FI serial numbers vs. build sequence for 380 units

**John D.** · March 31, 2016, 10:12 AM

Re: FI serial numbers vs. build sequence for 380 units

Michael, Thanks for posting the info from Nolands. We have to remember how old that survey really is.
Obviously some of those FI units have been moved around from car to car.
That's what is so nice about finding a real deal '65 FI car with the serial number and original ID tag on the plenum.

I am sure you will be getting some replies on this survey shortly.
Meanwhile the last 64 you listed with a 7017375R.Serial number of unit is 3015, car 17205. Glowing example of an error on the author or the imput from the owner. By the way. A 7017375 unit with number 3015 would not typically have an R after it would it. Thanks, John

**Mike L.** · March 31, 2016, 11:52 AM

Re: FI serial numbers vs. build sequence for 380 units

Michael, I have been involved with my own survey about this matter for about 30 years, and what I have are Vin #'s and fuel unit #'s for many cars before what you have posted. I'm a owner of a 64 FI car with a 375R unit. It looks to me as the change over to the 380 units was around the third week of Nov. 1963. @ Vin #4951. Not to say that there might have been some 375R units that slipped in later. I don't think that a ser. # 17205 could have a 375R unit when it left the factory.

**Michael J.** · March 31, 2016, 01:28 PM

Re: FI serial numbers vs. build sequence for 380 units

Thanks guys, as a new '64 FI owner I am just curious how the serial numbers sequences match up to build sequences. From Nolan's book's data (admittedly old) it seems to have little rhyme or reason to it, pretty random looking. I was really curious if that randomness was TFP, or if it was an artifact of many FI units being swapped around and changed over the years. Not many FI cars with 380 units anyway, but I was wondering if the way the FI units were stored and inventoried could explain the randomness of the ones that got pulled to be on a car as it was coming down the line.

**Dave S.** · March 31, 2016, 02:14 PM

Re: FI serial numbers vs. build sequence for 380 units

Originally posted by Mike Lombardi (9525)

Michael, I have been involved with my own survey about this matter for about 30 years, and what I have are Vin #'s and fuel unit #'s for many cars before what you have posted. I'm a owner of a 64 FI car with a 375R unit. It looks to me as the change over to the 380 units was around the third week of Nov. 1963. @ Vin #4951. Not to say that there might have been some 375R units that slipped in later. I don't think that a ser. # 17205 could have a 375R unit when it left the factory.

Mike,
Are you keeping track of serial numbers of 7380 units in 65 cars.?

**Wayne M.** · March 31, 2016, 03:19 PM

Re: FI serial numbers vs. build sequence for 380 units

Here's two more;

1964 VIN 14732 with 380 unit stamped 2087; (VIN stamped on front tube boss of plenum).

1965 VIN 06114 with 380 unit stamped 2642

**Wayne G.** · March 31, 2016, 05:23 PM

Re: FI serial numbers vs. build sequence for 380 units

I worked on this unit at the Bend regional in 1982. Original owner '64 Coupe, Vin #3132, 7017380 S/N 1882.
Hope this helps your Vin/S/N sequence project.

**Michael J.** · March 31, 2016, 05:37 PM

Re: FI serial numbers vs. build sequence for 380 units

That's the earliest VIN I have seen with a 380 unit. And that 380 unit is 882 units into the sequence. Just shows more randomness it seems. I also saw 380 units on '65s, in Nolan's lists, '65 VIN 6362 with #1681, VIN 7574 with #1843, and VIN 12320 with #1891. They are all over the place it seems.

**John D.** · March 31, 2016, 05:54 PM

Re: FI serial numbers vs. build sequence for 380 units

`Early "64 FI cars.There was a time in November of 1963 when St. Louis was installing both FI units. Even if I didn't have proof of this and you were a novice logic would dictate that this was feasible.
Frank Scibiaca of RP once told me that RP was under big time pressure from the St. Louis Assembly line to get rid of the cranking signal valve. Hence the development of the '7380 unit.
But meanwhile RP was loaded with 7375 units so they tried to unload as many as possible because the "bean counters" Franks language) were upset. But after very early December '63 it would not be typical to see a '7375 R unit on a real fuel car.
Waiting for Jim to chirp in on his extensive survey. John D.

**Michael G.** · March 31, 2016, 07:24 PM

Re: FI serial numbers vs. build sequence for 380 units

Here's a graph showing most of the 1964 cars in Noland's survey. The line is the expected relationship (based on this data). I threw out one car, 7297, because it was an extreme outlier. The rest look pretty good, but 2446 is 300 units higher than you'd expect, probably not yet in the plant during that build, so too high to have come on that car from the factory. The three units somewhat below the line could be real, but may have beed repaired at the plant (or some such delay)

You can extend that line for the lower cars that were 380s.

Screen Shot 2016-03-31 at 8.16.18 PM.jpg

**Michael J.** · March 31, 2016, 07:53 PM

Re: FI serial numbers vs. build sequence for 380 units

Michael. thanks for the graph, it shows the general correlation. There is a positive correlation (slope) between VIN and 380 serial #s as expected, but I would estimate the correlation coefficient is about maybe .7 or so. Not real high, and the SD is pretty large, so that if you had a car in the 14000 to 18000 range, it could be FI serial #s 1700s to maybe 2500s, so a pretty big spread, and again I do wonder how these FI units were inventoried, stored, and even delivered to assembly if you assume the FI units were built and delivered sequentially, and maybe they weren't. This would indicate they weren't, IMO.

**Michael J.** · March 31, 2016, 08:47 PM

Re: FI serial numbers vs. build sequence for 380 units

I made this graph and extended the VIN vs. 380 FI serial # into mid '65, and the correlation disappears.

Attached Files

VinVsSerial.png (39.3 KB, 48 views)

**Michael G.** · March 31, 2016, 09:16 PM

Re: FI serial numbers vs. build sequence for 380 units

Originally posted by Michael Johnson (49879)

Michael. thanks for the graph, it shows the general correlation. There is a positive correlation (slope) between VIN and 380 serial #s as expected, but I would estimate the correlation coefficient is about maybe .7 or so. Not real high, and the SD is pretty large, so that if you had a car in the 14000 to 18000 range, it could be FI serial #s 1700s to maybe 2500s, so a pretty big spread, and again I do wonder how these FI units were inventoried, stored, and even delivered to assembly if you assume the FI units were built and delivered sequentially, and maybe they weren't. This would indicate they weren't, IMO.

Mike, I disagree. You can't analyze this data as if it were untampered with. Because there are fake cars, it can safely be assumed that there are a number of data points that are clearly bogus and should not be used to establish the "best" fit line. Additionally, this data is unusual, in that there are many data points that were actually generated in production, but do not represent the standard production procedures. This is because there were units in the plant that got misplaced repaired and put on later cars. Those, from repairs, etc., are outliers well below the line, that, while they might be real fuelies, really shouldn't be used to establish the "typical" line.

To establish a "typical production" line, then, if you look at the fuel unit/car combinations in both 1963 and 1964, you will find that the vast majority are found to lie very close to an expected line. This suggests fairly rigorous first-in-first-out material handling (not perfect, but good for GM). Adding in the obvious outliers caused by repairs and fakes moves the line away from all those points and you can learn very little from the result.

In essence, many of the data points cannot be analyzed as if standard deviations were expected. While you can use a certain selection of the data to establish a meaningful line, pinpointing with some accuracy where your fuel unit / VIN relationship might be if you had a "normal" car, with this data, standard deviations don't mean anything. While any point below the line might be a real car, any point more than say 50 units above the line is probably not. That 50 points is not established by data, but by the likely number of units the plant typically stocked at one time - you can't put on fuel unit that hasn't arrived from Rochester yet...

**Loren L.** · March 31, 2016, 09:20 PM

Re: FI serial numbers vs. build sequence for 380 units

First, come to grip with the idea that the units were NOT on a First In, First Out basis, immediately causing sequential problems. Then add the question of how many units overall were on hand? 1325 in '64, 771 in '65? On hand at any one time, 100? Does another shipment of units arrive before the room is emptied? You're chasing a wisp......

**Michael G.** · March 31, 2016, 09:53 PM

Re: FI serial numbers vs. build sequence for 380 units

Michael, that 65 data suggests that something different was done the 1965 model year. No clue what, here.

Loren, the 63 and 64 data would tend to suggest that there was some semblance of first-in-first-out practiced, inadvertent or not. (not typical for GM in those days) Not rigorous, mind you, but a pattern did exist, in both years. 65 was clearly different.

Your points about stock-on-hand is valid. Certainly there was not much stock at any given time of this expensive, hard, to manufacture assembly. As I mentioned, you can't put a unit on the car that hasn't been made yet, so there is some finite limit to the highest fuel-unit serial number you should find on any given car. That number is not too far above the line, maybe 50, maybe 100 units, certainly not much more.

FI serial numbers vs. build sequence for 380 units

FI serial numbers vs. build sequence for 380 units

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment