Subaru Outback Forums banner

41 - 60 of 66 Posts

·
Registered
Joined
·
16,313 Posts
@johnre @DavidPeab
Picked up two original Subaru 3rd Gen. flashers at the pick-and-pull yard. Plan to look "more closely" at one, and use the second in a mock-up circuit to see what happens to the clicking/flashing rate, and the ability to deliver current, when the load is varied.
 

·
Registered
Joined
·
16,313 Posts
Wow! Thanks. Anxiously awaiting your results.
Working on it . . . but haven't yet been able to reproduce the weak/fading flashing symptom.

Question: When the turn signal problem first appeared, or during the trip, did you check whether or not the parking lights are working correctly, especially the right front amber clearance (i.e., parking) light, when the headlight switch is at the first or second On position? If not, perhaps Anna can do this, comparing the right and left sides at the front and back. (Maybe a video or photos so we can see as well.)
 

·
Registered
Joined
·
4,078 Posts
Discussion Starter #44
She’s making more videos to post. I’ll ask her to read your request. Although, I seems to recall the markers are illuminated by the headlights (amber lens overlay) and rear markers wrap from the tail lights.
 

·
Registered
Joined
·
16,313 Posts
This is an update on what I've found so far in bench testing the flasher. As noted above, I have not yet reproduced the fade out symptom, nor do the results point more specifically to the fault. Ideas/suggestions for further experimenting welcomed.

Note: the Turn Signal & Hazard Module (aka flasher) being used is the factory version from 2005-9 Subaru Legacys or Outbacks. Details: Subaru part number 86111AG010; Mfr: Denso. The rating on the cover is visible in the attached photos; basically on each side (right, left), 54 Watts (approx 4.2 Amps).

I'm bench testing using a 13.1 V regulated power supply, various power resistors to simulate different bulb wattage ratings and numbers of bulbs, a 3247A front turn signal bulb to provide visual confirmation of the flashing, and an oscilloscope to monitor the signal to the loads.

The flasher has two relays (photo below), one for each turn signal side, and an SE336 IC chip (for which I have not found any details/datasheet).

The relay switch contacts connect the main power coming into the flasher to their respective outputs going to the right and left side turn signal circuits. There's no device apparent between the supply and relay outputs that could limit the current, such as to cause the fade out.

The 3rd gen. turn signals consist of a 27 W bulb filament at the front, and a 21 or 23 W bulb filament at the rear. There's also the LEDs in the side mirror and the LED for the green arrow in the instrument panel, however these normally draw less than 50 ma, and are usually wired with a current limiter (e.g. resistor) that would prevent excessive current flow even if the LED were to short.

When operating as turn signals (left or right), with a load equivalent to the two bulbs used in the 3rd Gen on each side, the flasher IC switches the corresponding relay from off-to-on and on-to-off 85 time per minute. (This is the clicking we hear. It's also the spec'd rate according to an imprint on the cover.)

When either bulb on one side is disconnected (as if burned out), the flasher cycling rate changes to about 190 switches per minute. I found that with a load of about 3 Amps (more than either bulb alone) the flashing rate is at 85; when the load is 2.5 A or less, the flashing rate is 190. The flasher is obviously able to detect the load and switch to the higher rate when one of the two bulbs is burned out. I have yet to figure out how this works. (Incidentally, this suggests that for the flasher to indicate when a bulb is burned out, it's important to use bulbs that meet the original rating.)

When I increased the load to over 8 A (equivalent to doubling the number of bulbs on each side) the switching rate remained at 85 cycles per minute. So the flasher does not limit the current or cause it to "fade out" in the face of an excess load.

When operating in Hazard mode, there's a distinct difference. The 85 cycles per minute switching remains constant regardless of the number of bulbs; even with no bulbs connected!
 

Attachments

·
Registered
Joined
·
16,313 Posts
She’s making more videos to post. I’ll ask her to read your request.
Great. (Thanks @AnnaPeab ) We're all trying to help from a distance, and whatever we info "gets us closer" can only help.
Although, I seems to recall the markers are illuminated by the headlights (amber lens overlay) . .
Am I misunderstanding "markers are illuminated by the headlights"? Sounds as if the headlight bulb has to be on for the amber marker to be lit. If that's what's meant, then not so. The amber marker toward the outside of the headlight assembly has a bulb behind it. That bulb is independent of the headlight.
 

·
Registered
Joined
·
4,078 Posts
Discussion Starter #47
I’m waiting for Anna to send a pic of the markers. But, I do recall worrying about them before taking our cross country trip and being relieved that they were working.

Anna?
 

·
Registered
Joined
·
4 Posts
She’s making more videos to post. I’ll ask her to read your request. Although, I seems to recall the markers are illuminated by the headlights (amber lens overlay) and rear markers wrap from the tail lights.
Sorry for the delay and thanks to everyone for your help!

Here are two new video links:
1. Car On Hazards/Turn Signals Test
2. Car Off Hazards/Turn Signals Test

In the one with the Hazards from the front view, it looks like there's light on the right side but I think it's just the reflection from the left.

I'll be checking the sidelights that my dad has been talking about when it gets dark here within the next few hours.
 

·
Registered
Joined
·
16,313 Posts
I'll be checking the sidelights that my dad has been talking about when it gets dark here within the next few hours.
The videos are telling. Will get back to this.

Anna, for the sidelights, there's a rocker switch on the top of the steering column, between the steering wheel and the instrument panel. With the ignition Off (key out), and nothing else switched on, flip the rocker switch, and then go to the right front corner and check for any sign of light in behind the amber part of the headlight assembly.

After checking light at the right front, check the left front for the same. Then at the back, in each corner, in the red section of the light housings there should be two lights, one is facing the rear, the other more on the side. (Photos in post #51 below.)

Let us know what you find, or if you can, video around the car.

After this test be sure to flip the rocker switch back off, as otherwise the lights stay on and that will drain the battery.
 

·
Registered
Joined
·
16,313 Posts
Sorry for the delay and thanks to everyone for your help!

Here are two new video links:
1. Car On Hazards/Turn Signals Test
2. Car Off Hazards/Turn Signals Test

In the one with the Hazards from the front view, it looks like there's light on the right side but I think it's just the reflection from the left.
No problem, and still helpful.

In the videos, it's now clear (to me at least) that when the right side should be flashing there is a dim burst in the right side green arrow, and, possibly in the outside LED of the right side mirror. It's a brief burst that is gone by the time the flasher clicks off. That is clearly a fade away! Moreover, the flasher is clicking at the normal rate which suggests the problem is more likely a short of some sort rather than a bad connection (because, as in my testing, if there's a bad connection, the current is likely to be low and the flasher would switch to the faster rate of clicking.) Of course, this is presuming that the flasher module that's in there now has this same capability.

Let's see what there is at the four corners with the rocker switch set to on.
 

·
Registered
Joined
·
16,313 Posts
I'll be checking the sidelights that my dad has been talking about when it gets dark here within the next few hours.
When it's dark, if the rocker switch (Subaru calls it the Parking Lights switch) is turned on, this is what there should be:

Right front. The headlight is off, and there's only the one amber light, at the outer side/corner. There should be the same at the left front. (These are "clearance lights" in Subaru-speak.)

IMG_4231.JPG

Right rear. Here there's two bulbs, one in the larger red section, and one in the smaller, narrower red strip along the side. Same on the left rear corner. (These are "tail lights" in Subaru-speak.)
IMG_4230.JPG
If at the right front there isn't a bright light, as bright as the left front, check closely inside to see if the light bulb might be glowing dimly.
 

·
Registered
Joined
·
4,078 Posts
Discussion Starter #52
I FaceTimed Anna and both clearance lights (DSF and PSF) are working. Both tail lights (incl the "side "marker" DSR and PSR) are working with the virgin switch on and also when the headlights are on. The big question: Do the clearance light and the turn signal/hazards/remote loch-unlock indicator all use the same bulb? Could it be as simple as a short or burned out second filament in a two filament bulb? The shop manual shows the singal bulb listed as: Front turn signal, parking, front side marker light and the illustration looks like there is only one bulb socket on the side of the assembly.

The front right side marker works. It seems to use the same bulb as the turn/hazards. The turn/hazards on the right work intermittently, dimmly - even the mirror flashes dimly. There must be a bad or shorted filament. I say she replaces the front dual filament amber bulb (the one easy thing I did not do - wasn't in the position to remove the wheel well cover when I was on the trip). If it's not that, then I believe it is the BIU.

Taking bets.
 

·
Registered
Joined
·
341 Posts
The 2005 FSM lists the Front Turn Signal, Parking, Front Side Marker Light as 12 V — 27/8 W. The dual power rating suggests to me that there's two filaments in there.

The FSM also says
3.Parking Light Bulb
A: SPECIFICATION
The parking light bulb is integrated into front turn
signal light bulb as a unit
and
4.Front Side Marker Light Bulb
A: SPECIFICATION
The front marker light bulb is integrated into front
turn signal light bulb as a unit
The bulb shown at https://parts.subaru.com/p/Subaru_2008_Outback-30L-5AT-LLBean/BULB/49295585/84920AG030.html clearly has two filaments.

So yes, I think you're on the right track.
 

·
Registered
Joined
·
16,313 Posts
The 2005 FSM lists the Front Turn Signal . . .

OTOH, the non OEM replacements I can find online quickly are single filament.
It's a 2008, but essentially the same.

As noted earlier in this thread, the bulb is a 3457A, 27/8 W dual filament. (Page 12-9 of the 2008 Owners Manual.)

Do the clearance light and the turn signal/hazards/remote loch-unlock indicator all use the same bulb? Could it be as simple as a short or burned out second filament in a two filament bulb?
Yes, it certainly could be a short in the bulb.

At the front, in each headlight assembly, there's three bulbs; low beam headlight, high beam headlight and the clearance/turn signal bulb.

The latter is the dual filament bulb that was mentioned earlier in this thread. Here's a photo of one of those bulbs. a 3457A, that I was using in my bench tests. (Slipped off the bench -- lost part of the glass bulb, but provides a good, close look at the filament structure!)
IMG_4232.JPG

One of the filaments, the lower one (closer to the black plastic base), is the 27 W used for the turn signal as well as the Hazards and lock/unlock verification. The other filament, 8 W, is used for the "clearance" (aka parking/marker light) function.

In dual filament bulbs, it's not unknown for one of the filaments to break and contact the other, or short between the two supports. Often, that affects both functions -- the filament that is broken, and the other filament that isn't. That's why I wanted to find out if the right front and left front clearance lights are equally bright. Unfortunately, it seems that's not an indicator in this case.

I say she replaces the front dual filament amber bulb (the one easy thing I did not do - wasn't in the position to remove the wheel well cover when I was on the trip). If it's not that, then I believe it is the BIU.
Certainly, it can be replaced to get that out of the way. It's been suspect all along. (It's a bit tough to get to but it sounds as if you've been there before -- via the wheel well. Is this something Anna might be able to do?)

If that doesn't work, it doesn't mean the BUI becomes the prime suspect, unless the flasher itself is defective as well. (Earlier I understand you had replaced the current tap-to-pass flasher with a regular, generic unit but the symptoms were exactly the same. Is that correct?) There could be a short (intermittent, and/or "near short") to ground in the wiring heading out from terminal #2 of the flasher module.

The wiring can be checked for a short with a multimeter if the two turn signal bulbs (right front, right rear) are removed. With the bulbs out, the right side turn signal wiring goes to the sockets, the LEDs in the mirror, the LED for the right green arrow in the instrument panel, and to terminal #2 of the flasher. When the flasher is Off, that terminal shouldn't connect to anything.

With that, at the rear turn signal socket there's two wires going to two contacts in the socket. One of the wires goes to a ground. The other wire is the connection back to the flasher mentioned above. When measuring the resistance from each of the two contacts to a good ground, one should show very low resistance ( not more than an Ohm or so), while the other should show a higher resistance (probably in the hundreds of Ohms if not more.) If the resistance at both is low, that would confirm a short or near short that is loading down the right side turn signal circuit. If that's the case, next steps would be to remove the flasher (in case it's defective), and if that doesn't work, isolate parts of the wiring by separating connectors in the circuit. But let's see where changing that 3457A front bulb takes it.
 

·
Registered
Joined
·
4,201 Posts
The flasher is obviously able to detect the load and switch to the higher rate when one of the two bulbs is burned out. I have yet to figure out how this works.

When I increased the load to over 8 A (equivalent to doubling the number of bulbs on each side) the switching rate remained at 85 cycles per minute. So the flasher does not limit the current or cause it to "fade out" in the face of an excess load.
This part of it I'm able to address - per this image you provided of the board, the SE336 IC is clearly a high side current sensor and amplifier device, capable of detecting the lighting circuit current loads on each side of the car to some reasonable degree of resolution. You can see both the shunt resistors here (the large cylindrical devices with the color bands), and also the 10k Ohm pickoff resistors to the IC (the SMT parts with the "103" letters on the top). And I'm guessing it's three channel, given the three 10k Ohm resistors.

470738


Here's a good app note on a similar device from Analog Devices:

I'm wondering about its output. With 10 pins total on this mystery part, it might be that there is a low threshold comparator circuit as well inside this IC, which could alarm if the current sensed is too low using the three pins in the middle on the top. But there's not enough pins to also set up a high threshold comparator circuit, and get this information sent out as individual signals, which you of course confirmed with your measurements. And while they could have set up a digital bus through the three pins instead to get more information out, I doubt it. This is probably a single threshold deice.

But this setup means that there is always a finite resistance between the +12V supply and the loads - obviously not a lot as during your testing you managed to get 8 A out of this device by varying the loads. It could be that the resistor for the right side turn signal lighting is flaky and has a thermal characteristic that causes it to open circuit once a heavy load is drawn, but of course they're replaced it once and the problem stayed with the new one. And what is cause and effect - would the resistor flakiness be just a symptom here?

Too many more questions to be answered before David and Anna can get to the bottom of it, I'm afraid. But I at least addressed the question you raised above about how it was done.
 

·
Registered
Joined
·
16,313 Posts
@johnre

We're working in parallel on this and much appreciated.

You're right in identifying the SE336 as a current sensing device, which also controls the relays.

In searching for info on the SE336 I came across this paper on the Mouser Electronics site; it's not for the same IC, but it describes how a flasher unit senses the current to the bulbs, and can detect when one or more is out. In the schematic on page 2, the shunt resistor, R3, is only 18 milliOhms; a virtual short in most situations, but not this one!

sample flasher.jpg

I spent some time tracing out the circuit of the Subaru flasher, and then happened to find a schematic on-line that looked similar. So I used it as a base, and retraced the circuit board adding in values and identifiers on both the schematic and a photo of the board. (I took apart the second flasher unit so that the board is fully exposed. Note, the joint at R8 looks messy; I had lifted one end of R8 to locate a connection, but hadn't resoldered it when the photo was taken. Also, the dimples in the solder joints are from test lead tips -- there's a thin conformal coating on the board.)

Turn Signal & Hazard Module schematic.jpg Flasher pcb layout.jpg

In this case, the "shunt" resistor is 0.05 Ohms (measured with a low-R meter). It can be seen in the third photo attached to post #45 above. It's a rectangular wire standing above the board on the relay side.

It looks as if the SE336 is monitoring the total current to the bulbs passing through the shunt. Because this circuit uses separate relays, when in turn signal mode, only one relay would be activated, so the IC can distinguish between the current to two bulbs (plus the LEDs) or only one bulb. In this regard, Section 4 of the Mouser paper clarifies the role of the very low resistance shunt in determining the flash rate.

When in (4-way) Hazard/BIU mode, the SE336 apparently doesn't use the comparator to monitor current; rather, it simply switches both relays on and off at the standard rate without considering the number of bulbs.

Interestingly, paragraph 4.6 of the Mouser paper makes reference to the shunt plus the harness wiring limiting current (and therefore the voltage across the shunt) in the case of shorted bulbs. That too might explain why the flash rate doesn't change when additional bulbs are connected or, in the current case, when there might be a short downstream in the circuit.

Your "a high side current sensor and amplifier device, capable of detecting the lighting circuit current loads on each side of the car to some reasonable degree of resolution." combined with the actual schematic and external explanations of these types of circuits, seems to answer the question of how the flasher detects a missing bulb.

At this point I think there is a short (bulb, or wiring) on the right side. We know from the latest videos that the relay is switching, and that the LEDs light up when the relay audibly clicks on, but quickly fade off before the next click indicating the relay has switched off. That could be due to combination of the location of the short relative to the location of the LED connections, and the resistance of the wiring from the supply fuse (15 A).
 

·
Registered
Joined
·
4,201 Posts
Spot on with all of your observations, I think. I stopped on Mouser's last sentence in 4.6 as well ("The combination of shunt resistor and resistance of wire harness prevents pin 7 from a too high voltage in case of shorted lamps" and thought this through; it appears to be pretty much wishful thinking, depending on harness and fusing, but it may well be applicable to David and Anna's issue.

Now how does David use all of this information to debug the root cause of the problem? Not really sure.
 

·
Registered
Joined
·
16,313 Posts
Now how does David use all of this information to debug the root cause of the problem? Not really sure.
At this point the understanding of the system, and the symptoms, are consistent with a short or near short. As the bulbs can be a problem, especially the dual filaments, pulling them is the first thing. @DavidPeab is already on this, but doesn't have access to the car. This will depend on what can be done where the car is now located (Las Vegas).

With the right side front and rear turn signal bulbs removed, or one, and then if necessary the other as well, see if the LEDs (green arrow and mirror) light up properly in time with the flasher. (Or as mentioned above, measure for a short to ground on the "high side" wire.) As the bulbs are out, and if there's no short remaining, the flasher, operating in right turn mode, will tend to run at the higher speed, but the LEDs should still flash clearly. If there's a short in the system, it will operate at the standard speed, and the right side LEDs might only light very briefly and dimly, as is the case now.

[The front, dual filament bulb isn't easy to access in this generation. The FSM calls for the headlight assembly to be removed, and that means first taking off the bumper cover. It's possible to access it from the wheel well after moving the front part of the wheel well liner out of the way. Also, having looked at it just now in my 07, I believe it can also be reached, perhaps more easily, from the engine compartment once the front half of the engine air filter housing is removed.

The rear bulb requires the rear combination light housing to be removed to get to the sockets, but that's a few screws and a bit of prying to allow the housing to slide out of the fender. (It's shown in the Owners Manual.)]
 

·
Registered
Joined
·
16,313 Posts
We made it to Las Vegas with the obligatory stops at the Grand Canyon and such. The passenger side amber lights worked throughout the last 4 days of the 9 day trip, so I assumed the problem was solved. Unfortunately, my daughter told me it is creeping back intermittently.
The symptom in this case is unusual -- I don't recall ever working on a turn signal problem that involved the flasher turning on and off ("clicking") at the normal rate, while the bulbs barely lit up, if at all.

Given what we now know about the flasher, especially the different effect of a short vs. a bulb being burned out, I'm wondering if, when the turn signals appeared to be working during those last 4 days of the trip, the flashing/clicking was at the faster rate.

Were you actually checking the right side turn signal lights outside to see if all three were flashing?

[The goal all along is to fix a problem. But I also like to figure out what's was wrong. In some cases it could verify the cause before a part is changed, and in others it could add to a knowledge base that could be helpful in future. With this in mind, I looked more closely at the way the front dual filament bulb is oriented when installed in the headlight assembly. In my 07, and probably all the 2005-9, the bulb filaments are vertical, as in the photos below. With this positioning, it's possible that if the turn signal filament were to break, it could fall down onto the grounded support of the other filament, shorting the right side turn signal circuit to ground. (See second photo below.) That could lead to the flasher clicking at the normal rate, but little or no sign of light at any right side bulb.

If that were the case, because the broken filament is just resting on the grounded support, at times it might not make good contact. The result would be the same as a burned out bulb; the right side turn signals would appear to work, i.e., the green arrow inside would flash clearly, and there would be noticeable flashing lights on the outside. But in fact the right front bulb would not be flashing and the flasher clicking would be at the higher rate, as observed in the bench tests.]

bulb in place.JPG bulb orientation and layout.JPG

In the first photo above, we can see the two filaments (in this case, intact). When a broken filament is suspect, it might be possible to see that by looking at the bulb this same way.
 

·
Registered
Joined
·
4,078 Posts
Discussion Starter #60
Anna is going to Pep Boys to get the bulb replaced tomorrow. Of course, we’ll let you know if it fixes it. I told her to take “before” pictures of the wheel well to be sure they reassembled it correctly.
 
41 - 60 of 66 Posts
About this Discussion
65 Replies
7 Participants
plain OM
Subaru Outback Forums
Welcome to the Subaru Outback Owners Forum, we have tons of information about your Subaru Outback, from a Subaru Outback Wiki to customer reviews.
Full Forum Listing
Top