinstall process was painless, except for the magnets. They are held in place by a spring clip that has to be pressed into place. I found a screwdriver worked just fine for this task. It was just a little tricky to find the right leverage point. The plastic face went on without any issues. I found it easiest to start by aligning the pot side of the face, then gently closing it. That way the motor contacts are aligned when the face is lowered. Seemed to go easier that way.

The 038 went together the same way as the 474. No surprises. Word of caution: Care must be taken when installing the armature into the motor housing. It goes in with a spring fit. If the motor contacts aren't lined up just right when the screws are tightened, the motor will bind. I found that pushing the armature down manually and then screwing it tight worked.

Once they were both together, I put them on the tester and check the results.

038-2-Right

Before: MDC 61, milliamps 2350, lag 320 milliseconds

After: MDC 32, milliamps 2635, lag 269 milliseconds

474-5

Before: MDC 27, milliamps 3040, lag 240 milliseconds

After: MDC 39, milliamps 2562, lag 321 milliseconds

I can see from the results that some losses still are present in the old rebuilt unit. As seen from the new unit rebuilt from old parts, the MDC is higher but not as high as the old unit before rebuild. The components I did not change, mainly the pot pickup, plate pivot and gear. 

The performance of the rebuilt unit is much improved though. It went from an MDC 61 to 32. That's better than the two working units in my car. The amps are higher from the new brushes and the lag has been reduced considerably.

Here's a note from an owner (WBT in New Zealand) who successfully rebuilt his throttle bodies using the instructions in this article.

"It was a very straightforward process. I followed your guide and after unclipping the plastic cap on the new units I marked the top of the magnets for direction and right/left locations. Then I disassembled the Ferrari units and replaced all the removable parts from the new units. 

I included the new plastic end cap as well as that carries the potentiometer tracks which are a wear item so I figured that new would be better. The only parts that are original Ferrari issue now is the actual metal body and throttle butterfly and driving gear (they still appeared like new and moved freely with no play). 

The hardest part was reassembling the magnets and locator spring. A set of fine needle point pliers popped the spring back in place then I pressed it down flush once inserted with a wide blade screwdriver. Whole job done in an hour. If I had to do it again it would be 10 minutes for both. The rebuilt throttle bodies have been working flawlessly ever since.

As you say new OEM throttle bodies are now unattainable (and even more worryingly so are the Motronic ECU's which are unrepairable) so this fix is going to become essential for every owner in the very near future. The peace of mind knowing the throttle bodies are both reliable and now unlikely to smoke my ECU H-bridge is a very good feeling."

(A slight correction: The H-Bridge circuit in the ECU is actually repairable so save your ECU if it's been burned out by a bad TB)

UPDATE: Two years later.

I started having issues with my throttle bodies again. If I let the car sit for a couple of weeks and it was cold, sometimes I'd see limp mode. Turn the car off and back on again and all was well. That's how it started before.

I decided to make an adapter kit so that the plentiful 474 throttle body could fit directly on the 360. The kit consists of an adapter sleeve to replace the OEM sleeve (Ferrari 160701) and a connector pipe to move the engine vent from the OEM throttle body (Ferrari 171716) to the adapter sleeve. It works great, and I have the adapter kit available for sale at this link:

Ferrari 360 Throttle Body Adapters

UPDATE: More information.

I wrote an article on how to make a basic throttle body tester using inexpensive parts here:

Make a Bosch Throttle Body Tester