Mach-Lee
Well-Known Member
- First Name
- Lee
- Joined
- Jul 16, 2021
- Threads
- 203
- Messages
- 7,729
- Reaction score
- 15,299
- Location
- Wisconsin
- Vehicles
- 2022 Mach-E Premium AWD
- Occupation
- Sci/Eng
- Thread starter
- #1
Hope you guys enjoy this, I did a complete teardown and brief engineering analysis on a HVBJB. I’ve been working on this for a while, but bad timing since a bunch of other people happened to all post about it at the same time. Warning, this post will have a lot of technical jargon and engineering data! Component data and measurements for engineers is included.
First, this is part # NK4Z-10C666-C on the box sticker, which is the version for vehicles with standard range packs. The front motor connections will be unused on a RWD vehicle. Dated 23-JUN-2022 (post recall). This will not work in a GT, the connectors won’t fit. This part is designed for a peak power of 266HP/198kW/560A. The GT version is very similar but will have some larger components in places.
Description and Operation:
The High Voltage Battery Junction Box (HVBJB) is located inside the sealed battery pack, and is used to “switch” the high voltage leaving the pack on and off. This internal switch is necessary for safety. When the car is fully shut down or in an accident, the high voltage is completely shut off to the external HV cables feeding the motors and accessories. The electromechanical switching devices inside are call contactors, because they contain a set of contacts that open and close. Closed is on, and open is off. The contactors are controlled and monitored by the Battery Energy Control Module (BECM) which also lives inside the sealed battery pack. There are six contactors inside the HVBJB:
The HVBJB also contains four fuses for each of the high voltage circuits. The fuses are replaceable. There is a current sensor (round black circle in later photos) inside the HVBJB around the positive array terminal which measures all current flowing into and out of the pack. This is connected to the BECM and is used to calculate the % SoC in the battery (coulomb counting). Last, voltage sense wires are connected to the stages after each contactor and are monitored by the BECM. This allows the BECM to monitor if the contactors have opened or closed properly, and calculate the voltage drop across each contactor.
This is a photo of the HVBJB with all components and connections labeled:
Reference: The front of the vehicle is towards the bottom of the photo.
Acronyms in photo:
Array = Battery pack array
CONT = Contactor
ISC = Inverter System Controller (rear motor inverter)
eFAD = Electric Front Axle Drive (front motor inverter)
DCFC = Direct Current Fast Charge
DCDC = DC to DC inverter (provides 12V power)
PTC = Positive Temperature Coefficient cabin heater
You will notice the layout is somewhat symmetrical, all the positive components and connections are on the passenger side, and the negative components and connections are on the driver’s side (LHD). FYI the high voltage negative side is completely isolated from the chassis ground for safety. Negative is not ground in a high voltage system.
The part number on the actual part is NK48-10C666-AB. It is normal for the part number on the box and the on the actual part to be different, although they share the same base (10C666 is the base part number for a HVBJB).
The components on the top of the HVBJB use spade terminals, and can be removed in order to open the top cover. The bottom cover can also be removed to reveal the the internal wiring and busbars.
Cover removed. The current sensor (black donut) for battery current measurement has been taken off the positive array terminal and can be seen lower left. It has a 4-pin connector (C4815E). The white ceramic pre-charge resistor can be seen above. It is pretty hefty at 30W since a lot of power will flow through it for a brief second during precharge. The black auxiliary contactor bridges the gap in the small busbars below it when closed, and sends power to the DC/DC and SOBDM modules. The 50A fuse to the right of it is next in series. The very large 630A cube fuse bolted in is for the rear motor. The 150A front motor fuse is the largest orange one. And the last small 50A fuse is for the PTC heater and A/C compressor.
The metal between the white contactors are the busbars. It is 3mm thick copper sheet that is tin-coated. Very beefy. The large spade terminals for the array connections are 20x3mm.
The center circuit board is used to condense the contactor coil connections into a single 8-pin connector (C4815D) to the BECM. The contactors are ground-side switched. The yellow component is the precharge contactor, which connects the precharge resistor to the + main bus. R1 next to it is a 249Ω resistor in parallel with the precharge coil to suppress inductive kickback and increase operation speed.
Underside. You can see the bottom of the contactors and how they are bolted to the busbars. There are two small wires coming out of the white contactors, those are the wires for the control coil. When 12V is put on these wires, the contactor will close. They are not temperature sensors. There are no temp sensors anywhere in the HVBJB. The small control wires lead back to the center and connect to the bottom of the green circuit board with flag terminals.
The flag terminals scattered around that are not under the center board are the voltage sensing taps. These wires lead back to the orange connector (C4815C). The sense taps are connected to the busbar after each contactor stage so the BECM can detect the voltage after each one. If a contactor sticks open or closed, the BECM will see the voltage on the other side of the contactor where there shouldn’t be and set a stuck contactor code. When this happens all the contactors are disabled and you will see a Stop Safely Now message in the cluster. Since the DC/DC converter is powered through the Aux contactor, when that shuts down the 12V system will not be charged and the 12V battery will die after a while.
The last components we see are the four rectangles with leads, those are 0.1 µF suppression capacitors that are placed on either side of the DCFC contactors and connected to chassis ground. These should suppress EMI coming in on the DC charger that might interfere with vehicle systems.
I removed a DCFC contactor to examine the busbar connection. It appears the busbar shape near the contactor terminals has been maximized to greatest extent possible to fit the angled dividers on the contactor. It is the same shape on all four contactors. This should maximize heat dissipation. You can see the two coil control wires which are used to close the contactor. The contactors are hermetically sealed so the primary heat dissipation is conduction through the bolt terminals.
Overall Physical Specs:
Components:
Contactor electrical data:
Main contactor physical specs:
Component resistance:
Busbar specs:
Findings:
Overall the part seems very well constructed. The components are all automative grade, as expected. The contactors are all made by TE. All measured data suggests the four white main contactors are internally identical, the different suffix of the part numbers (-1, -2, etc.) is likely for minor differences in the external coil wire length and colors. They are a custom part, with no specifications available. They appear very similar to the TE EVC 500A main contactor (whose datasheet is available), but are physically larger and heavier. Another difference is the coil, based on inductance and resistance measurements the internal coil is much larger than the EVC 500 (22Ω/250mH vs. 3Ω/12mH). It seems probable the extra height and weight of these contactors is due to the larger coil alone. The larger coil is probably to reduce the amount of holding current required while the contactors are closed, so smaller wiring and connectors can be used. The size of the terminals is the same as the EVC 500, which would seem to suggest this may still be a 500A rated contactor.
All the contactor measurements seem nominal.
My milliohms meter isn’t great, the readings are slightly over the 0.5 mΩ spec but it's a difficult measurement to make and I wouldn't worry about it too much. The readings still seem low enough to support 500A with a minimal 0.35V voltage drop. However the heat dissipation would still be several hundred watts, so the contactors will get quite warm handling 500A. Heatsinking through the busbar is very important to control internal temps. There are no temperature sensors in the HVBJB.
The busbar design seems to have been made to maximize the contact area with the bolted contactor terminals for better heat transfer. There isn’t much further improvement that can be made to busbar coverage without making the overall HVBJB larger. Terminal nuts were correctly torqued to 10 Nm. The high-current busbars are 3mm thick. The main bus terminals for the battery array are 3 x 20 mm, which is very close to 1/8 x 3/4”. From a busbar sizing reference, this size busbar would support 325A continuous at 105ºC rating. It’s possible a higher temp rating (such as 125ºC) was chosen for the design, which would increase the current rating slightly.
I will try to post a wiring digram eventually, but another important finding is the DCFC contactors are in series with the main contactors. This means all four contactors must be closed in order to DCFC, and the current from DCFC will pass through all the main contactors and heat them.
Conclusions:
A reminder this is the HVBJB for a standard range pack with lower specs (approx 560A peak). And that HVBJB failures are caused by overheating.
Many of you are wondering if the HVBJB is properly engineered, and the short answer is I’m not qualified to answer that definitively. I can only give you my thoughts and opinions. Based on the size of the busbars connecting the pack, I would estimate the continuous rating of this particular HVBJB to be around 350A, which is about 60% of peak power. This should be adequate for everyday driving with occasional 560A maximum acceleration.
Prolonged or repeated acceleration will cause the contactors and busbars to heat up to temps over 100ºC for short periods, this is expected. In extreme cases, temps may reach 200ºC, which is too hot and will melt the contactors. As previously mentioned, heat from the contacts is mostly dissipated through the busbars. Since the busbars must remain isolated, convection and radiation would be the only means of dissipating the heat from the busbar. In a sealed pack, heat saturation could easily occur in the HVBJB with prolonged high-current usage. Thermal saturation remains a question of when, not if. To be successful, the design needs to absorb enough heat under anticipated normal acceleration conditions to not overheat, and slowly dissipate that heat while driving. Since active cooling of the HVBJB really isn’t an option, the only solution is to make the entire HVBJB bigger if the design is deemed inadequate. Unfortunately there is limited space in the front of the Mach-E pack, so only so much can be done to improve the design retroactively. This revised HVBJB seems to maximize what can be done within the confines of the existing design.
The choice to put the DCFC in series with the main contactors also surprised me because DCFC will only continue to heat up the main contactors during a charging stop. I’m not an expert in HV design, but it seems like the DCFC contactors could have been given a completely separate path directly to the pack. This would give the main contactors time to cool down during a DCFC session. And it should still be safe since it would require a double contactor fault before voltage would be seen at the exposed DC charge port pins. I feel this design decision has contributed to HVBJB failures since a number of people had a stuck contactor happen shorty after DCFC. The extra heating from DCFC puts the failing contactor over the edge.
In terms of the recall software interacting with the part, it can measure the resistance of the contacts by monitoring the voltage drop under load. Normally this voltage drop should be well under 1 volt, but if it rises much above that, it means the contacts are going bad and likely overheating. So a power cut can be made. Excessive voltage drop across the contacts is what triggers the “Service vehicle soon” message and 1/3 power cut. This should prevent further damage to the contactors while allowing the vehicle to continue to drive until replacement.
Preventing contactor damage with software is more tricky, and requires a thermal model of the contacts since there are no temperature sensors in the part. A thermal model such as this would require instrumented testing to develop. The software could integrate the pack amperage over time and correlate that to a heating model to “guess” the contactor temperature, and provide intelligent reductions in power when it thinks the contactors are getting too hot. I haven’t seen evidence of this behavior yet, perhaps in a future software release.
It is hoped the combination of improvements in the HVBJB hardware and better software will prevent the part from failing again in the future. Time will tell.
It would be interesting to see an original design HVBJB to understand what changes were made, but unfortunately those all go back to Ford.
Special thanks to @breeves002 for logistics. If you want higher quality pics, PM me.
First, this is part # NK4Z-10C666-C on the box sticker, which is the version for vehicles with standard range packs. The front motor connections will be unused on a RWD vehicle. Dated 23-JUN-2022 (post recall). This will not work in a GT, the connectors won’t fit. This part is designed for a peak power of 266HP/198kW/560A. The GT version is very similar but will have some larger components in places.
Description and Operation:
The High Voltage Battery Junction Box (HVBJB) is located inside the sealed battery pack, and is used to “switch” the high voltage leaving the pack on and off. This internal switch is necessary for safety. When the car is fully shut down or in an accident, the high voltage is completely shut off to the external HV cables feeding the motors and accessories. The electromechanical switching devices inside are call contactors, because they contain a set of contacts that open and close. Closed is on, and open is off. The contactors are controlled and monitored by the Battery Energy Control Module (BECM) which also lives inside the sealed battery pack. There are six contactors inside the HVBJB:
- Main +
- DCFC +
- DCFC –
- Main –
- Auxiliary –
- Precharge +
The HVBJB also contains four fuses for each of the high voltage circuits. The fuses are replaceable. There is a current sensor (round black circle in later photos) inside the HVBJB around the positive array terminal which measures all current flowing into and out of the pack. This is connected to the BECM and is used to calculate the % SoC in the battery (coulomb counting). Last, voltage sense wires are connected to the stages after each contactor and are monitored by the BECM. This allows the BECM to monitor if the contactors have opened or closed properly, and calculate the voltage drop across each contactor.
This is a photo of the HVBJB with all components and connections labeled:
Reference: The front of the vehicle is towards the bottom of the photo.
Acronyms in photo:
Array = Battery pack array
CONT = Contactor
ISC = Inverter System Controller (rear motor inverter)
eFAD = Electric Front Axle Drive (front motor inverter)
DCFC = Direct Current Fast Charge
DCDC = DC to DC inverter (provides 12V power)
PTC = Positive Temperature Coefficient cabin heater
You will notice the layout is somewhat symmetrical, all the positive components and connections are on the passenger side, and the negative components and connections are on the driver’s side (LHD). FYI the high voltage negative side is completely isolated from the chassis ground for safety. Negative is not ground in a high voltage system.
The part number on the actual part is NK48-10C666-AB. It is normal for the part number on the box and the on the actual part to be different, although they share the same base (10C666 is the base part number for a HVBJB).
The components on the top of the HVBJB use spade terminals, and can be removed in order to open the top cover. The bottom cover can also be removed to reveal the the internal wiring and busbars.
Cover removed. The current sensor (black donut) for battery current measurement has been taken off the positive array terminal and can be seen lower left. It has a 4-pin connector (C4815E). The white ceramic pre-charge resistor can be seen above. It is pretty hefty at 30W since a lot of power will flow through it for a brief second during precharge. The black auxiliary contactor bridges the gap in the small busbars below it when closed, and sends power to the DC/DC and SOBDM modules. The 50A fuse to the right of it is next in series. The very large 630A cube fuse bolted in is for the rear motor. The 150A front motor fuse is the largest orange one. And the last small 50A fuse is for the PTC heater and A/C compressor.
The metal between the white contactors are the busbars. It is 3mm thick copper sheet that is tin-coated. Very beefy. The large spade terminals for the array connections are 20x3mm.
The center circuit board is used to condense the contactor coil connections into a single 8-pin connector (C4815D) to the BECM. The contactors are ground-side switched. The yellow component is the precharge contactor, which connects the precharge resistor to the + main bus. R1 next to it is a 249Ω resistor in parallel with the precharge coil to suppress inductive kickback and increase operation speed.
Underside. You can see the bottom of the contactors and how they are bolted to the busbars. There are two small wires coming out of the white contactors, those are the wires for the control coil. When 12V is put on these wires, the contactor will close. They are not temperature sensors. There are no temp sensors anywhere in the HVBJB. The small control wires lead back to the center and connect to the bottom of the green circuit board with flag terminals.
The flag terminals scattered around that are not under the center board are the voltage sensing taps. These wires lead back to the orange connector (C4815C). The sense taps are connected to the busbar after each contactor stage so the BECM can detect the voltage after each one. If a contactor sticks open or closed, the BECM will see the voltage on the other side of the contactor where there shouldn’t be and set a stuck contactor code. When this happens all the contactors are disabled and you will see a Stop Safely Now message in the cluster. Since the DC/DC converter is powered through the Aux contactor, when that shuts down the 12V system will not be charged and the 12V battery will die after a while.
The last components we see are the four rectangles with leads, those are 0.1 µF suppression capacitors that are placed on either side of the DCFC contactors and connected to chassis ground. These should suppress EMI coming in on the DC charger that might interfere with vehicle systems.
I removed a DCFC contactor to examine the busbar connection. It appears the busbar shape near the contactor terminals has been maximized to greatest extent possible to fit the angled dividers on the contactor. It is the same shape on all four contactors. This should maximize heat dissipation. You can see the two coil control wires which are used to close the contactor. The contactors are hermetically sealed so the primary heat dissipation is conduction through the bolt terminals.
Overall Physical Specs:
Part No. | NK4Z-10C666-C |
Overall weight | 11.730 lb |
External dimensions (LxWxH) | 23.3 x 5.9 x 4.0 in |
Components:
Part | Qty | Manufacturer | Part No. |
Main contactor | 4 | TE | 2406901 |
Auxiliary contactor | 1 | TE | 2203997-1 |
Precharge contactor | 1 | TE | V23700-C0001-A408 |
24Ω 30W Precharge resistor | 1 | TE | SQH30 24R? |
Current sensor | 1 | LEM | DHAB S/161 |
630A Rear motor fuse | 1 | Eaton | EV1XSQ-630-TE |
150A Front motor fuse | 1 | Eaton | EV20-150-PX |
50A Auxiliary Fuse (3/8” spade) | 1 | Eaton | 10EV50F4TEJP9 |
50A PTC/ACCM Fuse (1/4” spade) | 1 | Eaton | EV10-50-P |
0.1 µF Safety Capacitors | 4 | KEMET | 413N310040M1K |
Contactor electrical data:
Contactor | Coil Ω | Coil mH @ 1020 Hz | Coil Hold Current @ 12V A | Pull-in V | Drop-out V | Closed contact mΩ |
Main + | 23.2 | 262 | 0.518 | 7.64 | 2.32 | 0.74 |
DCFC + | 23.0 | 222 | 0.521 | 7.55 | 2.38 | 0.70 |
DCFC – | 22.9 | 270 | 0.525 | 7.78 | 2.57 | 0.50 |
Main – | 22.8 | 225 | 0.526 | 7.65 | 2.54 | 0.70 |
Aux – | 21.6 | 22.7 | 0.555 | 6.55 | 2.39 | 0.43 |
Precharge + | 51.5 | 165 | 0.223 | 6.42 | 3.30 | 7.60 |
Main contactor physical specs:
Part No. | 2406901 |
Weight | 631.2 g |
Diameter | 51 mm |
Can height | 78 mm |
Overall height | 95 mm |
Threads | M8-1.25 |
Bolt flange diameter | 15.8 mm |
Nut | 13 mm |
Torque | 10 Nm |
Component resistance:
Precharge resistor | 23.5 Ω |
Front 150A fuse | 0.67 mΩ |
Aux 50A fuse | 1.81 mΩ |
PTC 50A fuse | 1.70 mΩ |
Busbar specs:
Main bar thickness | 3 mm |
Material | Copper bar, tin coated |
Front motor connection | 3/8 x 3/64" spade |
Main Array Spade Terminal | 20 x 3 mm |
Findings:
Overall the part seems very well constructed. The components are all automative grade, as expected. The contactors are all made by TE. All measured data suggests the four white main contactors are internally identical, the different suffix of the part numbers (-1, -2, etc.) is likely for minor differences in the external coil wire length and colors. They are a custom part, with no specifications available. They appear very similar to the TE EVC 500A main contactor (whose datasheet is available), but are physically larger and heavier. Another difference is the coil, based on inductance and resistance measurements the internal coil is much larger than the EVC 500 (22Ω/250mH vs. 3Ω/12mH). It seems probable the extra height and weight of these contactors is due to the larger coil alone. The larger coil is probably to reduce the amount of holding current required while the contactors are closed, so smaller wiring and connectors can be used. The size of the terminals is the same as the EVC 500, which would seem to suggest this may still be a 500A rated contactor.
All the contactor measurements seem nominal.
My milliohms meter isn’t great, the readings are slightly over the 0.5 mΩ spec but it's a difficult measurement to make and I wouldn't worry about it too much. The readings still seem low enough to support 500A with a minimal 0.35V voltage drop. However the heat dissipation would still be several hundred watts, so the contactors will get quite warm handling 500A. Heatsinking through the busbar is very important to control internal temps. There are no temperature sensors in the HVBJB.
The busbar design seems to have been made to maximize the contact area with the bolted contactor terminals for better heat transfer. There isn’t much further improvement that can be made to busbar coverage without making the overall HVBJB larger. Terminal nuts were correctly torqued to 10 Nm. The high-current busbars are 3mm thick. The main bus terminals for the battery array are 3 x 20 mm, which is very close to 1/8 x 3/4”. From a busbar sizing reference, this size busbar would support 325A continuous at 105ºC rating. It’s possible a higher temp rating (such as 125ºC) was chosen for the design, which would increase the current rating slightly.
I will try to post a wiring digram eventually, but another important finding is the DCFC contactors are in series with the main contactors. This means all four contactors must be closed in order to DCFC, and the current from DCFC will pass through all the main contactors and heat them.
Conclusions:
A reminder this is the HVBJB for a standard range pack with lower specs (approx 560A peak). And that HVBJB failures are caused by overheating.
Many of you are wondering if the HVBJB is properly engineered, and the short answer is I’m not qualified to answer that definitively. I can only give you my thoughts and opinions. Based on the size of the busbars connecting the pack, I would estimate the continuous rating of this particular HVBJB to be around 350A, which is about 60% of peak power. This should be adequate for everyday driving with occasional 560A maximum acceleration.
Prolonged or repeated acceleration will cause the contactors and busbars to heat up to temps over 100ºC for short periods, this is expected. In extreme cases, temps may reach 200ºC, which is too hot and will melt the contactors. As previously mentioned, heat from the contacts is mostly dissipated through the busbars. Since the busbars must remain isolated, convection and radiation would be the only means of dissipating the heat from the busbar. In a sealed pack, heat saturation could easily occur in the HVBJB with prolonged high-current usage. Thermal saturation remains a question of when, not if. To be successful, the design needs to absorb enough heat under anticipated normal acceleration conditions to not overheat, and slowly dissipate that heat while driving. Since active cooling of the HVBJB really isn’t an option, the only solution is to make the entire HVBJB bigger if the design is deemed inadequate. Unfortunately there is limited space in the front of the Mach-E pack, so only so much can be done to improve the design retroactively. This revised HVBJB seems to maximize what can be done within the confines of the existing design.
The choice to put the DCFC in series with the main contactors also surprised me because DCFC will only continue to heat up the main contactors during a charging stop. I’m not an expert in HV design, but it seems like the DCFC contactors could have been given a completely separate path directly to the pack. This would give the main contactors time to cool down during a DCFC session. And it should still be safe since it would require a double contactor fault before voltage would be seen at the exposed DC charge port pins. I feel this design decision has contributed to HVBJB failures since a number of people had a stuck contactor happen shorty after DCFC. The extra heating from DCFC puts the failing contactor over the edge.
In terms of the recall software interacting with the part, it can measure the resistance of the contacts by monitoring the voltage drop under load. Normally this voltage drop should be well under 1 volt, but if it rises much above that, it means the contacts are going bad and likely overheating. So a power cut can be made. Excessive voltage drop across the contacts is what triggers the “Service vehicle soon” message and 1/3 power cut. This should prevent further damage to the contactors while allowing the vehicle to continue to drive until replacement.
Preventing contactor damage with software is more tricky, and requires a thermal model of the contacts since there are no temperature sensors in the part. A thermal model such as this would require instrumented testing to develop. The software could integrate the pack amperage over time and correlate that to a heating model to “guess” the contactor temperature, and provide intelligent reductions in power when it thinks the contactors are getting too hot. I haven’t seen evidence of this behavior yet, perhaps in a future software release.
It is hoped the combination of improvements in the HVBJB hardware and better software will prevent the part from failing again in the future. Time will tell.
It would be interesting to see an original design HVBJB to understand what changes were made, but unfortunately those all go back to Ford.
Special thanks to @breeves002 for logistics. If you want higher quality pics, PM me.
Last edited: