Come up with a brilliant solution .......

[Franc Zabkar]

Maybe I am missing something here - it seems simple. Power transistor thru
resister from active side of tail lamp supply to earth - resistor set so
required current is drawn when transistor is biased on - transister base fed
thru zener so conducts when pulse arrives. Pulse sent, transistor conducts,
computer thinks happy thoughts.

David - or is this just too easy (and cheap - $5 worth of bits at the most)

Maybe I'm missing something, but wouldn't your transistor be on
whenever the lights are on?

-- Franc Zabkar

Please remove one 'i' from my address when replying by email.

 

[Franc Zabkar]

As far as we've been able to check this mess, yes, the current pulse is sent
when the ignition is switched on, and will keep on pulsing until the load
sense is satisfied. Vehicle diagnostics are performed (obviously, the lights
are only one of the tests) and any errors are indicated on the dashboard.
However - it would appear that the system has a continuous checking routine
to determine whether a lamp has failed during the run.

How continuous? Does the computer test the lamp immediately after
switching it on and then ignore the current reading until next time?
Or does it keep testing the current during the entire time the lamp is
on?

-- Franc Zabkar

Please remove one 'i' from my address when replying by email.

 

[Franc Zabkar]

I don't think there is any legitimate way around this, for the
following reason.

The whole idea of the computer check for blown lamps is to alert the
operator to a safety issue.

If you use ANY system which "fudges" the computer into thinking that
the led tail-lights are in fact incandescants (such as dummy loading,
transistor current loads etc) then what happens if somebody actually
disconnects the led lamps (try it out using a parallel incandescant
bulb)? With the dummy load in place the computer will think the led
tail lights are connected when in fact they are not. Such work-arounds
are potentially hazardous and should not be contemplated. Either the
manufacturer of the truck computer modifies his program to cater for
led lamps or you should stick to using incandescant bulbs.

Ross

I had considered the solution proposed by Craig Hart but refrained
from suggesting it because the requirement was for a "simple" method.
Anyway, if this level of complexity is allowed, then one could easily
satisfy the safety requirement by placing a high side current sensor
within the trailer itself.

PNP transistor
E C
+V o--|---\___/-----> to RC coupled, dummy load switching transistor
| |
| R
| |
|--R--|------> to LED lamp array


-- Franc Zabkar

Please remove one 'i' from my address when replying by email.

 

[Franc Zabkar]

I work in the trailer manufacturing business.
We use LED tail-light clusters.
The modern truck/tractors utilise a computer system to perform diagnostics
on various systems within the rig. One of these is the lighting system.
A pulse is sent to each tail light, the system monitoring the current. If
there is current flow, it is assumed the incandescent globes are OK.

Unfortunately, the LED lamps draw such a small amount of current that the
computer does not see a "filament" and flags a major fault on the truck
dashboard. In addition to the error message, the system continues to send
curent pulss in the forlorn hope that things at the rear of the rig have
improved. This causes the entire suite of LED lamps to flash like a
low-class disco!

There are ways around it. Some manufacturers have placed incandescent globes
in parallel with the LED lamps, others have used high wattage resistors.
Neither is acceptable for obvious reasons.

There must be a way to "tell" the computer that the LED lamps are fine by
emulating the current drawn by an incandescent globe, without using the
solutions noted above.

Do any of you outstandingly brilliant contributors have any novel and
innovative ideas?????

Maybe the idea requires a different approach. Just how smart is the
computer? Can you fool it into believing that the trailer is not
there? Will the computer continue to test the trailer in this case? Is
there a "trailer present" sense wire in the harness to the trailer? If
not, then how does the computer distinguish between the case where the
trailer is not present and the unlikely case where *all* lamps have
failed? Does the computer understand the difference between no current
draw and insufficient current draw?

If the trailer isn't there, will the lamps still work, ie does lamp
voltage appear at the connector? Does the computer sense the brake
switch and then supply power to a lamp, or can the brake switch
control the lamp independently of the computer? What happens if the
computer fails? Do the lights still operate? This may be a silly
question, but do the lights behind the cabin still operate when the
trailer is attached? If not, then this would indicate that the lights
are under computer control and that the trailer's presence is being
sensed.

-- Franc Zabkar

Please remove one 'i' from my address when replying by email.

 

[quietguy]

Since the computer sends a pulse to check the lights this would have to be
higher than the normal 'alight' voltage of the taillights etc. So you would set
the zener going to the base of the transistor a little higher than the 'alight'
voltage - eg the tranister only turns on when the pulse is sent

David

 

[Ken Taylor]

message

Since the computer sends a pulse to check the lights this would have to be
higher than the normal 'alight' voltage of the taillights etc.

How do you know this? And why do you think it has to be so?

Ken

 

[John G]

Maybe the idea requires a different approach. Just how smart is the
computer? Can you fool it into believing that the trailer is not
there? Will the computer continue to test the trailer in this case? Is
there a "trailer present" sense wire in the harness to the trailer? If
not, then how does the computer distinguish between the case where the
trailer is not present and the unlikely case where *all* lamps have
failed? Does the computer understand the difference between no current
draw and insufficient current draw?

If the trailer isn't there, will the lamps still work, ie does lamp
voltage appear at the connector? Does the computer sense the brake
switch and then supply power to a lamp, or can the brake switch
control the lamp independently of the computer? What happens if the
computer fails? Do the lights still operate? This may be a silly
question, but do the lights behind the cabin still operate when the
trailer is attached? If not, then this would indicate that the lights
are under computer control and that the trailer's presence is being
sensed.

-- Franc Zabkar

All these suggestions are a load of rubbish...
The ONLY way to fix this problem is have the manufacturer fix the
computer to correctly handle modern (LED) lamps.

If he does not want to comply then don't buy his product.

If it is a design rule requirement to do the test then he has to comply
or the OP has to stay with the original candles (incandecescent lamps)
that have to be tested because they fail often on trailers.

 

[Ross Herbert]

I had considered the solution proposed by Craig Hart but refrained
from suggesting it because the requirement was for a "simple" method.
Anyway, if this level of complexity is allowed, then one could easily
satisfy the safety requirement by placing a high side current sensor
within the trailer itself.

PNP transistor
E C
+V o--|---\___/-----> to RC coupled, dummy load switching transistor
| |
| R
| |
|--R--|------> to LED lamp array


-- Franc Zabkar

Franc (and Ken), both suggestions will work, BUT.... the "fudge"
components would have to be integrated as part of the LED lamp
assembly in order to conform to the relevant standards, I would
imagine. Not only that, if the LED lamp did go open circuit the
"fudge" components must not allow the computer to test the dummy load
and think that the LED lamp was still ok. As I see it,any sort of
dummy load would still allow a successful lamp test when the LED lamp
itself is open circuit. It would have to be a "fail safe" arrangement
to serve the intended purpose of the computer lamp test. I can't see
how it can be done using these methods while still providing a
meaningful test result.

Considering the reliability of incandescant bulbs I would stick to
using these in the trailer lamps. This way the automatic lamp test
will provide a meaningful test result as intended so that faulty lamps
can be replaced. After all, it is not as though a few extra watts is
going to seriously affect the alternator loading or battery capacity
on a truck.

Ross Herbert

 

[Ross Herbert]

All these suggestions are a load of rubbish...
The ONLY way to fix this problem is have the manufacturer fix the
computer to correctly handle modern (LED) lamps.

If he does not want to comply then don't buy his product.

If it is a design rule requirement to do the test then he has to comply
or the OP has to stay with the original candles (incandecescent lamps)
that have to be tested because they fail often on trailers.

I entirely agree......

Ross Herbert

 

[Ken Taylor]

Franc (and Ken), both suggestions will work, BUT.... the "fudge"
components would have to be integrated as part of the LED lamp
assembly in order to conform to the relevant standards, I would
imagine. Not only that, if the LED lamp did go open circuit the
"fudge" components must not allow the computer to test the dummy load
and think that the LED lamp was still ok. As I see it,any sort of
dummy load would still allow a successful lamp test when the LED lamp
itself is open circuit. It would have to be a "fail safe" arrangement
to serve the intended purpose of the computer lamp test. I can't see
how it can be done using these methods while still providing a
meaningful test result.

Considering the reliability of incandescant bulbs I would stick to
using these in the trailer lamps. This way the automatic lamp test
will provide a meaningful test result as intended so that faulty lamps
can be replaced. After all, it is not as though a few extra watts is
going to seriously affect the alternator loading or battery capacity
on a truck.

Ross Herbert

I agree - if the truck manufacturer isn't going to do a fix for their
firmware to accommodate LED lamps (and it probably isn't trivial, as you'd
have to somehow 'know' which lamps were LED and which incandescent) then any
fix is either potentially dangerous or self-defeating (ie. the globe in
parallel sorta kinda defeats the purpose of having an LED pack in the first
place).

Cheers.

Ken

 

[MC]

I had the same idea but declined to suggest it because the OP
specified a "simple" solution. I was thinking of something like this:


|---------------------|
| dummy R R = 50 ohm ???
| |
| Z |/ Z = 6V zener
o--|--C--|--|<|--|---R--| Q npn
| | | |\
low R D hi R | D = diode (grounded anode)
|_____|_______|__ ____|
_|_
=

-- Franc Zabkar

Please remove one 'i' from my address when replying by email.

How about

npn
C E
+V o---------------|---\___/-----> to to LED array
| | |
/ E | |
--| pnp |--R1-|
| \ C |
R2 | |
| RD C1
C2 | |
| | |
o----------------|

RD emulates the incandescent-lamp load.
Select R2 & C2 time-constant to satisfy the duration
constaint of the 'lamp-test' and then to disconnect
that power-hungry load.

Select R1 & C1 time-constant to eliminate the
'light-test' pulses from flashing the leds at
night and dusk.

Seems that you'll need to determine two things ;
1. what the pulse-duration of the lamp-test
actually is. (I suspect it'll be milliseconds
if the pulses are visible during daylight)
and
2. what the 'lamp-ok' current threshold is.
(10 mA, 100 mA, ???? )
{A cold incandescent has a much lower resistance
than a hot operating bulb.}


If one needs to verify that the led-strings are ok
then that will complicate things a lot because ;
1. the leds have such a fast response that the
eye will detect the test-pulses at night/dusk,
and
2. a led in a series-string could be short OR open.

 

[two bob]

put finger to

keyboard and composed:


Maybe I'm missing something, but wouldn't your transistor be on
whenever the lights are on?

Whats your point? If there was a normal globe in cct, it would do the same
thing.

 

[Franc Zabkar]

put finger to

Whats your point? If there was a normal globe in cct, it would do the same
thing.

My interpretation is that the transistor is DC coupled. Therefore, if
it is on whenever a pulse is present, then it would also be on
whenever the lights are on, so why use a transistor at all? The
proposal would only make sense if the base drive were AC coupled.

-- Franc Zabkar

Please remove one 'i' from my address when replying by email.

 

[quietguy]

I don't know 'this' - just guessing. And since the computer measures current
draw, it seems unlikely it is an RF or Audio pulse, but rather a DC pulse. And
to superimpose a DC pulse onto a 24v DC line, it seemed logical that it would
have to be higher than 24v in order to be detected.

But of course I am just 'supposing' - is that a problem for you?

David

 

[quietguy]

I was suggesting DC coupling, but thru a zener so the transistor only conducts
when the zener v is exceeded, by the pulse.

David

PS don't take my idea too seriously, as I may be way off track - just an idea

 

[Ken Taylor]

It's not a problem to me, it's just plain wrong. It's not the way lamp
current sensing is done.

Ken

message

 

[The Real Andy]

I agree - if the truck manufacturer isn't going to do a fix for their
firmware to accommodate LED lamps (and it probably isn't trivial, as you'd
have to somehow 'know' which lamps were LED and which incandescent) then any
fix is either potentially dangerous or self-defeating (ie. the globe in
parallel sorta kinda defeats the purpose of having an LED pack in the first
place).

Cheers.

Ken

Easy, measure the voltage drop across the led array, if it exceeds
that of the led(s), no load is switched on. If it is as expected,
switch on the load for half a second.

V+ -----+
|
RES
|
+-------Vsense_led
|
LED
|
LED
|
GND

If the two leds have a foward voltage of 1.2V each, then Vsense_led
should be 2.4v. If the device fails, Vsense_led will be the same as
v+, providing that the input impedance of the measuring curcuit is
reasonble high.

I would also be cautious about the dirty signal that will probably be
found on the power line. This whole circuit will no doubt need a good
dose of filtering.

 

[The Real Andy]

Easy, measure the voltage drop across the led array, if it exceeds
that of the led(s), no load is switched on. If it is as expected,
switch on the load for half a second.

V+ -----+
|
RES
|
+-------Vsense_led
|
LED
|
LED
|
GND

If the two leds have a foward voltage of 1.2V each, then Vsense_led
should be 2.4v. If the device fails, Vsense_led will be the same as
v+, providing that the input impedance of the measuring curcuit is
reasonble high.

I would also be cautious about the dirty signal that will probably be
found on the power line. This whole circuit will no doubt need a good
dose of filtering.

After playing with a few circuits on paper, my conclusion is that you
will need 1 opamp or comparator with 2 transitors. It may be possible
to get away with one transistor.

Dare I say it, a 555 in monostable mode may even proved an elegant
solution.

All these parts can be had in smd packages now, so adding it to an
existing circuit should not consume much real estate.

Might have to download LT's spice and have a play.

 

[The Real Andy]

The Brake and Indicator lamps are 24V 21W, the Park lights are 5W

Is it possible to grab a trace of the pulse on a cro? Are you using
off the shelf inicator/brake led assemblies or manufacturing your own.
Can you provide a link to the manufacturer of the assemblies if you do
not construct them yourself

 

[quietguy]

Ken, since you say you have the inside knowledge here it would be helpful if
you could let us know how it is done, so we could be more hopeful to the OP

David