Pete's RX-8 Time Attack project

For you to showcase what you've done to your car and give other members ideas and inspiration.
njkmr
Club Member
Club Member
Posts: 12644
Joined: Mon Oct 30, 2017 7:08 pm
RX-8: Kuro
Colour: Kuro Black
Location: Lancs
Has thanked: 234 times
Been thanked: 223 times

Re: Pete's RX-8 Time Attack project

Post by njkmr »

Just read this from start to finish.
It was a bit like gorging on a "box set" on TV but once you start reading it your hooked...! :xmasthumright:
Fascinating journey and very informative, although like some others have said , I don't get all of the techy stuff but the detail is fantastic all the same .
Looking forward to the next "episodes"
Keep up the good work.
Regards
Rob.
PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

Thanks Rob. :xmasthumleft: . I can't believe this is 17 pages already.

Brace yourselves for the next instalment. It's all about flame front speeds, and why torque drops away as the red line approaches....... :-k :xmasgrin:
These users thanked the author PeteH for the post (total 2):
13Black (Mon Dec 18, 2017 9:47 pm) • njkmr (Mon Dec 18, 2017 9:49 pm)
User avatar
brix79
Forum Administrator
Forum Administrator
Posts: 11009
Joined: Mon Sep 03, 2012 9:51 pm
RX-8: R3
Colour: Sparkling Black
Location: Caerphilly
Has thanked: 133 times
Been thanked: 446 times

Re: Pete's RX-8 Time Attack project

Post by brix79 »

Episodes. Now there's an idea. We want video evidence Pete! :xmascool:
John
Forum Admin Team

2021- // 2009 Sparkling Black R3 | Pettit Exhaust | 20mm Eibachs

2017-21 // 2010 Aurora Blue R3 | Gone, never to return, never to be forgotten.
2011-17 // 2005 Titanium Grey 231 | Gone. But Lives(d?) on.
PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

Next episode:!

The new engine comes without an ECU, so I need to find a stand-alone ecu to run it. The choices are many: Adaptronic, Haltech, Microtech, Link, DTAFast, Motec, AEM.

If anyone has any specific experience I would be grateful to hear your experiences.

Anyway, whichever I choose I will need to write my own base map. I have been researching, and doing sums, and it's all been very interesting. This post will describe the basics behind the ignition map.

Ignition timing is all about trying to fire the mixture at a time that creates the most torque. That basically means firing the mixture so that as much of the expansion stroke as possible is taking place with a fully burned mixture, so that the gas can do the maximum possible work on the rotor. In an ideal world you would instantly generate all the heat (and therefore pressure) at a fraction of a degree past the Top Dead Centre (TDC) point, and let all that pressure work on the rotor for the entire period of the expansion stroke, before exhausting the gas after BDC. In reality you can't burn the mixture instantly, it takes a finite amount of time, so the general rule for piston engines is that you aim to have your combustion fully completed by about 20degrees after TDC (ATDC). The equivalent e-shaft angle in a rotary engine is 30deg ATDC. This is the first important number to remember.

The next important aspect is the speed at which combustion takes place. This is referred to as the "flame front speed", and is not easy to establish. The literature is sparse, and usually refers to low speed single cylinder test engines in labs. But values of between about 10m/s and 35m/s are typical for a turbulent mixture. I will use a value in the middle of this range, 24m/s, for reasons that will become obvious. This is the next important number. It should also be said that the flame front speed slows down when the mixture is weak, or low density (i.e. on part throttle). The flame front speed also speeds up when mixture turbulence is higher.

Next we need to know how far the flame front has to travel. 13B rotors are 80mm wide, and about 180mm long, but they have two spark plugs arranged approximately in the centre of "their half" of the rotor. So the maximum distance the flame has to travel is approximately 40mm to 45mm. So we now know the speed of the flame, and the distance it has to travel, so we know the time it will take to reach full combustion: 0.04m / 24m/s = 0.002s.

We now know how long we have for combustion, and we know it must be done before 30deg ATDC, so it is now simple maths to calculate when we need to initiate combustion. The answer is:
Ignition Map.jpg
Ignition Map.jpg (24.1 KiB) Viewed 1411 times
I have put the theoretical ignition advance, and the actual RX-8 map on the plot, and they match very well. There are really good reasons for the parts of the curve when they don't match:
- At low engine speeds the mixture is lower density (poor volumetric efficiency, and heat loss), and the combustion speed is slower than 24m/s, so the timing cannot be as retarded as my calcs suggest. But you don't need to retard that far anyway, since having full combustion closer to TDC is actually a good thing, if you have the time, which you do at slow engine speeds. In my final calc sheet I will adjust my timing maps based on rough corrections for flame speeds, and optimum firing.
- The timing works really well for the peak torque value (at 5500rpm). At peak torque the timing can be set based purely on physics. Everything works nicely, and there is plenty of time for ideal combustion. This is why peak torque is peak torque!
- Beyond peak torque the real ignition map cannot advance as far as the ideal theoretical timing suggests. If you tried to run the theoretically optimum timing at high revs then you would initiate combustion well before TDC, and the mixture burn would become unstable, as the burn would be taking place while the mixture is still compressing. The sudden increase in pressure and temperature causes a run-away situation, and the mixture self-ignites at all places at once. This is called detonation, and rotaries must never detonate. So to avoid this the ignition map is held at almost the same advance, all the way from peak torque to peak power. This is good for avoiding detonation, but does mean the mixture is not fully burned by 30deg ATDC. In fact at peak revs the mixture may not be fully burned until about 70deg ATDC, so the gas cannot do its maximum work against the rotor, and torque is lost. This is a primary reason why torque drops at higher revs, and eventually the power starts to drop as well.

So it should be clear now why there is power to be gained by advancing the ignition timing of a rotary (or any engine), but why doing so brings a greater risk of detonation destroying your apex seals.

A final note: This timing map is only considering the full throttle condition. At part throttle the mixture can be very weak, and that necessitates a much more advanced spark. A proper ignition map is a 2D table that considers both engine speed and engine load. This is when things get really complicated!

I think I'm happy that I can now make a base map for both ignition timing, and fuel injector timing. :xmasthumleft:
These users thanked the author PeteH for the post:
warpc0il (Tue Dec 19, 2017 1:51 pm)
User avatar
Clive
Sponsor Trader
Sponsor Trader
Posts: 1712
Joined: Sun Jan 30, 2005 6:39 pm
RX-8: Supercharged
Colour: Lightning Yellow
Has thanked: 449 times
Been thanked: 1054 times

Re: Pete's RX-8 Time Attack project

Post by Clive »

May help to know unlike piston engines every one I know has never found gains by playing with ignition timing on Rotary's and I am talking about some big 800 BHP plus Rotarys.

Leaving the timing alone takes away a huge variable, because as you will know every change, even a small one, means the air fuel ratio has to change relative to where the spark is performing.

Choice of ECU / don't forget the Adaptronic will keep all your RX-8 chassis functions if you want them. I don't think there is another ECU that will integrate totally with the RX-8 chassis.
REgards
Clive


http://www.racingbeateurope.com

We aim to help and assist our customers make the right choices for their Rotary, whether they be standard or modified. If you need further assistance, please do not hesitate to contact us:-
sales@racingbeateurope.com or call 01474 777 800
PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

Thanks Clive. I am planning to stay close to the RX8 timing maps. I don't want to take risks with the timing. I just thought it was an interesting technical exercise to work out the physics behind it.

I agree about the Adaptronic, and have been talking to the manufacturer to make sure it can support my 20B (It can). I am resigned to losing some functionality, but would like to keep PAS and TC. But if a compatible ecu is offered with a really good sponsorship deal then I would look at that instead.
User avatar
Clive
Sponsor Trader
Sponsor Trader
Posts: 1712
Joined: Sun Jan 30, 2005 6:39 pm
RX-8: Supercharged
Colour: Lightning Yellow
Has thanked: 449 times
Been thanked: 1054 times

Re: Pete's RX-8 Time Attack project

Post by Clive »

As I have dealt with Adaptronic direct for some four years now, PM with the guys name to make sure your dealing with the correct person.
Providing you have a 231 Chassis running a tripple wont be a problem as you will have enough injector feeds also, steering, dash DSC
etc will also be functional using the Adapronic ECU
REgards
Clive


http://www.racingbeateurope.com

We aim to help and assist our customers make the right choices for their Rotary, whether they be standard or modified. If you need further assistance, please do not hesitate to contact us:-
sales@racingbeateurope.com or call 01474 777 800
350matt
Club Member
Club Member
Posts: 3158
Joined: Mon May 15, 2017 9:09 pm
RX-8: 231
Colour: Strato Blue
Location: Northants
Has thanked: 49 times
Been thanked: 587 times

Re: Pete's RX-8 Time Attack project

Post by 350matt »

I think that Motec have quite a bit of experience with rotaries as they were 1 of the 1st aftermarket ECU's setup to cope with rotaries and the UK office is very approachable
plus there's the Cosworth Sq6 which should also cope
but if Adaptronic can provide a base map then this will save a lot of time

how are planning on mapping it Pete? on the rollers then tweak on track based on log data?
If the Adaptronic has fuel trims that can be applied to each rotor then it would be worth putting a lambda boss on each primary so you can dial in the fuelling just so on full throttle
you will often pick up 2-3% on a piston engine by doing this and lot more if the inlet manifold is a large one with a single entry
if the ECU also has ignition trims then this too is worth playing with but its trickier to dial in without dynamic pressure measurement
how many knock sensors can the ECU cope with? as if you can run 1 per rotor that'd be splendid as then you can setup the knock control to pull back individual spark trims to squeeze the most out of it plus also provide the best protection to the engine
PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

Thanks Matt. It's a useful discussion. I'm still learning. My current thought are:
350matt wrote:
Wed Dec 20, 2017 9:31 am
I think that Motec have quite a bit of experience with rotaries as they were 1 of the 1st aftermarket ECU's setup to cope with rotaries and the UK office is very approachable
Motec have a great name, and I know it can run a rotary, but they are expensive and poorly featured compared to others. I havn't ruled them out, but more rotary-focussed ecus currently look a better bet.
350matt wrote:
Wed Dec 20, 2017 9:31 am
plus there's the Cosworth Sq6 which should also cope
but if Adaptronic can provide a base map then this will save a lot of time
I will look it up, but I suspect the same comments as for the Motec will apply. I'm not expecting a base map from anyone. I will write my own.
350matt wrote:
Wed Dec 20, 2017 9:31 am
how are planning on mapping it Pete? on the rollers then tweak on track based on log data?
Hopefully 99% done on the dyno. EuroSpec are a massive plus for me, so dyno time isn't a problem.
350matt wrote:
Wed Dec 20, 2017 9:31 am
If the Adaptronic has fuel trims that can be applied to each rotor then it would be worth putting a lambda boss on each primary so you can dial in the fuelling just so on full throttle
you will often pick up 2-3% on a piston engine by doing this and lot more if the inlet manifold is a large one with a single entry
Multiple wideband lambda gets expensive! If I had individual throttle bodies then I would consider it, but I don't, so I'm sure I'll tune with only one. I'm currently planning to run the car open loop, so probably won't even run a lambda on the track. This is Club racing, so I can't approach it in the same way as the pros.
350matt wrote:
Wed Dec 20, 2017 9:31 am
if the ECU also has ignition trims then this too is worth playing with but its trickier to dial in without dynamic pressure measurement
how many knock sensors can the ECU cope with? as if you can run 1 per rotor that'd be splendid as then you can setup the knock control to pull back individual spark trims to squeeze the most out of it plus also provide the best protection to the engine
I'm just playing safe with ignition. The gains from optimising are small, and the danger of blowing the engine with detonation are large. Knock sensors could be useful, but they need to be tuned to recognise knock. That means making the engine knock, and that means likely destroying a 20B! I don't think anyone is keen, from the conversations I've had so far.
User avatar
Ryan Rotary Performance
Former Trader
Former Trader
Posts: 1616
Joined: Fri Jan 27, 2017 8:06 pm
RX-8: Supercharged
Colour: Brilliant Black
Location: Rugby and Stonham Aspal
Has thanked: 69 times
Been thanked: 226 times

Re: Pete's RX-8 Time Attack project

Post by Ryan Rotary Performance »

20B!! excellent news here Pete and very exciting!!! I take it you already have the engine ready to go?

As for ignition timings I concur with the general consensus that Mazda has already done most of the hard work for you, you can advance quite a bit lower down the rev range but up top I tend to be quite conservative with what I do. Besides the lower ranges will not concern you that much.

As for the ECU, I will be a supplier for another company soon which is neither Adaptronic or Motec, their system is very much the same as Adaptronic and I'm yet to bottom out prices and support, some news will be released soon.


Good luck with this one, exciting stuff!
These users thanked the author Ryan Rotary Performance for the post:
PeteH (Wed Dec 20, 2017 11:16 am)
User avatar
warpc0il
Spin Doctor
Spin Doctor
Posts: 32531
Joined: Sat Sep 06, 2008 10:56 pm
RX-8: 231
Colour: Lightning Yellow
Location: Groomsport, Co Down, NI
Has thanked: 826 times
Been thanked: 3537 times
Contact:

Re: Pete's RX-8 Time Attack project

Post by warpc0il »

Have you considered Syvecs?
Dave
The Spin Doctor ™
uǝǝɹɔs ɹnoʎ ʇɹǝʌuı ǝsɐǝld :ɹoɹɹǝ
PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

Thanks Carl. Yes, the engine is at EuroSpec, and we'll start working out the installation in January, when I take the car down. I'll be interested to see what you end up doing with ecus.

Dave. Yes, I have had a quick look at Syvecs. EuroSpec have been doing a lot with them recently (their Civic Type R was a development mule for a plug and play ecu). They are on the list, but I'm not sure how rotary friendly it is.
350matt
Club Member
Club Member
Posts: 3158
Joined: Mon May 15, 2017 9:09 pm
RX-8: 231
Colour: Strato Blue
Location: Northants
Has thanked: 49 times
Been thanked: 587 times

Re: Pete's RX-8 Time Attack project

Post by 350matt »

When looking for an ECU I'd 1st check out their I/ O list to check all the things you want to drive can be driven ( this isn't usually an issue these days but it dosen't hurt to ask)

the big one for the rotary is running the twin coil setup and ensuring that the ECU processor can run fast enough to deliver 2 ignition events at high engine speed as a lot of ECU's struggle past 10,000rpm on a conventional 12 cylinder engine ie with 6 firing intervals per crank rotation
the rotaryness of the 20B means its a similar case especially as some ECU's will treat a 2nd spark event as another cylinder firing

so a 13B is already 12 sparks / rev a 20 B is 18 sparks / rev?
PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

350matt wrote:
Wed Dec 20, 2017 12:44 pm
so a 13B is already 12 sparks / rev a 20 B is 18 sparks / rev?
No. They are all one spark per rev per coil. So a 13B is 4 sparks per rev, and a 20B is 6 sparks per rev.
User avatar
Dr. FrankenRex
Committee Member
Committee Member
Posts: 8169
Joined: Thu Jan 08, 2015 1:30 pm
RX-8: R3
Colour: Velocity Red
Location: Cotswolds
Has thanked: 531 times
Been thanked: 1558 times
Contact:

Re: Pete's RX-8 Time Attack project

Post by Dr. FrankenRex »

Potentially stupid question, but how does that work?

I.e. does one 'rev' count as one combustion or does it count as a full revolution of the eccentric shaft? If the latter then you've actually got 3 ignition cycles on each rotor don't you? So 6 sparks per rotor, or 12 across both as Matt350 says?

Interested to hear an explanation as I'm now confused! :lol:
DOC 787B

Club Information Coordinator - Media Team Lead


LM30 VelociBRAPTOR - Josh Scott magic sauce porting, shiny big brakes and a big ole spoiler

Doc's Auto Barn - https://www.youtube.com/channel/UC7BW_h ... Ym6eGqEIWg
PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

Rotors spin at one third e-shaft speed. I thought that was common knowledge :?: One bang per rev per rotor. Each bang has to take place at the same e-shaft angle in order for the "crank" to be at the right angle to generate torque. Nothing else would work. You can design four, five, etc. chamber rotors, but you then have to gear the rotors down by four, five, etc.

And, to clarify, your rev counter reads the e-shaft rotation speed, not the rotor rotation speed.

Peak rotor rotation speed is about 3000rpm. People don't think 4kg of rotor is spinning at 9000rpm, do they? :xmaseek:
User avatar
Dr. FrankenRex
Committee Member
Committee Member
Posts: 8169
Joined: Thu Jan 08, 2015 1:30 pm
RX-8: R3
Colour: Velocity Red
Location: Cotswolds
Has thanked: 531 times
Been thanked: 1558 times
Contact:

Re: Pete's RX-8 Time Attack project

Post by Dr. FrankenRex »

I did know that the rotor spun at 1/3 the e-shaft speed, but I thought there was still 3x ignition cycles per full rotor rotation. Like a dis:

Image
DOC 787B

Club Information Coordinator - Media Team Lead


LM30 VelociBRAPTOR - Josh Scott magic sauce porting, shiny big brakes and a big ole spoiler

Doc's Auto Barn - https://www.youtube.com/channel/UC7BW_h ... Ym6eGqEIWg
User avatar
Harrizone
Founder Member
Founder Member
Posts: 7123
Joined: Fri Oct 31, 2003 5:23 pm
RX-8: 231
Colour: Winning Blue
Location: Southend-on-Sea
Has thanked: 42 times
Been thanked: 361 times

Re: Pete's RX-8 Time Attack project

Post by Harrizone »

PeteH wrote:
Wed Dec 20, 2017 2:58 pm
People don't think 4kg of rotor is spinning at 9000rpm, do they? :xmaseek:
Could have quite a gyroscopic effect to keep the car 'balanced'. :-k :lol:
Now a distant memory unfortunately.
PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

I'm losing track of the question....?

One rpm on a rotary means one rpm of the e-shaft, and 1/3 rotation of the rotor. Matt's concern was about firing 18 times per rev at 10,000rpm. But this doesn't happen. It fires 3 times per rev at 10,000rpm on the leading, and 3 times per rev on the trailing. The same firing frequency as a 6 cylinder twin spark engine doing 10,000rpm.

I think we are all agreeing. It's plenty of sparks, but should be totally fine for a rotary specific ecu.
PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

Dr. FrankenRex wrote:
Wed Dec 20, 2017 2:26 pm
does one 'rev' count as one combustion or does it count as a full revolution of the eccentric shaft?
Same thing.......
PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

Harrizone wrote:
Wed Dec 20, 2017 3:07 pm
PeteH wrote:
Wed Dec 20, 2017 2:58 pm
People don't think 4kg of rotor is spinning at 9000rpm, do they? :xmaseek:
Could have quite a gyroscopic effect to keep the car 'balanced'. :-k :lol:
Perhaps, but I think seeing your e-shaft disappear through your front wing could be quite "de-stabilising"!
User avatar
Dr. FrankenRex
Committee Member
Committee Member
Posts: 8169
Joined: Thu Jan 08, 2015 1:30 pm
RX-8: R3
Colour: Velocity Red
Location: Cotswolds
Has thanked: 531 times
Been thanked: 1558 times
Contact:

Re: Pete's RX-8 Time Attack project

Post by Dr. FrankenRex »

PeteH wrote:
Wed Dec 20, 2017 3:17 pm
One rpm on a rotary means one rpm of the e-shaft, and 1/3 rotation of the rotor.
That explains it. I get it now, sorry team :D

Thanks for explaining!
DOC 787B

Club Information Coordinator - Media Team Lead


LM30 VelociBRAPTOR - Josh Scott magic sauce porting, shiny big brakes and a big ole spoiler

Doc's Auto Barn - https://www.youtube.com/channel/UC7BW_h ... Ym6eGqEIWg
350matt
Club Member
Club Member
Posts: 3158
Joined: Mon May 15, 2017 9:09 pm
RX-8: 231
Colour: Strato Blue
Location: Northants
Has thanked: 49 times
Been thanked: 587 times

Re: Pete's RX-8 Time Attack project

Post by 350matt »

hello Pete

you've probably seen this but it was worth listening to as it goes into a bit more detail about the 20B and its variants and whats involved in a transplant

PeteH
Former Member
Former Member
Posts: 5849
Joined: Mon Apr 11, 2016 1:26 pm
RX-8: 231
Colour: Custom
Location: West Sussex
Has thanked: 172 times
Been thanked: 787 times

Re: Pete's RX-8 Time Attack project

Post by PeteH »

Thanks Matt. Yes, I have seen it and it's not a bad summary. It did prompt me to look at if I've got an A or B spec motor (I haven't looked yet, and maybe I'd prefer not to know!).

A lot of the difficulties he mentions are applicable to road car transplants, but less so for race cars. His suggested budget looks about right, but is there really anyone who thinks they can wedge a 20B into an 8 for $2k??

I was subscribed to Rotors, but they make about two videos a day, and many of them are just his odd opinions on random stuff, so I unsubscribed. :roll:
350matt
Club Member
Club Member
Posts: 3158
Joined: Mon May 15, 2017 9:09 pm
RX-8: 231
Colour: Strato Blue
Location: Northants
Has thanked: 49 times
Been thanked: 587 times

Re: Pete's RX-8 Time Attack project

Post by 350matt »

its funny that there aren't more 20B MSP engines, I think
looks like hayward rotary did one a few years ago

and yes $2K for any engine transplant is dreaming
in my experience its more like 8-10K for even a simple put a carb fed rover v8 into a capri!

made a decision on the ECU yet?

also reading through your answer about the multi lambda stuff, i didn't mean to have 4 full time sensors ( although that is nice) rather have 1 sensor that you move around to get the data you need on full throttle ramps / mapping then when you come to run the car normally revert to just the 1 in the collector so the individual rotor trims then get left in the map as a passive value
that works pretty well in my expierence
Post Reply

Return to “Personal Car Projects and Mods”