Zed.

Looks good, might follow youre lead with this...

look at some of the best pannel-beaters, use anything to hand

as an apprentace I made a louver-tool from a block of hardwood & used a router to, well, rout a groove (slot?) in it
place the block under the metal to be louved & using another piece of wood / hammer tap the metal into the groove (the edge of the metal will sit against the side of the slot so keping things square)

^^ I'm dislexic so hold no responsibility for spelling

Rich.

Tchaps

Ha ha feel free. One thing the tool doesn't do well is the first louvre in a row, it doesn't really form the leading bend line well. But when you are on the second + louvre in a row you don't have that leading bend line, you are only forming the sides really and it does that well and pretty uniformly.

I think the block of wood tool would be superior, just think you could get a more uniform and crisp form line but you need full access from both sides.

Hopefully it comes out good with a bit of final finishing. The silver doesn't hide anything so need to pay attention, but fingers crossed!

Tchaps

I've switched up cam positions again, this time retarding to my original intake max lift position of 114ATDC, but dialling in more overlap by retarding the exhaust max lift to 108BTDC.

The logic behind this, apart from that I havent tried it yet, is that I hope to keep the power climbing all the way to the limiter, and see if the increase in overlap makes better use of the increase in power from 6-6.7k that in my very optimistic world is being caused by my exhaust when I run slightly more overlap. We shall see.

After more and more research into supercharger specific cylinder head setup, this later intake valve closing and later exhaust opening timing seems to be recognised as a good setup on a supercharged car. The caveat being it places more emphasis on exhaust port flow as there is less time for exhaust blowdown once the valve is opened. Mx5 / Miata seem to have quite a high intake to exhaust valve size bias anyway, and I did smooth out the ports on my current head, so hopefully I'm in a good position there. Likely that gains can be found with more attention to seat angles and that will be the main focus exhaust side on the next head.

Speaking of the next head, the hoover should be arriving in the next few days, and I have decided that I'm not going to change cams. They would be the most expensive part of the whole head project by far, and I think there is a high chance of me moving the power band too high and too far over the rpm limit of the supercharger. I think the current peak tq and hp rpms are perfect for my sc pulley ratio, so the effort on the head will go into developing the seats and short side radius, maintaining pretty much the current port cross sectional area to ensure I don't drop velocity, and just work on flow efficiency. Also if it is my header increasing tq between 6-6.7k that is perfect, right between peak tq and hp to give me a nice meaty power band from 5krpm to 7.5krpm, so I don't want to end up moving where the head makes power outside of that window.

Tchaps

So I've been out and tested the retarded 114 int / 108 ex settings and there are some semi interesting results.




So red is the 114 / 108 medium overlap, blue is the other retarded intake 114 114 low overlap that I started with, and it shows that in the very top rpms the retarded intake cam keeps making power right up to redline, whereas the advanced intake cam runs power is rolling over here.

Through the midrange its a bit harder to make any particular conclusions if I'm honest. The green advanced high overlap seems to win through 6-6.7k.

Through peak torque 5-6krpm the blue 114 114 low overlap wins out.

I must apologise - I went through all my dyno runs yesterday to clear out and just keep the good ones and when I have loaded back in the 104 118 run it did not make the power it did in my previous graphs. I think I hadn't selected the correct profile in virtual dyno, because when I did for this graph it was lagging. Part of the reason I did the tidy up because I just had too much data to keep track of properly.

The other point of note really is that as I've said before the advanced intake results in leaner AFRs and also a drop in manifold pressure suggesting an increase in VE, it just doesn't necessarily translate directly into a gain in power across a wide rev band like the above suggests. I think this is to do with the compromise you then make on the exhaust valve opening (advanced) or running high overlap so you can retard the valve opening, which only seems to win in a specific part of the rev range (6-6.7krpm).

I think my results show that opening the exhaust cam too early is detrimental at high rpms, see difference between orange and green traces, orange opens the exhaust valve 8 crank deg earlier than green.

Overlap seems to be detrimental at peak torque, possibly over scavenging from the exhaust and blowing a bit of charge out the exhaust rather than keeping it all in the cylinder? However as per above it looks like it could benefit at 6-6.7krpm, maybe allowing the exhaust to scavenge and help the VE above peak torque.

I think my next test will be to keep the intake cam where it is and retard the exhaust another couple of degrees to increase overlap further, and see if this results in a power gain 6-6.7k and also see what it does to power around peak torque.

If it results in a better top end (6krpm+) but drops the 5-6krpm performance over the reduced overlap 114 114 timing its really then just a choice of what torque curve you want really. Sacrifice a little low end for the high end or a little high end for the low end.

I'll see what the curves look like when I have the data.

Tchaps

Still testing out a bit more overlap, 114 intake 106 exhaust.

In the meantime here's a little video of the choppy idle with the overlap dialled in, listen with headphones in because on laptop speakers you loose the bass of the exhaust.

It keeps me amused at traffic lights. Not sure I am mature enough to back to low overlap if it looks to be the best performance

Tchaps

Bit more work on the louvres today. I decided I was going to do a set for both sides of the hump, one for the headers and one for the supercharger. Figured it would look best symmetrical, the more venting the lower the underhood temps and there isn't anything vulnerable to water underneath the vent positions. I'm very happy with how these sets came out. Definitely more even than the ones at the front of the bonnet. Bit of finish filing and a bit of paint repair and they'll be perfectly presentable I think.

Tchaps

I managed to get a decent run in with the low overlap. It changes the fuelling tune quite a lot, cruise runs leaner with low overlap, but then runs richer in boost.




But in terms of power there doesn't seem to be anything in it, my theory is I'm in some mid zone where the pay off of the low overlap (increased mixture capture, earlier EVO for extended exhaust blow down) is balanced by the advantage of the increase in overlap (later EVO, longer power stroke)

Online on what seem to be reputable sources of information I have seen both early and late EVOs recommended for supercharger applications. I think its just one of these things you need to try out for yourself, for your engine and tune and refine based on your own data.

I'm going to stick my cams back to the original 114 / 114 and get another run with the increased rev limit at some point.

I don't think it has been a waste of time, I found that the advanced intake cam was not good for the top end (and I could feel that in the car, it felt less eager right at the top) and that my retarded, later closing intake was better for top end without any negatives through the midrange.

EVO and overlap findings are less solid. Maybe I would find gains with a shorter duration intake cam to reduce overlap while keeping the later EVO. I think this would be a good experiment.

Anyway, it is making excellent power. With the little bump in rpm and cam fettling up into the top end reliably into the 285+whp zone, and if VD could handle the last few hundred rpm up to the soft limiter of 7600 we should see 290whp. It just seems to clip them short or if I crop the log as the limiter comes in it screws up the result. It pulls like a train up to the limiter and we seem to have found a few more top end ponies without losing any from elsewhere so mission successful.

Tchaps

I had a look in the back of the supercharger yesterday and all seems to be clear from debris after the last month and a half of driving. Oil in the back of the supercharger is also a lot less, just a sort of film rather than pooling, so really happy with that.

I also scoped the bores because I was feeling a little nervous from all the on boost lean events I had getting the car back up and running but there doesn't seem to be any pitting and the edges of the pistons all look nice and sharp.

The reason I was checking again was I had noticed some puffs of smoke on first acceleration after a long stint of vacuum again. When I had the supercharger off last time I did notice a bit of oil on the stem of cyl 4 intake valves, in comparison to the others. When I bore scoped yesterday cyl 4 looked to be slightly wet / oily in comparison to the others, so I think / hope the puffing is the cyl 4 valve seals.

Got some more seals ordered and will tackle the fun job of replacing them before the next trackday in July. Hopefully its not a valve stem clearance issue and the seals get me by until I sort the new head out.

Tchaps

This morning I looked back on my old dyno sheets from the two actual dyno sessions I did last summer, the first one where I was on the low rpm 4 rib pulleys then the second session with the pulleys I had made.

Just double checked and reloaded that old excel graph I did ages ago to show the comparisons between the two days. I then took the best exhaust only run I have recorded approx 275whp, and then the latest best run with WI on, which is 286whp, and added them to the graph too.



I couldn't get a trendline that fits torque curves nicely but basically the red squares are original 'small' 4rib pulley torque values, black squares are the 'big' 8 rib pulley torques. Up to 6000rpm big pulleys show a decent torque increase as you would expect, but then they close to virtually the same at 6700rpm, to produce the peak whp and the mystery of the missing power from last summer.

Then you can see the new torque values with the new exhaust and WI installed lifting power above 6000rpm, to what I expected to see last summer when I installed the new pulleys.

As I said some of the trendlines don't fit well but I left most on as I thought they helped the eye sort out an otherwise fairly hard to read graph.I am pretty happy with how closely my figures gained from virtual dyno tally up to the lower rpm figures from the actual dyno, if anything the the WI on torque (upgrades) is a little low in the beginning. I am confident as I can be that my virtual dyno figures are representative and would correspond with what the actual dyno would record.

I just wanted to put all the stages of power on one graph for easy(ish) consumption, hopefully it helps others in the future if they bump into a similar power cap that I did.

I think its funny though, I was so annoyed not to get my big pulleys for that first dyno session, but if I had had them I probably wouldn't have ended up with the figures I have now. I would have thought, 255whp, eh that's about right and called it a day. Its only because I had already made basically the same peak power with 4psi less manifold pressure that I figured there must be more on the table somewhere.

Hopefully I can get it to 300whp at some point, maybe a bump in comp and some flow bench work with the updated head will chip me over.

Tchaps

I've been doing a bit of forum searching and also going on something Warpspeed said on Miata.net with regards to detonation threshold on a supercharged engine vs a turbo engine.

Looks like the reduced backpressure due to lack of exhaust compressor enables hot spent gasses to be more quickly evacuated away from the cylinder and therefore cylinder temperatures are lower and detonation risk lessened. Makes sense really.

Mild boost Rotrex setups (10psi area) were discussed being able to achieve MBT, and I am not a million miles away from that, only a few psi higher.

The findings from the dyno tuning last year should have told me this but I didn't have the experience to interpret this at the time. Adding 4deg and not achieving any visible increase in torque while also not knocking should have really told me this but there we go. Not saying I am right on MBT currently but I think I am near enough within the sort of accuracy I can measure.

Then it also goes and explains the results with the WI, and the lack of additional torque from timing. So plan for the moment is I will just run the WI switching fuel only and not timing, regain another 3 degrees knock buffer for track sessions. No point adding timing if it doesn't do anything.

Longer term I guess there are two choices; when I do the head either just give it a mild skim to say 10:1 comp and keep it pretty safe, still really only using WI for a bit of safety and a small torque bump or go all out, high comp pistons to say 11.5-12:1 and run it as a proper all out water injection build. Rough maths would certainly land me over 300whp by going up to that higher comp, the system with failsafe is there, and with the 10L tank and low water fuel ratios it lasts a long time. I don't think it would be too much of a pain to live with it's just more risk vs really not a huge leap in power. Have to have a think.

Anyway I kind of feel like the 'mysteries' ('missing' power of the pulleys and timing vs torque insensitivity) of the build have all been answered now, just took me a while to clock on

Hopefully others find that helpful / interesting, I certainly have.

3rdCarMX5

iTrader: (1)

Just wanted to stop by and say I have been following along and you have done a great job, and I appreciate your methodical updates. You have achieved what my goals are (were) for my similar setup.

Any chance you would share your msq? I'm really interested in your Accel enrichment, timing map, and general idle settings.

Tchaps

Thanks mate, yes knock yourself out, see attached.

I biased accel enrichment slightly towards tpsdot rather than mapdot in the end, it removed a small mild lean spike on tip in. Map builds so quickly it seems to be better to lead slightly with TPS and try to be ahead of the game.

Spark map 1 is non WI, spark map 3 is WI. WI is only operational 3000rpm (if its quick enough) up and above 150kpa

Idle settings are just a simple open loop warmup for the valve then I use the idle advance control to precisely control idle speed. Works really well I find, easy to tune out rpm droop etc. I tried closed loop for the valve but it didn't seem to play nicely with the idle advance at the time. The setup I currently use works so well I didn't put any more development into closed loop to solve the weirdness, I just went back to the setup that worked. Low overlap the car idles at 50kpa and with higher overlap it idles at 55-60kpa. Its quite high because of the draw back of a twin screw, always compressing always drawing power so the engine is throttled open a bit to sustain idle. One of the quirks.

3rdCarMX5

iTrader: (1)

Thanks, I'll download and look around! The tpsdot approach to tip in makes sense to me. I was trying to get fancy and I was running into transitional knock.

Spaceman Spiff

A reference tune will be super nice to have on my end as well regardless of ecu, thanks man! The impractical coldside sc club owes you one.

Tchaps

No problem, yes I certainly saw an improvement in my logs, although I never (as far as I am aware) transitional knock previously.

Tchaps

No problem matey, ah what a fine club it is too! Need to come up with an emblem of some description, maybe featuring some scratched knuckles and lost fuel rail spacers.

Just keep in mind a couple of things on the timing tables - table 1 is well proven under extended periods of load now on my setup with a chargecooler, with a reasonably rich mixture (11-11.4AFR) without WI and with WI at a leaner 12:1 mixture.

Timing table 3 has been fine under road pull conditions with WI on and at fuel mixtures up to 12.5-12.7 (the leanest I have gone by accident while testing cams). I have had the timing up a couple of degrees higher than these current settings without knock. But I haven't tracked the car with these spark timings. As discussed it doesn't seem to make any difference over table 1 anyway at my boost level (1bar) so I will probably be turning spark table switching off and just enjoy the great big fat knock buffer on track.

Subtext is approach with caution, but you knew that anyway

Spaceman Spiff

I'm game, who knows, maybe we'll be more accepted when the site formally becomes miatakswap.net, but If we get logos you'll need a special one for actually doing it correctly (you know, with one of those magical things they call a charge cooler).

The handlebar mustache that "designed" the kit on my car spent a ton of money doing some things right (e.g. cast intake manifold, cnc aluminum [aluminium] tensioner bracket), and then just called a wet fart good enough for most the rest. E85 and water/meth are going to have to do my cooling for me 🤦‍♂️.

In any case, yeah I just wanted something to compare to and make sure the Haltech base map wasn't totally off the mark. Thanks again!

Tchaps

Well us Brits like an underdog just as much as we like saying aluminium correctly a nice badge or a special hat would be much appreciated!

Tbf the 300+hp car Tom made, that was on E85 and as far as I can tell we make almost exactly the same peak hp and very similar tq curves.

On his youtube video the dyno plot shows 310-315crank hp, from my virtual dynos I must be making very similar figures at the engine but I am adding a bit of water, so I'm not doing it with only petrol. I make about 10hp less on petrol only.

I think he is using the TVS900 on that car - which has pretty much the same capacity as my charger, his only 0.03L bigger per rev, naff all. Not sure on the pulley ratio he is using on it though, I think you can spin a TVS a bit faster than I can spin my Autorotor, but not by much.

I just find it interesting because it's an obvious comparison, we've gone about it different ways but made to all intents and purposes the same ultimate power.

If you look at petrol only performance, i.e detonation limited (?maybe not limited on my setup?) I certainly think the chargecooler wins out - I'm not aware of any non-intercooled supercharger setups on these engines making 270-275whp on petrol only, although I will eat my new hat if one comes up.

No problem on the maps hope they are vaguely useful.

Spaceman Spiff

You guys might be making the same power but I'd put my money on his eating the piston rings on #4 by the time he's done with a track day.

Eric Anderson's SSM build is a good example of both a great build and the slight packaging jenkiness that normally has to happen to accommodate a charge cooler. Your packaging is the closest to an OEM V engine like setup I've seen (which seems to be preferable).

https://www.miataturbo.net/race-prep...m-miata-62672/

Tchaps

Well I'll take the compliment, thanks I was lucky in finding both a compact supercharger and chargecooler in terms of width, which really helped the packaging,

Uh don't talk about rings. I'm using a bit of oil, smoking a little bit on cold start, particularly if you give it a little rev. It clears up on hot oil but sometimes smokes a bit when you give it beans after sitting in traffic for a bit.

I had a look down the bores and I couldn't see anything horrible going on with the pistons, but piston 4 was a little wet on top. All other pistons were dry. This was after the car had sat for a couple of days.

I think atleast intake valve seals on cyl 4 are leaking, and actually when I last installed the supercharger I noticed that these valves were a little wet with oil.

So symptoms all suggest valve seals.

Hopefully it is just seals - I've got some OEM ones to go in - and there's no need for me to pull the head to do that.

Valve stem to guide tolerance was all in spec when I built the head about 3 years ago, not sure if Mx5/Miatas are prolific valve guide eaters? The direct nature of the valve train suggests to me they aren't but I am of limited experience from only this car. I guess I'll do a calibrated wiggle test when I'm doing the seals.

Hopefully its not guides or rings as that would require a more invasive adventure and I have some events coming up pretty soon.

I didn't realise Eric was A/A cooled coldside, and I haven't read his build thread so thanks, should keep me entertained for an evening