Need replacement for an obsolete mosfet/component

[MikaJuur]

I'm having trouble finding a component to be able to repair an EML box for a BMW850i from 1991.
It's ON4239. I haven't found any datasheet to compare the old one with a modern replacement.
Does anyone have some insight?

[fzabkar]

I expect it's a special order from Philips.

FWIW, ON823 is an NPN Darlington pair. TIP132 is claimed to be a "slightly different" substitute.

https://www.radiomuseum.org/tubes/tube_on823.html

Perhaps you could consult Philips databooks of that era (1990)?

https://mirrorservice.org/sites/www.bitsavers.org/components/philips/_dataBooks/

https://mirrorservice.org/sites/www.bitsavers.org/components/philips/_dataBooks/1986_S04a_Philips_Low-Frequency_Power_Transistors_and_Hybrid_Modules.pdf

[coromonadalix]

sadly  you may need to buy an used  bmw module  ??

is there any markings near the transistor leads   like ebc  or sdg ??

[MikaJuur]

Unfortunately there are no markings. As I have figured out from the repair manual for the car this component is the main switching transistor supplying a square wave for regulating the electronic throttle.

[fzabkar]

FWIW, this "Bosch Motronic 0261200183 per Audi V8 3.7" unit uses the same PHON4239 transistor, only it's described as a Bosch device:

https://www.facebook.com/ClubItalianoElaborazioniElettroniche/posts/riparazione-centralinebosch-motronic-0261200183-per-audi-v8-371-gestore-segnali-/336165950163268/

This web site claims that the ON946 can be substituted with BD202, although it is "slightly different":

https://www.radiomuseum.org/tubes/tube_on946.html

[james_s]

What does the blown transistor do? It's entirely possible that some generic replacement will work just fine.

[fzabkar]

Unfortunately there are no markings. As I have figured out from the repair manual for the car this component is the main switching transistor supplying a square wave for regulating the electronic throttle.

I'm wondering how a single transistor can control the throttle position. The Bosch literature suggests that the throttle body has an electric motor, which would mean that it would need to rotate in both directions. This in turn would require a H-bridge driver circuit.

Bosch Motorsport | Electronic Throttle Body datasheet:
https://www.bosch-motorsport.com/content/downloads/Raceparts/Resources/pdf/Data%20sheet_68749835_Electronic_Throttle_Body.pdf

Bosch Electronic Throttle Body engineering drawing and pinout:
https://forums.haltech.com/download/file.php?id=1352

A square wave drive would be consistent with a stepper motor.

The only alternative scenario I can think of would be if the throttle valve were restrained by a return spring, and if a solenoid were acting against the spring by controlling its voltage via PWM.

[MikaJuur]

Yes the throttles are controlled by stepper motors. The V12 egine is divided in 2 sides, cylinders 1..6 and 7..12 each side controlled by the EML (electronic throttle control) through some more electrical boxes
I have traced the source and drain pins from the on4239 to the EML connector and they go to pin 18 and 35.

Electrical troubleshooting manual for the BMW850 (E31): https://www.dropbox.com/s/x0i9bk9ipdu39f7/BMW850%20etmdrm.pdf?dl=0
From PDF page 80 and forward I have found out the pin assignments.
Pin 18 is input  "Voltage supply plus,   connection=DME main relay cylinder 1..6,   signal=on-board voltage"
Pin 35 is output   "Drive throttle actuator,  connection=actuator cylinders 1..6,   signal=Square-wave signal"

(on page 69 the EML throttle Valve is described. It seems to be a PWM signal as fzabkar suggested)

[fzabkar]

"By way of a servo motor, the EML throttle valve is opened by the output stage in the EML control unit against the force of a spring. The output stage activates the servo motor with a square wave voltage which has a fixed frequency. The different settings of the EML throttle valve are achieved by altering the ration between the cut-in and cut-out duration."

AIUI, pins 35 and 36 of the EML control the throttles for each half of the engine, so there must be two transistors of the same type in the control box. Also, the motor is a servomotor, not a stepper.


Edit:

The following thread suggests that "PH 0N823 M 9108" is the transistor that controls the throttle for cylinders 1 - 6 in a 1992 850i:

http://www.bimmerboard.com/forums/posts/566567

[MikaJuur]

"By way of a servo motor, the EML throttle valve is opened by the output stage in the EML control unit against the force of a spring. The output stage activates the servo motor with a square wave voltage which has a fixed frequency. The different settings of the EML throttle valve are achieved by altering the ration between the cut-in and cut-out duration."

AIUI, pins 35 and 36 of the EML control the throttles for each half of the engine, so there must be two transistors of the same type in the control box. Also, the motor is a servomotor, not a stepper.

Thank you. I missed that. It is true, it is a servo motor. I assumed it was a stepper because I read some manual earlier with decription over the Siemens EML IIIS. In my car it is a Bosch EML.

On the board I've found
4 x BUZ10 (one blown)
2 x ON4239 (one blown)
3 x ON823
1 x TO-220-5 I haven't checked yet because it has a riveted heatsink covering the component.

I really appreciate your help guys. You have given me a lot of info to go on.

[fzabkar]

On the board I've found
4 x BUZ10 (one blown)
2 x ON4239 (one blown)
3 x ON823
1 x TO-220-5 I haven't checked yet because it has a riveted heatsink covering the component.

I would start by metering the good transistors and determining whether they test like BJTs or MOSFETs or Darlingtons.

Your experience and one other source would suggest that two transistors per side are involved in throttle control. Could the ON4239 be a PNP high side switch while the ON823 is PWM-ing the low side? That would provide some redundancy if either transistor were to fail with a C-E short.

[MikaJuur]

Your experience and one other source would suggest that two transistors per side are involved in throttle control. Could the ON4239 be a PNP high side switch while the ON823 is PWM-ing the low side? That would provide some redundancy if either transistor were to fail with a C-E short.

Exactly. I just found this:

Function
The M70 used in the E32 and later in the E31 is the first BMW engine to use an electronic throttle
control instead of a mechanical connection between the accelerator pedal and the throttle valve. The
German term is Elektronische-Motor-Leistungsregelung (EML). The EML consists of a pedal position
sensor, i.e. a potentiometer connected to the accelerator pedal, an EML control unit and electric
actuators on the throttle valves. The DK are held in the closed position by a spring and can be opened
up to an angle of 88°.

The DK motors move the butterfly valves via reduction gears. They are supplied with a pulsed direct
current whose cycle length determines the torque output and thus the opening angle. The motors can be
controlled in both directions, i.e. opening and closing. The motor of DK I receives its operating voltage
via contacts 17 and 35 of the EML control unit. For the DK II, contacts 16 and 34 are used. To open the
DK, positive voltage is applied to contacts 17 and 16, whereas contacts 35 and 34 are grounded. This is
reversed to close the DK.


The shafts of the buterfly valves are connected to a potentiometer which reports the opening angle back
to the EML control unit. In addition, a throttle position switch with a series resistor opens at 17°. A
circuit diagram is shown in Figure 1. The values of the resistors can vary from unit to unit [4][5]. The
accelerator pedal is also connected to a position switch which closes at 9°. Both do not influence the
EML’s function, they only serve logic monitoring purposes. In normal operations, both switches should
not be open at the same time. However, if this is the case, the EML control unit reduces the engine
power by closing the throttle valves.

I would start by metering the good transistors and determining whether they test like BJTs or MOSFETs or Darlingtons.

Is there some easy way to figure out if the part is a BJT or MOSFET or Darlington?
I have ordered a Peak Atlas DCA Pro DCA75, but it hasn't arrived yet.

[fzabkar]

Now it sounds like a H-bridge, which means that there would be 4 transistors per side. :-?

I would use the diode mode of your multimeter to measure the BE, BC, GS, GD junctions of each transistor. I would also determine which transistors are connected to ground and which are connected to the positive supply.

[MikaJuur]

Hi guys!

I made a simple schematic over the connections if someone is interested.

I tried my DCA75 Pro and got the following...

ON823:

• NPN Darlington BJT
  Pins 1-2-3 B-C-E
  hFE=35 at Ic=5,01mA
  Vbe=1,245V at Ib=5,00mA
  VceSat=0,583V at Ic=5,0mA and Ib=1,00mA
  IcLeak=0.000mA
  Rshunt=5530R
  with protection diode
• (The presence of the resistive shunt across the base-emitter will mean the gain measurement (hFE) will be very low (as the resistive shunt will dominate the base terminal current at the low levels used by the DCA Pro)

ON4239:

• PNP Darlington BJT
  Pins 1-2-3 B-C-E
  hFE=34 at Ic=5,01mA
  Vbe=1,096V at Ib=5,00mA
  VceSat=0,497V at Ic=5,0mA and Ib=1,00mA
  IcLeak=0.000mA
  Rshunt=5950R
  with protection diode
• (The presence of the resistive shunt across the base-emitter will mean the gain measurement (hFE) will be very low (as the resistive shunt will dominate the base terminal current at the low levels used by the DCA Pro)

[fzabkar]

That's an unusual H-bridge. :-?

Since the ON4239 appears to be the complementary PNP equivalent of the ON823, and since other sources recommend an NPN TIP132 as a substitute for the ON823, I would look for the complementary PNP equivalent of the TIP132. According to the previously referenced Philips databook, the PNP equivalent is TIP137 (page 11).

If I were doing this job, I would seriously consider replacing all 4 transistors in the H-bridge, even if they test OK. I would also examine the throttle motor. Are the two throttle actuators interchangeable? If so, then it might be an idea to swap them, just in case the problem reappears. Then you will be able to determine if the problem stays with the actuator or the H-bridge.

[MikaJuur]

That's an unusual H-bridge. :-?

See page 2 of this PDF on how the EML controlles the throttles. https://www.dropbox.com/s/xem9orgx81mobb9/199708IS_electronic_throttle_control.pdf?dl=0

Since the ON4239 appears to be the complementary PNP equivalent of the ON823, and since other sources recommend an NPN TIP132 as a substitute for the ON823, I would look for the complementary PNP equivalent of the TIP132. According to the previously referenced Philips databook, the PNP equivalent is TIP137 (page 11).

I was thinking the same.

If I were doing this job, I would seriously consider replacing all 4 transistors in the H-bridge, even if they test OK. I would also examine the throttle motor. Are the two throttle actuators interchangeable? If so, then it might be an idea to swap them, just in case the problem reappears. Then you will be able to determine if the problem stays with the actuator or the H-bridge.

Yes I will change all transistors. I'm pretty sure there's nothing wrong with the throttles. The original fault was corrosion on the pcb where the main connector is soldered on the board. It made the EML restrict the car into limp mode. Unfortunately while measuring the EML unit before opening it some pins got shorted and killed one ON4239 and BUZ10. A small job just got bigger

[fzabkar]

It appears that the asymmetry in the H-bridge is explained by the fact that there is a spring working with or against the direction of motion. Interesting design ...

[james_s]

I have no experience with these electronic throttles, but at a glance it sounds very similar to the Bosch motorized valves used to regulate idle speed in many older European cars. I have two different systems in my Volvos, the earlier one uses a valve with two coils fed with voltage from the center, ground one of the other terminals and the valve opens, ground the other and it closes, these are driven by a PWM signal. The newer system uses a 2 wire valve with a spring return to close, it is driven with a PWM signal to open it. It sounds like this throttle system is similar in concept to the latter.

Incidentally, the connector for the idle air control valve in the Volvo 740 is the same as the connector on the throttle position sensor and I found out the hard way one night that if you accidentally swap these it blows a transistor in the ECU that controls the IAC valve. In that case it was a fairly standard NPN Darlington and I found something in my junkbox that worked as a replacement.

[MikaJuur]

Solution:

First of all thanks to everyone who helped me out with this, both understanding the throttles/EML and finding equivalent components.

I changed the components as below. In this application they seem to work as the original ones. The throttles are working as before. 

BUZ10 ---> STP24NF10 N-channel mosfet
ON823 ---> TIP132 Darlington NPN power transistor
ON4239 ---> TIP137 Darlington PNP power transistor

and I also had to change one shorted zener at the burnt out mosfet.