Survey Feet vs Feet export to Google Earth

Tim is correct in that a point with coordinates in survey feet will not be at the same location as a point with the same coordinates in international feet. Think of it this way, If you and me both start at the same place and I walk 1000 survey feet north and you walk 1000 international feet north, we will not end up at the same place. I will be further north then you are since a survey foot is longer that an international foot.

If I place a point in google earth at 1000, 1000  survey feet and a second point at 1000, 1000 international feet they should not be located at the same location in google earth.

It sounds like you placed information in the drawing using international feet values but the drawing had survey feet units as the working units. Relating back to what Tim said, If you place a point that should be at 1000, 1000 feet in a drawing and the drawing is in meters, the point will be at 1000, 1000 meters. This is not the same location as 1000, 1000 feet.

You are placing the information using international feet values for the coordinates into a survey foot drawing. The information is not going to be at the location that are expecting it to be but at the actual location (wrong location) that you placed it.

The drawing units and data are all Survey feet from State Plane survey data. They reference in and will reproject correctly to other dgn files even if they use a different geographic coordinate system. However they don't export to KML like I expect them too. 

I do realize that by changing the units to International Feet and then moving the data back to the Survey Feet coordinates the data is then no longer in the correct place, but doing that is what gets the KML data to export closer to the imagery in Google Earth. Thus why it seems that the KML data is not projecting correctly on export. Here is an example below where the data in green is in the correct location in Survey Feet while the Red is changed to Feet and then shifted back to the survey feet coordinate so it's in the wrong geolocation. However you can see that the east west alignment of line work is much better on the data in the incorrect location than it is when exported from the correct location in Survey Feet. The north south is off by about the same in both. I realize it will never match the imagery perfect but I would expect it to be better without having move the linework to the incorrect geolocation. 

State Planes (GRS 80) uses a different Datum than Google Earth (WGS84).  The long/lat in one is not the same point as the long/lat in the other.

State Planes (GRS 80) uses a different Datum than Google Earth (WGS84).  The long/lat in one is not the same point as the long/lat in the other.

Is there a conversion or work flow you can recommend?

what  happens with synchronise a point in mstn  to google earth does google earth say find a light pole from the survey in survey feet?

What are your screen shots? .. where is the imagery from?

There is probably the same amount of shift when using the Synchronize Google Earth View command. It's hard to tell exactly since it's only visual with no graphic overlay to compare. 

The screenshots are from google earth with an exported KML displayed on it.  

I get that, but shouldn't MicroStation be smart enough to do the proper translation on export to KML so the data is projected in the correct location in Google Earth? 

Yes, with Geographic Coordinates defined it's always worked for me. I've spent more time than I will admit tracking down accuracy standards for aerials. They are not all the same. Google aerials remain undocumented. Some are pretty good, say within 2 feet, some will be off much more.