BP4W and BP05 heads flow
 

Found this recently on another forum (mx5nutz), and since it contains some good info/measurements, I thought I would share here for future reference as well. All credits go to na8cgee over there!

This project is to see how much flow/power I can make with thorough testing of the cylinder heads via a flow bench (which I am currently designing/looking to build shortly).

So to start the project off, I sourced a BP05 (93-98 1.8 mk1) and BP4W cylinder head. These were stripped of their valves on two cylinders and I used a horizontal bandsaw to make cuts across the ports. The end result was good, it helps me to visualise what I want to achieve, shows me the thickness of the casting, allows me to take accurate measurements and also leaves 1 individual cylinder to perform porting/development on.

Anyway here's some cross sections of the heads, first up is the BP05 mk1 head:

Attachment 228170

Attachment 228171

Second up is the BP4W:

Attachment 228172

Attachment 228173

There are many differences on the inlet side of the casting between the two but the exhaust side is very similar. Everything being discussed beyond this point is a comparison between stock castings.

Inlet side

There are three main differences on the inlet side.

The first is the port angle to the valve: the BP05 has a port angle of 12 degrees vs the BP4W at 17 degrees. The higher the angle, the more the port ‘see’s’ the valve which as a general rule means more flow.

The second is the port floor, the BP05 head has a lower port floor and is fairly flat/un-shaped vs the BP4W which has a higher floor and is shaped to improve flow to each cylinder. This goes in the BP4W’s favour as it should help improve flow at low valve lifts/low port speed.

The third is the port shape into the bowl transition / short side radius. The BP05 head has significantly better shape in this area compared to the BP4W, which will help to evenly spread the airflow around the whole valve circumference. The short side radius on the BP4W is pretty horrible with no real radius at all.

Exhaust side

Not too many differences between the two castings on the exhaust side. You can see the bean counters became a factor when casting the BP4W as there is significantly less material around the port (This is most evident around the port exit near the manifold flange). The castings are not great, significant restriction from the valve guide area and both SSR (Short side radius) are pretty poor.

What's the plan?

My rough/guess ported BP4W made 217cfm at .400" (10mm) lift, I know this head has achieved good power on my current engine so as a minimum I want to at least achieve this but hopefully exceed it. I'm fairly confident I can achieve the same peak flow at the same lift regardless of the castings I use, but its the area under the curve that i believe can be improved on the most.

You'll see in the pictures some black pen marks on the inlet side of the ports, the solid lines is the safety margin (This is a 2mm line that I can port upto but still retain the 2mm material thickness I believe is required). the dotted lines is my rough guess to how I want the modified port to look although this does not include the epoxy I will be adding into the port, it just facilitates it.

My initial thoughts are that the BP05 is the best casting to continue with going forward, these are the reasons why:

1) SSR - the radius on the BP4W is bad, you can see in the pictures its almost a cliff edge rather than a radius!

2) Material thickness - the BP05 has significantly more material thickness in the areas that need significant modification (Port entry, bowl area and valve guide intrusion)

Here's an interesting one though, the BP4W has significantly better port angle of entry. I don't yet understand how significant this will be but my basic understanding is a higher port angle = better flow around the whole valve diameter. BUT, is a better SSR worth more than a better port angle? I don't know but flow bench testing will tell me!

I had 3 main goals to achieve before I build my flow bench:

1) understand the limitations of each head casting
2) create moulds of the intake ports from each casting (in as cast form)
3) perform some maths calculations to compare stock casting measurements Vs theoretical ideals

Head casting limitations:

With the inlet ports of each head casting cut in half, I wanted to find how much material I could safely remove from the port roof whilst retaining enough material to be of solid construction. This meant a lot of grinding, somewhat crude but got the job done. In the end I found that the BP05 casting has less thickness about 1 inch into the port on the roof, to the point where I thought I had left 2mm of material I actually broke through (glad this happened on my test head!).

Attachment 228174

The BP4W head on the other hand didn’t have this thickness issue at the same point and allowed me to gain significant port angle to the valve face:

Attachment 228175

(Please excuse the excess grinding around the valve seat area, I was seeing how much material was around the seats when they were removed to see if I could fit bigger valve seats in!)

This ability to improve the port roof, I think, will allow me to build a better short side radius on the port floor by adding material whilst retaining the port cross sectional area I need to make power. This is just applying theory at this stage but hopefully in the coming months I can prove that theory with improved low lift flow.

The following pictures shows some dimensions I have taken when studying each casting:

Attachment 228176

I think the BP4W head with its potential 25 degree port angle will be the better of the two with the figures I have at this stage, but the flow bench will prove/disprove it!

Here’s some pictures of the improved port angle, the easiest way to see the
Difference is how much of the valve guide is visible on the modified half:

Attachment 228177

Attachment 228178

And here’s the comparison between modified BP05/BP4W ports for port angle:

Attachment 228179

Create moulds of intake ports:

This part was very easy and quite fun, the first step was to remove the valves and clean the inlet tract with brake cleaner to remove as much carbon as possible.

Then the valves were placed back into the head and the castings were placed so that the inlet port was facing up/silicone would run down to the valve area. Then you mix up the lovely silicone, it’s really easy to work with (my first time using it). I just placed the silicone & hardener in the correct ratios into a paper cup, stirred until the colour was even then pierced a hole in the bottom of the cup and let the silicone run into the ports - allowing it to fall in the port from a reasonable height to help remove air from it. Leave them to harden for 12-16 hours (or in my case overnight in the house, I suspect it would be too cold in the garage to cure). Once fully cured, remove the inlet valves then using a screwdriver and silicone lubricant spray, try my hardest to remove the mouldings! They certainly took some prying/lots of silicone spray to remove but got there eventually. Here they are fresh out of the ports/labelled up:

Attachment 228180

The next step was to study them for any signs of a short side radius, first up is the BP05:

Attachment 228181

Next the BP4W:

Attachment 228182


As you can see, the BP05 has a much better short side radius. It’s actually better to say the BP4W doesn’t really have a sort of short side radius lol. More like a cliff edge into the valve!

The next was to view the port angles, BP05:

Attachment 228183

BP4W:

Attachment 228184

Finally the last job was to cut each mould into sections to compare the cross sectional area of each section against each other.

Attachment 228185

Attachment 228186

They were both cut up at the same points for comparison reasons.

I have taken measurements of the CSA of each moulding and it’s quite interesting the differences between the two. I haven’t yet quite completed my 3rd goal which was performing the calculations, I still need a good few hours to sit down and go through it all.

On another front, my flow bench design is coming together, follows a few guides including Dtec (Google it there’s an online PDF you can view/print - it’s very good) and I’m following David Vizard’s books as well which are also good.

[Cont.]

So with the Christmas period been and gone, I’ve made some good progress with my flow bench build.

I probably should have taken more photos during the build process, but it’s basically a MDF box with 2”x 2” support for strength. It uses 6” PVC pipe work (to ensure the flow bench can flow large amounts of air and is not limited by the pipe work)

The flow bench itself contains four 2 stage vacuum motors (the type from a Henry wet/dry vac), using 2 digital Manometers and calibrated orifice holes. It also uses 3 different size valves to bleed off air to achieve the correct test pressure as required.

The build itself follows the same principle of the DTEC bench but I’m using digital manometers rather than an actual water manometer. I’ve done this mainly for easy of install and improved accuracy.

I only have a few pictures (I forgot to take them!) but here’s a picture with the main side cover not fitted, which shows the orifice holes/air chambers:

Attachment 228157

And the individual orifice tiles which have been sealed in (flowed at 10” water on a superflow calibrated flow bench):

Attachment 228158

And finally with it all sealed up and the 3 valves in place, still need to make the orifice access door:

Attachment 228159

I’ve tested it to in excess of 50 inches of water (digital manometers ran out of range!) and it the box didn’t collapse which was good. The bore adapter was fun to make, it’s just MDF and a 83mm ID aluminium bore pipe which is the same stroke height as the actual engine. It also used the head bolt holes to align the bore as per the head gasket would do.

Attachment 228160

Everything works but I need to make a valve opener so I can test at various lifts, I’ll use a manual height adjuster and a dial gauge to measure the distance - shouldn’t be too long till I can start messing around and making some gains!
I've performed in excess of 30 tests on the BP4W test head, ranging from the standard setup to the outright bonkers ram it full of putty and see what happens!

I'll share with you some of the interesting comparisons that I found when messing with these heads.

So first up, how much did the BP4W head flow in stock form on my bench, at 28":

0.1 - 74
0.15 - 112
0.2 - 140
0.25 - 165
0.3 - 179
0.35 - 189
0.4 - 195
0.45 - 200
0.5 - 205

These were relatively low numbers but remember flow benches are like dyno's to some degree, its the Delta between two tests which is the most important here. or in other terms my flow bench seems to be a bit of a heart breaker!

The first interesting test was to fill in the injector opening within the port with putty, such a small addition of putty but the results were interesting:

0.1 - 80
0.15 - 115
0.2 - 144
0.25 - 169
0.3 - 187
0.35 - 197
0.4 - 204
0.45 - 207
0.5 - 213

This gave a nice increase in CFM across the board with little to no Impact on port cross sectional area = Easy win. Here's the picture of what it looked like:

Attachment 228161

Considering the easy increase in CFM from the injector putty, I went slightly mad and decided to see what would happen if you really raised the inlet port entry:

Attachment 228162

Then the result of mass removal of aluminium:

Attachment 228163

I removed a lot, gained an additional 3-4mm of port entry height although I have significantly increased the port CSA as well at this point. So with re-adding the injector port putty I flow testing the head again:

0.1 - 68
0.15 - 104
0.2 - 132
0.25 - 157
0.3 - 182
0.35 - 192
0.4 - 199
0.45 - 205
0.5 - 209

So this basically killed the low lift flow of the head but assisted at the top end (Although only above 0.4 lift which is my max cam lift at 10mm). So without replacing the 3-4mm of material on the port floor I was at a loss with this move, although it did show that raising the roof of the port was improving the flow at the high lift areas.

At this point, I decided to go for a full port on the head, and then rough cut in, awaiting clean up. If you use the valve guide as a reference for how much is being removed:

Attachment 228164

Attachment 228165

So after a quick clean up/blend pass, I got the following results:

0.1 - 81
0.15 - 113
0.2 - 141
0.25 - 165
0.3 -185
0.35 - 195
0.4 - 205
0.45 - 210
0.5 - 217

This setup seems to be a good balance between increased port size but retaining low lift flow. this is also without any additional putty in the port to recover the 3-4mm port entry so I knew I could find some additional low lift flow by filling that area in/shaping it accordingly., maybe even more additional flow across the test sweep? who knows!

I'm really happy with these results, mainly because this is using the stock valve seat angles and stock valves, just shows how much of a good increase can be achieved, but I reckon there is a lot more available - I just need to find it!

Here's a nice little excel graph which shows the difference between the best stock results and the full port results

Attachment 228166

So I approached a local engineering/engine shop near to me, Lockside Engineering and started to talk about my goals/what I wanted to achieve. They came on board with me and we began the process of choosing the valve seat width & 3 angles for the inlet side and backcutting the inlet valves for testing on my flow bench!

The two options I went for:

1.1mm valve seat and 60/45/39 angles

Attachment 228167

1mm valve seat width and 60/45/38 angles

Attachment 228168

Without going into too much detail, I chose very similar angles because I wanted to see how much gain would be seen by going to a thinner valve seat (But still within OEM Manual limits). I could have gone thinner than 1mm but then reliability/valve sealing comes into question which is something I wasn't interested in answering!

The inlet valves were back cut at 30 degrees on both tests.

So here are the results:

Attachment 228169

Well hello low lift flow gains! As expected the 1mm valve seat width has allowed for a large increase in low lift flow compared to the 1.1mm valve seat. More so than I expected if I'm honest, but I wanted to see how much of a trade off would be to have the extra re-assurance/durability of the additional seat width vs more performance with the smallest width seat.

The angles I chose were based on a lot of research, they were the most 'accepted' standard for improving low lift flow with heads with a poor short side radius. If I'm honest, there will be additional gains by using different angles I'm sure of it, but at a cost of £40 to test each 3 angle combination, I had to take a calculated best guess before choosing them! I wish I had the money/time to test out some other theories I have from my research but I don't!

Studying the above graph, it shows that the port itself is the limitation beyond 0.35 lift (above 8.9mm lift) which makes a lot of sense considering the max OEM cam valve lift is around 9mm lift for the BP4W head if my memory serves me correctly.

So did he ever build a head with this information and dyno it? While I know little about the technical side of headwork or porting, Emilio's quote here:
Makes me wonder if all this effort was worth it.

AFAIK no, but if an update pops up, I will repost it here. I know that as-is it is only half the picture, but it was an interesting read for me nonetheless.

In this particular example, the BP4W head's intake-ports have considerable casting-flash. I wonder what improvement would be netted, if the port geometry were to be left essentially untouched, but the casting-flash were to be ground-down, and the wall surface-roughness reduced (perhaps not all of the way to some "mirror" finish).

Wow some cool info on this thread! I have one question though. What is the angle of the flange on the intake side with respect to the vertical? It slightly faces downward and was wondering what the angle was. Thanks