Volkswagen Eos Forum banner
1 - 10 of 10 Posts

·
Registered
Joined
·
8,486 Posts
Discussion Starter · #1 · (Edited)
This question keeps coming up and there's a lot of hype from led bulb sellers about stopping lamp errors and wrong myths written about Canbus.

Our cars are designed for tungsten bulbs or with OE Xenon HID headlamps. A tungsten bulb has a very low off resistance of an ohm or two and takes the calculated voltage/wattage current when it's lit. V.W have 2 ways of checking a lamp in your car system. The first at key on checks the lamp cold, is fitted and the bulb isn't blown, the second test when the lamp is powered checks its current against the rated tungsten bulb wattage.

Leds can take less current compared to tungsten bulbs for the same light output and when they are off they have a higher resistance compared to a tungsten lamp. A led replacement can either produce a lamp fail at IGN on (the cold test) or when it runs at lower current (lamp hot test), or both.

What does CANbus compatible mean? Well, the lamp has nothing clever to change how CANbus sees it and what dash lamp warning is given. Lamps whether tungsten or led are analogue devices and there is no digital data in or out of them. Each important lamp in the V.W system is connected to a solid state relay chip inside the BCM (Convenience/Electronic module). MY07 is fitted with BCM 3C0 937 049 AE same as Golf V and I have a spare to peek inside. Others with later EOS may not have this version but the principles may be the same.

Here's a board photo:

Circuit component Hardware programmer Electronic component Electronic engineering Font


The large chips located mainly under each connector are the current switching solid state relays coming in different channel numbers and current values. Higher currents will be for large loads like headlamps. For those old timers who think their car is full of plug in clacking relays - re-think modern car electronics!

Here's the main chip list:
VND920 2ch 36A each, VNQ830 4ch 6A, VND830 2ch 6A, VN920SP e 1ch 30A, TLE63-23GP_ 6ch low current.

Each chip channel has a current sensing output but they don't all use the same method. The 2ch 36A chips VND920 (headlights) just have one analogue voltage output sensing current. 1 amp is approximately 0.5Volt. Whereas the 4ch 6A chips VNQ830 internally detect a bulb connected when off and a current threshold when hot outputing just a logic status bit. Off state bulb connected detection uses a selected resistor value to set the trip threshold which is >105mA. These chips used for smaller bulbs don't appear to have current sensing so I think bulbs other than headlamp bulbs are only monitored for fail open circuit NOT actual running current. It's therefore feasable to ensure a led replacing a small bulb takes a minimum load current of >105mA at 9V. or 1 Watt. This could be most small led replacements fitted with an internal parallel resistor.

As an aside, you can work out from a wiring diagram which chip relays are switching which circuits. If you ever get a wierd problem of two circuits staying always on and they are being switched from the same chip, you can guess your BCM module has a bad relay chip.

Canbus: No mention from me yet! Canbus is only a monitoring detection system for lamp failure. The SSD relays in the BCM module provide an analogue output corresponding to lamp current or a logic status, the BCM firmware then collects these voltage levels or logic status from each 'channel' (a single load) to generate the CANbus control words and then errors. Comparing the different chips functions for large headlight and small lamps, large bulbs are hardest to deal with because actual load current is being sensed and the BCM CPU will have levels pre-programmed in code corresponding to normal running. Of course different manufacturers, models and BCM units can do things differently.

There is no quick and easy solution to this frustrating problem of trying to replace tungsten bulbs with leds but here are some thoughts:

Small bulbs should be easier because AFIK the chips only need to know they are taking a minimum current equivalent to about 1W power.
A fix might disable bulb monitoring completely (remove the dash bulb!) but get bulb errors faults stored.
A fix could disable bulb monitoring selectively for some bulbs not others.
A fix could let the system see a led bulb like a tungsten bulb and still signal a failure (or over current) (Perfect solution).
A fix could add a parallel resistor to smaller bulbs which is feasible but its value should be chosen to keep the lamp fail off but put it on if the leds fail.
A fix for larger headlamp bulbs could add a high power resistor in parallel with the led to simulate the 55W running power of the tungsten lamp it replaces. This resistor will get hot and would have to be bolted to metalwork or cooled with a fan. There would still be an issue with cold lamp detection but cold check warnings may be supressed with coding - unfortunately it's all warnings not per lamp. Suppressing, ignoring lamp or disabling lamp warnings defeats the purpose of letting you know a lamp has failed. In the ideal world any fix would see the lamp as normal but still warn if it failed.

So far there are no simple smart solutions for headlamps. I found some work done for Skoda. It wasn't in English and you would only want to experiment on a spare BCM module. The BCM has the current monitoring relays and CPU to monitor each lamp relay. This work using VAS 5040 and PC ODIS (V.W proprietary) looked at an original copy of the V.W module firmware data file and modified each lamp channel block to selectively disable lamp fault messages. The author had identified the addresses for each lamp. This fits one of the fixes above - selectively disable a lamp fault in the BCM firmware.

It's not ideal because you have no led lamp failure monitoring. The perfect step would have been to identify and change the current sense voltages values for headlamps and change a few resistors for smalled leds replacing bulbs. I could play with my spare BCM unit and try changing their resistors detecting open load, but it seems headlamp bulb conversions are what everybody wants to do and that's the hardest perfect solution to find. High power headlamp bulb monitoring is only 2 X 2 channel SSD chips and 4 x current sense outputs. A hardware mod. wouldn't be too difficult. E.g A for a lower current 25 Watt led to be seen as running at the same power as a 55W lamp, the current voltage output sensed from these chips would need to be twice its present value. This should be doable with an experimental on board hardware hack using a SOIC quad opamp level changer on a small surface mounted board involving no changes to the BCM firmware code, but it requires on board micro surgery. For now, most will have to put up with generic claims of CANbus compatibility and large hot load resistors. Or go to HID with adaptation already available and deal with any lesser problems of aftermarket HIDs tripping lamp overload, because they may not incorporate soft start in their design.
 

·
Registered
2008 Volkswagen VR6 Eos
Joined
·
749 Posts
This question keeps coming up and there's a lot of hype from led bulb sellers about stopping lamp errors and wrong myths written about Canbus.

Our cars are designed for tungsten bulbs or with OE Xenon HID headlamps. A tungsten bulb has a very low off resistance of an ohm or two and takes the calculated voltage/wattage current when it's lit. V.W have 2 ways of checking a lamp in your car system. The first at key on checks the lamp cold, is fitted and the bulb isn't blown, the second test when the lamp is powered checks its current against the rated tungsten bulb wattage.

Leds can take less current compared to tungsten bulbs for the same light output and when they are off they have a higher resistance compared to a tungsten lamp. A led replacement can either produce a lamp fail at IGN on (the cold test) or when it runs at lower current (lamp hot test), or both.

What does CANbus compatible mean? Well, the lamp has nothing clever to change how CANbus sees it and what dash lamp warning is given. Lamps whether tungsten or led are analogue devices and there is no digital data in or out of them. Each important lamp in the V.W system is connected to a solid state relay chip inside the BCM (Convenience/Electronic module). MY07 is fitted with BCM 3C0 937 049 AE same as Golf V and I have a spare to peek inside. Others with later EOS may not have this version but the principles may be the same.

Here's a board photo:

View attachment 24685

The large chips located mainly under each connector are the current switching solid state relays coming in different channel numbers and current values. Higher currents will be for large loads like headlamps. For those old timers who think their car is full of plug in clacking relays - re-think modern car electronics!

Here's the main chip list:
VND920 2ch 36A each, VNQ830 4ch 6A, VND830 2ch 6A, VN920SP e 1ch 30A, TLE63-23GP_ 6ch low current.

Each chip channel has a current sensing output but they don't all use the same method. The 2ch 36A chips VND920 (headlights) just have one analogue voltage output sensing current. 1 amp is approximately 0.5Volt. Whereas the 4ch 6A chips VNQ830 internally detect a bulb connected when off and a current threshold when hot outputing just a logic status bit. Off state bulb connected detection uses a selected resistor value to set the trip threshold which is >105mA. These chips used for smaller bulbs don't appear to have current sensing so I think bulbs other than headlamp bulbs are only monitored for fail open circuit NOT actual running current. It's therefore feasable to ensure a led replacing a small bulb takes a minimum load current of >105mA at 9V. or 1 Watt. This could be most small led replacements fitted with an internal parallel resistor.

As an aside, you can work out from a wiring diagram which chip relays are switching which circuits. If you ever get a wierd problem of two circuits staying always on and they are being switched from the same chip, you can guess your BCM module has a bad relay chip.

Canbus: No mention from me yet! Canbus is only a monitoring detection system for lamp failure. The SSD relays in the BCM module provide an analogue output corresponding to lamp current or a logic status, the BCM firmware then collects these voltage levels or logic status from each 'channel' (a single load) to generate the CANbus control words and then errors. Comparing the different chips functions for large headlight and small lamps, large bulbs are hardest to deal with because actual load current is being sensed and the BCM CPU will have levels pre-programmed in code corresponding to normal running. Of course different manufacturers, models and BCM units can do things differently.

There is no quick and easy solution to this frustrating problem of trying to replace tungsten bulbs with leds but here are some thoughts:

Small bulbs should be easier because AFIK the chips only need to know they are taking a minimum current equivalent to about 1W power.
A fix might disable bulb monitoring completely (remove the dash bulb!) but get bulb errors faults stored.
A fix could disable bulb monitoring selectively for some bulbs not others.
A fix could let the system see a led bulb like a tungsten bulb and still signal a failure (or over current) (Perfect solution).
A fix could add a parallel resistor to smaller bulbs which is feasible but its value should be chosen to keep the lamp fail off but put it on if the leds fail.
A fix for larger headlamp bulbs could add a high power resistor in parallel with the led to simulate the 55W running power of the tungsten lamp it replaces. This resistor will get hot and would have to be bolted to metalwork or cooled with a fan. There would still be an issue with cold lamp detection but cold check warnings may be supressed with coding - unfortunately it's all warnings not per lamp. Suppressing, ignoring lamp or disabling lamp warnings defeats the purpose of letting you know a lamp has failed. In the ideal world any fix would see the lamp as normal but still warn if it failed.

So far there are no simple smart solutions for headlamps. I found some work done for Skoda. It wasn't in English and you would only want to experiment on a spare BCM module. The BCM has the current monitoring relays and CPU to monitor each lamp relay. This work using VAS 5040 and PC ODIS (V.W proprietary) looked at an original copy of the V.W module firmware data file and modified each lamp channel block to selectively disable lamp fault messages. The author had identified the addresses for each lamp. This fits one of the fixes above - selectively disable a lamp fault in the BCM firmware.

It's not ideal because you have no led lamp failure monitoring. The perfect step would have been to identify and change the current sense voltages values for headlamps and change a few resistors for smalled leds replacing bulbs. I could play with my spare BCM unit and try changing their resistors detecting open load, but it seems headlamp bulb conversions are what everybody wants to do and that's the hardest perfect solution to find. High power headlamp bulb monitoring is only 2 X 2 channel SSD chips and 4 x current sense outputs. A hardware mod. wouldn't be too difficult. E.g A for a lower current 25 Watt led to be seen as running at the same power as a 55W lamp, the current voltage output sensed from these chips would need to be twice its present value. This should be doable with an experimental on board hardware hack using a SOIC quad opamp level changer on a small surface mounted board involving no changes to the BCM firmware code, but it requires on board micro surgery. For now, most will have to put up with generic claims of CANbus compatibility and large hot load resistors. Or go to HID with adaptation already available and deal with any lesser problems of aftermarket HIDs tripping lamp overload, because they may not incorporate soft start in their design.
As always, excellent write up and detailed.

I might add for the less technical of the readers;

If your going to add LED's make sure you order LED's that specifically says CANBUS compatible. These types of LED's will not trigger a bulb out fault. I have replaced all my halogen bulbs with LED's.

As Vox mentioned, if you use a LED to replace the headlights or fogs, they also should be CANBUS compatible, if not, you would need to wire in some appropriate resistors so as to not trigger a bulb out condition. Again, I have wired in resistors for my tail light donut modification and for the LED's I am running in my fogs. As for my fogs, it would of saved me some time if I would of purchased the correct LED's.

Now as to HID's, you still need to purchase a HID conversion kit which is CANBUS compatible. You can not change the adaption settings via VCDS to Xenon. There are several factory adaption Xenon settings; Xenon with shutter - This will disable your existing halogen high beams AND you will get a MFD fault "NO AFS Found". The second xenon setting - "Xenon without shutter" will allow you to use your existing halogen high beams, but again, since the Xenon setting expects to see other computer modules for the AFS, you will get a MFD fault message "AFS NOT FOUND". This I have tested, but since my HID kit has a inline CANBUS box, there was no way for me to test if these Xenon settings actually disabled the bulb out fault in the dash.

A 35 watt aftermarket HID with CANBUS works best and trouble free without making any adaption changes in VCDS. These HID systems work incredibly well in our halogen projector systems, with a nice clean cut off and substantially higher light output. No worries about blinding on coming traffic. HID bulbs are a tight fit and require a H7 adapter ring, so they will twist lock into our headlights.

LED's for the headlights are problematic, they don't provide the same light output as HID's, some are made poorly and don't replicate the light pattern of the halogen bulb, other LED manufactures claim their LED's do replicate the halogen light pattern. LED's also are very large and most have a fan and heatsink and due to the size, will not fit inside the pre-facelift headlight buckets and allow the rear door to be installed. You would either have to leave the door off, or cut a hole in it to allow for the added depth of the LED.
 

·
Registered
Joined
·
8,486 Posts
Discussion Starter · #3 ·
I agree with advice on HIDS. I have 4 35W sets I took off a mobike - single globes not Bi-Xenons. The ballasts are OE spec. , Osram branded and have a lot more protection in them than the cheapo Chinese one's. The problem with HIDs is they aren't so easy to de-install and re-install if you fall foul of the annual testers. But I might try them out. On the face of it leds should be easier to swap back to tungsten, but the removal of heat and lamp detection issues puts me off. That's the advantage of 35 Watt HID, they run cooler than 55W halogen, the small area arc akin to tungsten is in the right place for the beam pattern and they are adequately air cooled in the headlamp housing.
 

·
Registered
Joined
·
44 Posts
Just to add to the above - having something say "CANbus compatible" doesn't necessarily make it so. Look for the fine print. "May not work in some European Vehicles" or "May require additional decoders, available separately".

I've done a fair bit of trial and error with various globes in different lighting sockets throughout the car. Some with success, others not, and some only after a lot of additional stuffing around. So, buyer beware and if it doesn't work, speak with the seller.

As to the aftermarket HID kits, there are also some smarter alternative kits now being offered. Much more "plug and play" than older generation units, some kits now have globes complete with H7 bases (so no adapter rings) and have been designed to fit in factory enclosures. So too, the method of connecting the kit allows for the use of the factory plug, albeit plugging in to a wiring harness - to ballast - back through harness, and connect to globe. However, these kits allow for immediate reversal and, if you keep a halogen globe set in the boot (or truck as the case may be :) ), should you suffer a failure with one of your new HIDs you can pop it out, unplug, fit the halogen, plug it in, and be on your way.

Example:

35W Kit: H7 JTX HID Kit 35W 12V 24V XENON | eBay
55W Kit: H7 JTX HID Kit 55W 12V 24V XENON | eBay

There's also a 70W Kit but not recommended for polycarbonate lenses. For the completely over the top, there's a 150W Kit but really only for specific offroad applications.

Oz
 

·
Registered
Joined
·
8,486 Posts
Discussion Starter · #5 ·
The semantic sales descriptions used can be misleading as cars do bulb failure monitoring in different ways. If you take an analogue globe tungsten or led with 2 wires and use its characteristics to measure if its working, or working correctly, I've shown that measurement is done before any CANbus gets the answer and gives the result. You either do something at the bulb end that doesn't involve CANbus or you block the CANbus messages after they have been given.

I tried to understand how a globe is characterised looking for a solution that monitored a replacement led and gave the correct warnings for failure, rather than not have them at all which is what a 'CANbus decoder' might do. I concluded that Tungsten/led compatibility can only be done via programming or a hardware change in the BCM module where the lamp is first measured and before CANbus gets the result. If you think your globe is CANbus compatible, prove it by removing one or 2 leds in its cluster to simulate a failure? You should get a cold lamp out warning at ignition on and in some cases a hot lamp fail, both of which store fault codes. Many doing lamp mods treat lamp fail warnings as a nuisance, but like V.W I think they are worth keeping.

For all the advertising hype, if you replace tungsten globes with leds you can expect system warnings and if you don't see them, your lamp warnings have been disabled so you might as well remove the dash bulb or stick black tape over it!

All Xenon arcs emit a certain amount of U.V light and the globe is designed with a U.V filter built in the glass. U.V is a killer for polycarbonate. When you buy a cheap Xenon globe kit you don't know the quality of the U.V filter or even if it has one. The higher the power the more U.V emitted.
 

·
Registered
Joined
·
8,486 Posts
Discussion Starter · #7 · (Edited)
what kind of bulbs do I need to buy for my 2012 Eos? :)
If you want bulbs that are fairly reliable and guaranteed to work correctly in V.Ws system, you will buy the tungsten (or Xenon bulbs if you have them) which V.W put in your car design when it was sold and listed for your car model by parts sellers.

If you want to change the bulbs for something else which V.W doesn't specify or recommend, then you can read the results others have had, take your chances and hope what worked for them will work for you. But they won't refund your money if they don't meet your expectations.:)
 

·
Registered
2008 Volkswagen VR6 Eos
Joined
·
749 Posts

·
Registered
Joined
·
8,486 Posts
Discussion Starter · #9 ·
They are all teeny weeny torpedo style lamps. Most aren't monitored for failure in our system and those that are only need a small load balance resistor. You don't see the larger 21W lamps. But if you want to change your interior color they would look cool.
 

·
Registered
2008 Volkswagen VR6 Eos
Joined
·
749 Posts
They are all teeny weeny torpedo style lamps. Most aren't monitored for failure in our system and those that are only need a small load balance resistor. You don't see the larger 21W lamps. But if you want to change your interior color they would look cool.
These are a pure white and substantially brighter than stock. I run white, but I believe blue is an option
 
1 - 10 of 10 Posts
Top