924Board.org

[MikeJinCO]

I wrote up my work and it is 3 pages long. I'm trying to past it in a .txt format. PORSCHE 924

Head Flow Testing and Development

I started the plan for head development when wintering in Mexico in 2012-2013. I envisioned building an engine with a normal 924NA head using high compression pistons with a standard CIS system on a modified stock intake with a Audi 5000 throttle body and some MSDS headers I got on Ebay. I planned to do the head development myself so I would need some sort of a flow bench to work with

My reference for most of this work has been David Vizard's book “How to Port and Flow Test Cylinder Heads”. It provides a lot of information on flow benches, testing and things to do. Probably the most important advice that applies to the 924 is that if the port is an inherently bad design to concentrate on the valve inlet area and throat first.

I didn't get very far before it started getting complicated. First I got Nick Thompson's head with two severly burnt valves. I built my “block “ using a block of wood the same size as the head and boring a cylinder of close to the correct size of 86.5mm. I then bored a side hole for the exhaust into a fairly large shop vac. I then built a valve press using a 7/16” 20 tpi bolt, or .050 per revolution. With my first series of tests I was not getting a good seal to my “block” so have basically discarded those results as I was just learning how to do this, I also used a 60” wp magnahelic gage that I had around, it wasn't accurate enough so I made a 6' U tube manometer using plastic tubing, colored water and 2 yardsticks.

I then came up with a 931S2 motor complete, cheap and all of the parts have been in excellent condition so the question came up as to do I build a motor using the S2 pistons and a NA head or get special pistons and use the 931 head, remembering that a new set of 931 rings will cost about 1/3 or more than a set of Diamond or JE pistons with rings and pins, plus being lighter. My initial tests(still using my lousey sealing system) showed that the 931 head flowed about 10-13% more than the NA head. Then I discovered the real problem. My NA head off of a '79 had 38mm valves. So I got Grenadiers to send me a NA head with 40mm valves and my enclosed tests are the result of my tests now reflecting better head to block sealing using some gasket material coated with wheel bearing grease to get a better seal.

I hve no idea as when will get all done as I don't have an unlimited budget to develop both an NA and 931 head at the same time, and I have developed a serious back issue that will require surgery on Nov. 4th which will put me down at least until March, and possibly end my mechanicing career due to having at that time a 6 to 7 level back fusion.

These tests are done using a 6' U tube manometer. All data is in pressure drops, not CFM as I have tried to make a calibration plate and so far been unsuccessful. I'll keep trying. The math formula is to divide the square root of P2 by the square root of P1 and that gives the percentage difference in flow. What I have done is plot the pressure graphs and then lay them on top of each other and so far they simulate the flow graphs that have been done on professional flow benches. The 931 head tends to improve at highe lifts and the NA head gets peak flow then backs off at about .400 lift.

All these tests except the 38mm were conducted in as received condition, no lapping, cleaning, etc. I wasted some money on the 38mm head before I found out it was the small valve.

931 Head, 1.375” dia. by 2” inlet cone, average of 3 tests.

Closed 44.625
.050 lift 36.3125
.100 29.0
.200 19.50
.300 15.0
.400 13.0
.450 12.5

I don't go to .500 as my depressing device is approaching its limit and the other charts I have seen don't change much at the higher valve lifts. The inlet cone has a 1/4” radius on the front and fits the ports “reasonably”.

40mm Valve NA Head/w air horn
average of 3 tests

Closed 46.5
.050 38.5
.100 30.5
.200 20.312
.300 15.6
.400 12.5
.450 13.0

40mm NA with Weber intake manifold, no air horn
1 run only

Closed 46
.050 38.5
.100 30.0
.200 20.25
.300 15.0
.400 11.25
.450 12.5

The Weber manifold does not match the ports very well, upper right port intrudes almost 1/8”(3mm), while the lower left manifold overlaps about the same. Also not sure about that .450 reading, or it could be the turbulance of not having a radiused inlet.

40mm NA with stock intake manifold.

Closed 44.5
.050 37.5
.100 30.5
.150 25.25
.200 21.25
.300 15.625
.400 13.25
.450 13.25

On graphing you can see the fall off at higher flows due to the added restriction. I once thought that the way to improve the manifold would be to saw it apart, smooth and slightly enlarge the runners then weld it back together. I think that was a correct thought.

I have both a manifold set up with the original '77 CA smog(very small) throttle body and an Audi 5000 throttle body. And would like to test them, but I only have 5 days from when I get home until going to the hospital so it probably won't get done. One of my first tests using much worse methods showed the small throttle body to be very restrictive.

Final Test
38mm NA with air horn
average of 5 runs

Closed 46.55
.050 39.675
.100 32.55
.150 26.55
.200 21.90
.300 17.125
.400 16.60
.450 16.825

Big differences here, wouldn't it be nice to see how a 44mm intake would compare? Would the port then become the limiting factor.

If some one who knows graphing could lay these curves out using valve lift on the x axis and Delta P on the Y, then overlaying the starting values it results really show the air flow differences.
_________________
Mike


'67 MG Midget Dp
'71 Ocelot Dsr Kawasaki 1000(under rebuild)

[vatoisimo]

are you saying that the smaller diameter opening flows more air?

[!tom]

I plotted the numbers using Octave (a GNU Matlab clone), which seems to have a few bugs when it comes to annotating plots. The title is only partway there, and the X-axis label is nowhere to be seen.

As requested, the plot has pressure drop on the Y-axis vs. lift on the X-axis. In order to align all of the plots, they have all been shifted so that the pressure drop at 0 lift is 0, so you are seeing the reduction in pressure drop from closed on the Y axis. Basically, I subtracted the 1st observation from each set of observations.

Here's a PDF of the plot. Someone else can turn it into a picture if you'd like it posted that way.
_________________
78 924 NA
5-lug

[MikeJinCO]

Itom, thanks for the plot, I was going to work on that today. BTW I had that turned upside down(on graph paper,the bottom of the graph 50 and top 0) so you get a curve that looks like the typical cfm vs lift curve. Vatismo the 38mm has a much higher PD at the higher lifts than any of the 40mm valve which means that it has significantly lower air flow. And to make it more significant this head had the throat above the valve bored larger and much smoother the the rough cast throat I found on my '77 when I had the head off.

What I noticed about both NA heads is that the valve is slightly recessed at about the same angle as the valve seat. Per D. Vizard that is a problem area as the gas needs room to expand as it enters combustion chamber. I'd look for a way to break that angle back by about another 15 degrees. Then do a 3 or 4 angle valve job and test again. The 931 head has the valve flush to the top of the combustion chamber so that extra angle cam be integrated into the valve seat. Unfortunately I won't have the opportunity until next spring.
_________________
Mike


'67 MG Midget Dp
'71 Ocelot Dsr Kawasaki 1000(under rebuild)

[!tom]

Like this?

_________________
78 924 NA
5-lug

[ideola]

Fascinating stuff. Good work Mike. Best of luck with the health issues, hope it's not the end of your wrenching.
_________________
erstwhile owner of just about every 924 variant ever made

[MikeJinCO]

Thought an update is necessary. I'll try to write a full report soon on my progress, but I've got too many projects going at once to make progress quickly on any of them-and I'm retired..

Thanks for all the good wishes, my back operation was very successful and I'm better than anytime in past 15 years.

I've been working on a 38mm valve head that I got from Nick Thompson and by smoothing and blending the valve inlet area, then filling the bottom of the port runner with epoxy(Splashzone-A788) and then reworking the top of the port I was able to significantly improve the low lift flow and increase the high lift flow up to slightly better than a stock 40mm head.

I'm looking for a wasted 931 head as I will completely trash it trying to get it to work for a NA application. The Weber manifold ports are quite different and the design philosophy is radically different. The NA head is actually not a bad design taking into account that the air has to bend 490 degrees after it leaves the throttle body. I think now that the 931 head works as well as it does by my tests because the ports are a little smaller so it maintains velocity better. Other parts of the design like the curvature of the bottom of the port is terrible for NA application so the sweet spot may take a full head worth of filling and grinding.

I'm doing all this A: for fun I am one of those goofy mechanical engineers, and B I think the gains in combustion efficiency from the improved combustion chamber and much better squish are well worth going after. Obviously new pistons will be required, Grenadiers has blown up my old ones anyway.

More soon.
_________________
Mike


'67 MG Midget Dp
'71 Ocelot Dsr Kawasaki 1000(under rebuild)

[MikeJinCO]

Second attempt to post, maybe this time I'll get it right.








My attempt this time around was to try to develop a little more usable and standardized data and then actually modify a head. First was to improve the system on my mega dollar ($75 by now) flow bench. I made a head gasket out of standard gasket material and fit it to both my block (bored out wood) and the head. Of course, I found them a bit off so adjusted them to get the head centered on the bore, which should help due to the constantly shrouded valves. I also used wheel bearing grease as a sealant and it has given me much more consistant numbers.

For my original tests I used a wood block with a curved entry to make an inlet nozzle. This had mixed results due to the various fitments. I decided to change to using a handformed entry out of modeling clay for everything and the results have been “interesting”. I do not have a dial indicator setup valve depression so I use a 20 tpi bolt, so individual data points particularly below .100” are not the greatest.

Problems I have run into include cylinder head design, financial and time. Per David Vizard the first thing to do is to develop a good valve seat. My local guy says he has a great cutter he uses for the local circle track folks, but paying him for the set up and time to do one valve is just out of the question - and besides he is qumm into the head with an entry angle of 45 deg, same as theite busy making a living. The design problem is that on a NA head the valve seat is recessed MOL 1 valve seat. For low lift flows this angle needs to be made into at least 60 deg.
The top corner of the ground intake valve also needs to be tapered to improve the flow angle off the valve. Again set up time and cost. For the entry I have found on ebay some cheap/junk router cutters that should be able cut this minimal amount of aluminum. We'll do the valve work when grinding the valves. So the bottom line is I can't really get a good valve seat and entry first.

My first round of tests was with a standard 40mm intake NA head. I did 6 runs and then did the math using both averages and dropped out the high and low and then average. I also did a little dropping of surprisingly few obviously bad numbers. This head which I got from Grenadiers was uncleaned with not cleaned up valves-just as off an old motor. It is currently out getting cleaned and blasted.

Results: With out the air horn the NA head did not flow nearly as well as my initial tests. The Weber manifold made very minor improvements a higher lifts(air horn effect?). And, surprise of all the stock intake manifold helped out quite a bit at the high lifts. That says something about needing port velocity rather than volume. I did not expect to see this.

I have purchased bought some silicone molding compound then will take this manifold, saw it into 3 pieces and take some molds probably DanB or 924RACR will get these when I'm done. Then I'll clean up the manifold ports, glue/weld/somehow it back together and run it again.



Second Phase: Take the 38mm valve head and see what can be done with it. I could go on for pages about the design philosphy, if your that interested go find David Vizard's book. This head had been cleaned although the valves just on my wire brush. The port volume was approximatel 101cc.

I first tested it then took out my porting kit($30) and worked on blending the throat machined surfaces into the casting and smoothing out the bowl and to break the badly angled corner on the head face. This tapering was not well. This head was pre-planned as a throw away when done, so no great damage here. The results show a pretty decent improvement at lower lifts with no effect at higher lifts just as predicted. These results were accomplished with as poorly a formed, cheap, wide factory style valve seat as can done. Head volume up very slightly, .5 to 1 cc.

I then scarified the bottom of the port so that I could add epoxy. I used 60 grit roll and it worked pretty well, the corner is tough to get. The epoxy I used is Splashzone a good marine epoxy and recommended theough internet searches. The Benelli(?) that Gegge's fellow used is probably better but with very limited availability in this country, I found a supplier in Wyoming who used it for repairing oil field equipment, a machinable epoxy on super teroids and probablye priced to match. The Splashzone is easily shapable with abrasives but pretty difficult to work with due to its stiffness in those small areas of the port. When I am done I will have a lot of epoxy left over for the next person to do this.

I filled the lower curve let it cure for several days and then proceeded to rework the the surface all the way out to the port surface on side ofthe head. When finished as well as I could, the tight part of the curve is difficult. I then started opening up the top on either side of the valve stem. The head is designed properly for this vertical(to the cylinder) valve application as the cylinder wall side is larger so that the air takes the long way around and moves in toward the center of the cylinder rather than crashing into the partially shrouded cylinder wall. When I got done I retested and got the following graphs. One anomoly here was that at .150 lift I start getting a hormonic which peaks at.200 lift with reduced flow and then it completely disappears at .250 lift. I don't have any answers for this except that it probably isn't good.




_________________
Mike


'67 MG Midget Dp
'71 Ocelot Dsr Kawasaki 1000(under rebuild)

[MikeJinCO]

I hit submit by mistake, fortunately it work ok tho incomplete. One more bit.

The bottom line from this is I now have a cylinder that flows much better. Better than even than a head with the larger 40mm valves. I need to get the other head back and work with it. I'm not terribly comfortable with these results. I do well realize that my results are significantly different from the few other reports I have read, I am just reporting what I measured. I have 15 pages of data and graphs if anyone is interested.

As I go thru these tests I think about all the “internet experts” who say that Porsche intentionally made the 924 motor bad, etc, etc. Porsche engineers actually followed the still recognized design practices for this type of head design. The design and manufacturing budget with VW being a great consideration it is resonably state of the art for the day, and when they took over the project there was probably no money for a complete head rework, Porsche was still a pretty small player back then, my book has total Porsche production in the year prior to the 924 at 9000 units.

I took engineering compressible flow courses at U. of Missouri in about 1971 and things like CFD programs hadn't been dreamed of, our computer an IBM 360/50 took the second floor of the EE building and operated off punch cards. Developing a new head then took lots time with hand made models of everything including all all the cooling and oil passages, not a minor endeavor.

Where to from here? First to rework the 40mm valve head as much as possible as I still don't know if I'll use it or the 931 head on my motor. Secondly to get a second scrap 931 head to work on as it has pretty ports for a pressurized application, but they are completely wrong,in theory, for NA use even though my first tests showed it flowed slightly better-explain that one to me. I figure it will take destroying a head to get even close and even then I may not be able to get it done well enough to use. 931 heads are rare enough to avoid scrapping my good one if I can't get it to work. If I can get it to work then the improvement due to the increase in combustion efficiency should make it worth while. If I find time I will also take the 38mm head and rework the other intake ports to find out how possible it is to duplicate this work as I do have the some experience of one intake port.

Bottom line, I'm enjoying doing this, I just wish I could get my toy car beasty on the road. I end up going very slowly as I also have a new project of a older Ford truck and camper being set up for a 2 winter trip thru South America, it is quite involved due to different services available(ie 220v 50cy power vs the US 110v, 60cy) and the total lack of spare parts for the truck and camper.
_________________
Mike


'67 MG Midget Dp
'71 Ocelot Dsr Kawasaki 1000(under rebuild)

[MikeJinCO]

I made a poly model of a stock 40mm NA intake port and have some pictures.












The port size in the cut off manifold is about 43mm. At the port it is about 32mm so that is a nice progression, it is a fairly rough casting that was easily cut on a large wood working bandsaw, It could easily be cut apart, but the welding back together could be a little more difficult requiring preheating, extensive cleaning. A pro welder would be best for that.

The intake is about 1mm larger than the head except about 10-15 degrees clockwise from top where it gets to maybe 3 mm and that little soft lip on the bottom of the port which probably has very little effect. I'd say porting to the intake gasket is probably a mistake. Time to make your own gasket as a pattern.

I'm still looking for a scrap 931 head.

I'm going to take some modeling clay and fill in the bowl pocket behind the valve stem and see what happens on my 38mm head. The more continuous curvature of 931 is the major improvement to that port IMO, as the bottom curvature appears a little worse. I will look a straigtening the cylinder wall pocket a little more, but don't want to increase the area much.

I'm still looking for a scrap 931 head.
_________________
Mike


'67 MG Midget Dp
'71 Ocelot Dsr Kawasaki 1000(under rebuild)

[Rasta Monsta]

Fully intending to send it...sorry for the delay.
_________________
Toofah King Bad

• WeiBe (1987 924S 2.5t) - 931 S3

[MikeJinCO]

Thanks, Let me pay you for some time for packaging also. Nick T send me his by mail. Don't need cam bearing caps or anything like that. Put some clay in the 38mm head and the valve bowl looks much better, I'll test it soon.
_________________
Mike


'67 MG Midget Dp
'71 Ocelot Dsr Kawasaki 1000(under rebuild)

[nickthompson]

The work you are doing is completely awesome.

[MikeJinCO]

I tested my further modified 38mm NA head yesterday. The mod was to fill the valve stem pocket with some modeling clay and give the port a nice smooth curvature down to the valve. When I put it in there it really looked good.

Unfortunately the results were mixed. I got more or less the same flows up to .350 valve lift and then it started improving a little. At .500 the pressure drop went from my previous 12.0" of water to 11.5 which is an 8% increase, however the area under the curve is really very small which means to me that the improvement in volume in the cylinder would be very small.

However, since it now looks more like a 931 port it is interesting that the flow curve now more closely follows the 931 curve of increasing at higher lifts rather than the flattening more typical of the NA head.

I did reduce the port volume while doing this which might have something to do with the result, It could well be that I have just reached the practical limit for my non-professional setup and experience. I also don't have a really good valve seat and profile which should have a much greater impact.

I'm going to take the head down to the machinist to talk about how to remove(or partially remove) the valve guides, do this and the get the guides back in when place without breaking the epoxy back out.
_________________
Mike


'67 MG Midget Dp
'71 Ocelot Dsr Kawasaki 1000(under rebuild)