Double vision (ghosting) at night through windshield?

[vandacca]

Well, I'm not exactly sure what can be done other than slightly redesigning the vehicle, so the glass isn't at an angle that causes ghosting. I'm actually on my 3rd windshield and while the ghosting isn't as bad as the loaner X I was driving, it's still pretty noticeable.

It's comments like these that make me think that the problem is with Quality Control, rather than the design. If there are windshields in circulation with minimal ghosting, then I have to believe that this issue could be resolved with tighter QC.

 

[ohmman]

It's comments like these that make me think that the problem is with Quality Control, rather than the design. If there are windshields in circulation with minimal ghosting, then I have to believe that this issue could be resolved with tighter QC.

I continue to look across the windshields horizontally when I see others in parking lots, and the "bubbling" that I'd pointed out before seems to vary quite a bit. I'm not blaming that directly for the ghosting, but I think it does show that there's some variance in the manufacturing process.

 

[EV-lutioin]

Latest from my service center manager... he said he agrees that the ghosting is really bad in my car (he took it home for the weekend and drove it at night) and he said he begged Tesla engineering several times to help me out with this issue and Tesla told him that they are not going to replace the windshield because the windshield is within tolerance for secondary image separation based on the test they performed. This is so aggravating, I am not sure what to do now...

 

[Screwbal]

I can say when mine does show up I won't tolerate issues and if Tesla won't rectify things I'll raise hell through whatever channels get it fixed or draw the most attention. My current S is absolutely flawless and I didn't chose to change it for a more expensive X to wind up with a pile of *sugar*... We'll see how things look come March/April when my X gets delivered.

 

[rainman50]

This is the main reason why I will not buy a Model X. I have a Model S now and was thinking of getting an X for my wife.

 

[AudubonB]

Well in that case let's request a detailed evaluation by a independent third party testing agency of each one of our windshields that provides evidence that they precisely meet the regulations for secondary image separation listed below.

Federal Register :: Federal Motor Vehicle Safety Standards; Glazing Materials


10. Visual Distortion Paragraph S5.15 of ANSI Z26.1 requires glazing materials used as windshields to undergo visual distortion and optical distortion tests. The purpose of these tests is to ensure safe driver visibility. To conduct the visual distortion test, the sample is placed in front of a light source and a circle is projected through the test specimen onto a screen. The tester then records the separation between the primary and secondary image. The separation of the secondary and primary image is not allowed to exceed 3.95 minutes of arc or 8.9 mm (0.35 in). The procedure for the optical distortion test specifies that the sample be placed 7.62 m (25 ft) from a light source and moved toward the light source and away from the screen positioned behind the specimen at 127 mm (5 in) intervals. Each time the sample is moved, the tester observes the showdown pattern on the screen. The performance requirements of the test require that no light and dark patches representing a secondary image appear on the screen before the sample has been moved 635 mm (25 in) toward the light source. The test procedure requires that the sample be keep parallel to the screen at a right angle to the light source. Proposed Change The GTR visual distortion test, reflected in paragraph S6.11 of today’s proposed regulatory text, is conducted at the angle of installation rather than at a perpendicular angle. The latter is currently used in paragraph 5.15 of ANSI Z26.1. Since distortion is a function of the angle of incidence, the agency tentatively believes that testing at the angle at which the glazing will be installed is a more accurate representation of real world driving conditions. We note that the curvature of modern windshields at the margins makes it impractical to test the entire windshield for optical distortion at the angle of installation. The GTR specifies three vision measurement areas, reflected in S6.15 of today’s proposed regulatory text, on which the optical distortion test is performed, which are designed to capture the area of the windshield used by the driver to see the forward roadway. The vision measurement areas used in the GTR are based on SAE J941, Motor Vehicles Drivers Eye Locations, JAN 2008. SAE J941 defines a range of eye positions developed from a statistical analysis of 2,300 drivers’ physiological data (with a male-to-female ratio of one to-one) performing a straight ahead driving task.29 Elliptical contours defining a range of eye positions were developed from a statistical analysis of this physiological data. These contours, or eye ellipses, offer a representation of a driver’s eye location and can be used to determine what a driver could see in the straight ahead driving task. The optical distortion test in the proposed GTR applies different vision testing areas to differing classes of vehicles. These vision testing areas are referred to in the GTR as Zones A, B and I. The defined vision testing areas Zones A and B apply to vehicles with a gross vehicle weight rating (GVWR) of 4,536 kg (10,000 lb) and less also referred to as light vehicles. Zone I applies to vehicles with a GVWR over 4,536 kg (10,000 lb). Zone A is defined as the area on the outer surface of the windscreen bounded by four planes. The first plane is parallel to the Y axis passing through V1 and inclined upwards at 3° from the X axis (plane 1 in Figure 18). The second is a plane parallel to the Y axis passing through V2 and inclined downwards at 1° from the X axis (plane 2 in Figure 18). The third plane is a vertical plane passing through V1 and V2 and inclined at 13° to the left of the axis (plane 3 in Figure 18). The fourth plane is a vertical plane passing through V1 and V2 and inclined at 20° to the right of the X axis (plane 4 in Figure 18). The four planes correspond to an area forming a box directly in front on the driver’s forward eye position.30 In order to determine the extended Zone A, the part of the windshield subject to the optical distortion test, the box formed by the four planes is extended to the vehicle’s center line and then to the area of windshield symmetric to Zone A on the opposite side of the vehicle’s centerline. The extended Zone A represents an area of the windshield extending horizontally across the center of the windshield. The area of the windshield that comprises extended Zone A must exhibit a maximum of 2 degrees of arc when subjected to the optical distortion test. Reduced Zone B consists of area along the bottom third of the windshield bounded by extended Zone A on the top, plane 9 (in figure 19(a)) on the bottom and plane 3 (in figure 19(a)) and a plane symmetrical to plane 3 on the opposite side of the vehicle centerline on the sides as well as the areas in the upper corners of the windshield separated from each other by the opaque area where the rear view mirror is mounted. The area of the windshield that comprises reduced Zone B must exhibit a maximum of 6 degrees of arc when subjected to the optical distortion test. Zone I, the defined vision testing applicable to vehicles with a GVWR over 4,536 kg (10,000 lb), is determined from the ‘‘O’’ point which represents the driver’s eye location. The ‘‘O’’ point is a point 625 mm above the R point which is determined using the three dimensional vehicle reference system described in ISO Standard 6549, Road Vehicles—Procedure for H- and R-point determination, December 16, 1999. Zone I is comprised of the area of the windshield bounded on the sides by vertical planes extending 15 degrees from the right and left of the O point and on the top by a horizontal plane extending from the O point to 10 degrees above horizontal and on the bottom by a horizontal plan extending from the O point to 8 degrees below horizontal. The area of the windshield comprising Zone I must exhibit no more than 2 degrees of arc when subjected to the optical distortion test. We tentatively believe that testing only in these areas sufficiently assesses the windshield’s optical properties, given that the eye ellipses appear to offer a good estimate of the windshield area typically used by the driver and taking into account practicality considerations. The performance requirements for Zones A and I are more stringent than Zone B because Zones A and I represent the area of the windshield used most by the driver to observe the forward roadway. Zone B is also the area of the windshield closer to the edge where the windshield displays greater curvature. Given that the agency is testing the windshield at the angle of installation rather than at a perpendicular angle, we have tentatively concluded that allowing a maximum of 6 degrees of arc in the reduced Zone B at the margins of the windshield is a reasonable approach to ensuring safe visibility through the windshield. We believe that other than specifying an area of the windshield to be tested, the procedure and performance requirements for these tests are equivalent with those currently included in FMVSS No. 205. The secondary image test in paragraph S6.12 of today’s proposed regulatory text specifies two test procedures, only one of which the glazing must meet to satisfy the test’s requirements. The first test measures secondary image separation by projecting the image of a target through the windshield being tested and recording the secondary image shift of the target. Other than only applying this test to the defined vision testing areas described above, we believe that this procedure is substantially the same as the procedure specified for testing secondary image separation in paragraph 5.15.2.1 of ANSI Z26.1. The other is a collimation-telescope test. When a test piece exhibiting a secondary image is placed between the collimator and the telescope, a secondary image will appear on the polar co-ordinate system. The secondary image separation of the test piece can be determined by measuring the distance of the secondary image from the center of the polar co-ordinate system. This procedure differs from the procedure in ANSI Z26.1 where an image is projected through the test piece and secondary image separation is determined by visual inspection. The agency solicits comment on these proposed changes. We note that in its previous comment, Solutia expressed concern that the GTR’s method of testing the windshield using the installation angle ‘‘does not provide for testing the optics for a driver looking down or to the sides. A fixed angular test methodology can appropriately represent skewed driver vision (down or to the sides) for all vehicles, and reduces the test burden and ultimately costs for manufacturers.’’

I tried wading through that reference, but My Eyes Ghosted Over

 

[ohmman]

I tried wading through that reference, but My Eyes Ghosted Over

Heh. It took me a while when it was first posted as well.

Here is my summary of the methodology in the text.

A windshield, not mounted in a Model X but mounted on a moving cart, is placed 25 feet from a light source, and directly against a receiving screen (think a piece of white paper, for simplicity). The windshield is then moved toward the light source 5 inches at a time. The specification requires that no secondary image ("ghost") appears on the screen prior to the windshield being moved 25" toward the light source. So 5 test positions. The problem with this test is that the glass is tested perpendicular to the light source. That is, not at the installation angle. So for the Model X, that would equate to looking at the stars and not seeing ghosting. It doesn't translate very well.

There is another method that they then reference, the GTR visual distortion test. This one is measured at the angle of installation, which of course is more appropriate. The GTR measures three different areas on the windshield which reflect drivers' eye positions and are reflected in this document. The test is effectively the same, but they measure at installation angle and in those specific zones. For the Model X, a vehicle with GVWR under 10,000, they measure two specific zones that they term Zone A and Zone B. In Zone A, a maximum distortion of 2 degrees of arc is permissible, and in Zone B (the margins of the windshield), a higher maximum distortion of 6 degrees of arc is permitted since the driver isn't continuously looking through there.

In the end of the quote, they suggest using a more technical measurement instead of a human measuring on a visual screen.

So what does that all mean for us? I think we can make a test that simulates this and tests the 2 degree limitation through the center of the windshield. It won't be perfect, but it'll be something. Here's how it looks.

Print out the attached PDF on US Letter (8.5"x11") paper. Ensure that the circle has a diameter of approximately 0.872". Follow the placement instructions and see if the ghost escapes the grey circle. I calculated the diameter of the grey circle with the assumption that the paper will be 25" from the windshield. I used:

diameter = 2 * 25" * sin( 2 degrees / 2 ) = 0.872"

If that looks incorrect to anyone, please do correct me and I can update the PDF. Sadly, the X is out driving around and I don't have a laser pointer, so I can't test this just yet. LMK your thoughts.

--Note, deleted PDF while I revamp it. I have another idea that'll be easier.

 

[AudubonB]

This, if you have correctly extracted it (as I'll take for granted), is of utmost importance:

The problem with this test is that the glass is tested perpendicular to the light source

I've read - not hyper-carefully but better than skimming - through this thread's 800 posts and my belief is most posters have little to no experience in or knowledge of refractive physics. That is by no means an indictment, but....


...to the extent that neither Tesla's design/engineering team nor, worse yet, the optical engineers at AGP Glass properly have anticipated the refractive aberrations inherent in such a raked windshield is a very large miss, indeed.

Any test that passes its light perpendicularly through the test glass is utterly useless at best, and deceptive to the point of being dangerous at worst.

The first time I drove X Audie Orationem Meam in the dark, a few days ago, the dreaded refractions are indeed there. I've made a point out of not telling Jenny; I'm going to use this as an experiment to learn how long it takes her to notice them.

There should be a relatively easy fix, using just the right optic coating to re-refract the image from air // glass // standard-safety-stuff-if-present // glass // air. The potential problems I quickly can come up with include:

• de novo cost
• retrofit cost
• ability for any exterior coating to stand up specifically to wiper action
• color distortion <===this likely is not a big issue; microscopy, camera, telescope & binocular lenses all are nigh-perfect after many decades of research and innovations here

Any interior coating is less susceptible because of the more protected environment.

The basics are simple: the incident of light refraction is a function of the different refractive indices (duh) of different media. Atmospheric air (to consider that as a unique substance, which of course it is not but let it pass) has one such index; windshield glass another. Assume there to be just those two substances: "air" and "glass". Light passes from air to glass and gets refracted ("bent") as a result of the difference of those indices; upon passing out of the glass and back into air it is refracted again.

All well and good....actually, all poor and bad....BUT in that visible light is itself composed of a spectrum of wavelengths, different wavelengths ( = "colors", as all of you well know) refract at different angles.

ALL this needs be accounted for and rectified; in the case of the Model X windshield, the angle at which the driver and the windshield interact must also be accounted. This is where the magic of coatings comes in. Today's coating substances go a long, long way toward correcting these refractions. There is a lot more to the problem - internal reflections, birefringence, inhomogenties - but all can be solved to a great degree - far greater than what we presently are seeing in, through, and of our Model X windshields.

 

[FURY]

I am not an aviator, but there are reported Model X windshield like observations in certain aircraft. Visual anomalies have to have been are a tremendous obstacle in the military aircraft fighter type aircraft. I suspect, that at least at some point, "dogfights" had to have adequate workarounds or the pilot was extremely "focused" on which was which, so to say. An Immelmann, as an example, would seem to require a lotta visual adaptation and discipline...

So what was the military fix...

Thank you very much

Fury

 

[EV-lutioin]

I am not an aviator, but there are reported Model X windshield like observations in certain aircraft. Visual anomalies have to have been are a tremendous obstacle in the military aircraft fighter type aircraft. I suspect, that at least at some point, "dogfights" had to have adequate workarounds or the pilot was extremely "focused" on which was which, so to say. An Immelmann, as an example, would seem to require a lotta visual adaptation and discipline...

So what was the military fix...

Thank you very much

Fury

Here is some interesting information, apparently the issue of ghosting has been addressed with the ongoing development of heads-up displays of both military aircraft and passenger vehicles.

DuPont developed a wedged polyvinyl butyl (PVB) layer in windshields to minimize ghosting in heads-up displays in 2001.

DailyTech - MIT Seeks to Enable the Ultimate Heads-Up Display for War and Peace

Patent US5013134 - Ghost-free automotive head-up display employing a wedged windshield

 

[8XATL]

It's comments like these that make me think that the problem is with Quality Control, rather than the design. If there are windshields in circulation with minimal ghosting, then I have to believe that this issue could be resolved with tighter QC.

Maybe you're right but it's kind of like a comment one other guys wrote on here, Tesla isn't going to make every single Model X windshield right, even though they should.

 

[8XATL]

Here is what I see at night
View attachment 212792

Uhhhmmm....Maybe I don't have a right to complain. Mine isn't nearly this bad and doesn't even show up very well when I take a photo...still annoying.

 

[EV-lutioin]

I found a Tesla service manager who is very sympathetic to our ghosting issues. He told me that he vigorously challenged upper management in a meeting yesterday about the issue. He basically said that this has to be resolved one way or another (this was after seeing how bad the ghosting was in my MX after he took it home at night). However, he said only a few people are complaining, so it would be helpful if more people contacted him to make his case with Tesla engineering stronger. So, if you have a chance, give Randy Filippi a call or email him and let him know you have ghosting too.

[email protected]

Sacramento-Rocklin | Tesla

 

[Terapin]

I just picked up my Model X last week and my windshield has some noticeable defects as far as wavy-ness goes. One is right in the centre of the windshield and it is noticeably rippled. There is a line of wavy-ness across the whole windshield that is very mildly distorting. I'm wondering if this is something they will remedy, and if not, if it's something I am likely to get habituated to. I've never had this issue is a car before! It is fatiguing on the eyes to correct for the distortion constantly. It's like having prescription glasses that aren't quite right.

So far it seems the ghosting is very minimal - my eyes post-Lasik were much worse for a few years!

 

[FURY]

I just picked up my Model X last week and my windshield has some noticeable defects as far as wavy-ness goes. One is right in the centre of the windshield and it is noticeably rippled. There is a line of wavy-ness across the whole windshield that is very mildly distorting. I'm wondering if this is something they will remedy, and if not, if it's something I am likely to get habituated to. I've never had this issue is a car before! It is fatiguing on the eyes to correct for the distortion constantly. It's like having prescription glasses that aren't quite right.

So far it seems the ghosting is very minimal - my eyes post-Lasik were much worse for a few years!

Hi Terapin!

So maybe a renaming for this vehicle or at least the windshield would be the Model LasiX

Thank you very much

FURY

 

[Amp Dup]

I just picked up my Model X last week and my windshield has some noticeable defects as far as wavy-ness goes. One is right in the centre of the windshield and it is noticeably rippled. There is a line of wavy-ness across the whole windshield that is very mildly distorting. I'm wondering if this is something they will remedy, and if not, if it's something I am likely to get habituated to. I've never had this issue is a car before! It is fatiguing on the eyes to correct for the distortion constantly. It's like having prescription glasses that aren't quite right.

So far it seems the ghosting is very minimal - my eyes post-Lasik were much worse for a few years!

I got my Model X in late December. Right away, I noticed a distortion - wrinkle, ripple, waviness - in the driver's visual field. I also noticed the ghosting maybe a 2-3 severity. Mostly on LED lights.
I took the car to the Service Center and the Service Adviser saw the distortion right away and said we'll get you a new windshield. (I never mentioned ghosting)
They ordered the windshield and I took the car in a week or so later. They had the wrong windshield. Since my car had HW2 sensors, they had to re-order the windshield to get the right one. I mention this because I believe it to be an example of the current 'state of the art'.
They installed it last week. No distortion but still ghosting, severity 2-3.

 

[FURY]

Hi EV-Lutionin!

Here is some interesting information, apparently the issue of ghosting has been addressed with the ongoing development of heads-up displays of both military aircraft and passenger vehicles.

DuPont developed a wedged polyvinyl butyl (PVB) layer in windshields to minimize ghosting in heads-up displays in 2001.

DailyTech - MIT Seeks to Enable the Ultimate Heads-Up Display for War and Peace

Patent US5013134 - Ghost-free automotive head-up display employing a wedged windshield

THAT is the crux of the issue for these type of windshields as well and is information that TESLA no doubt has had for some time as well. The Patent must be strong enough that TESLA can't get around it, at least in a cost effective manner, to date. So, unless there is enough sales impact/other action, to compel TESLA to get a redesign, which I believe is going to be the path they ultimately choose, or another remedy, they won't give up the ghost...it seems.

Thank you very much

FURY

 

[ohmman]

No distortion but still ghosting, severity 2-3.

This severity rating - is it the number of images that you see or on some scale (1-10, for instance?)

I haven't forgotten about my promise to make a home-testing kit. I'm feeling like all of our windshields will fall within the guidelines, though, based on preliminary tests. The test distance from the front of the windshield isn't enough at 25' to make the ghost separation wide enough.

 

[Amp Dup]

Sorry. A few posts back or a few pages back, someone proposed a 1-5 scale. 1 being no ghosting, 5 being terrible. That's what I was referring to.
I can live with it. My wife (who is no fan of Teslas, by the way) has only ridden in the passenger seat and has never mentioned seeing ghosting. (It's probably just a matter of time till she does. Then I'll certainly hear about it!)

 

[ohmman]

Sorry. A few posts back or a few pages back, someone proposed a 1-5 scale. 1 being no ghosting, 5 being terrible. That's what I was referring to.
I can live with it. My wife (who is no fan of Teslas, by the way) has only ridden in the passenger seat and has never mentioned seeing ghosting. (It's probably just a matter of time till she does. Then I'll certainly hear about it!)

Got it. Sorry. I still don't know how to gauge a 1-5 rating either. Terrible to one person might be what I have. To another person it might be so many images that you can't tell which one is the real car. So there's a lot of variability. I get three images (real image, ghosted image, and a faint ghost above that), and when headlights approach me in the other lane, the first ghosted image is well above the roof of the car.