http://www.jandssafeguard.com/safeguard.html

Next DIY Project Paul ? ;)

I don't think so...

I'll use it as a display first, and listen with headphones as well. If it is not capable of real time knock control with an advanced base map then the individual cylinder control will be relatively valueless, as would allowing more advance say for a quick 1/4 mile blast vs a hot trackday or top speed run. I will test gradually and work up and stop if I lose trust in the unit to do what it says. Will be interesting.

I've emailed the designer of the Safeguard for his comments and will post up his reply.

Some of the product claims...

"....every cylinder is monitored over 125 times per second and the ignition timing adjusted accordingly. Attempts to reprogram either factory or aftermarket ecus have proven fruitlless because without closed-loop control, it is impossible for the supercharger supplier to pre-ordain what will, or will not, happen in the engine. It is simply impossible to predict the point at which detonation will occur in each cylinder under every circumstance. Tests done on dynos to set individual cylinder timing are maddeningly complex and time consuming and do not relate accurately to real world operating conditions."

It seems that people are successfully using these units without reprogramming their timing and fitting turbos or superchargers to high compression NA motors and experiencing reliability. Very interesting... if it is after the (significant) event then you would expect a catalogue of failures?

[Edited by john banks - 9/4/2003 9:01:05 PM]

Yes, it is adjustable. Needs a lot of careful experimentation. It maybe will end up only being insurance where it takes out 1 or 2 degrees occasionally, like the standard Subarus end up doing, or saves your ass if the water runs dry of with a bad tank of fuel, or a punishing day like TOTB, or where you suddenly loose your coolant etc etc.

ISWYM re a single event, but is it really a single event that leads to the hot spots that cause pre-ignition (assuming that this is what really causes the damage rather than mild knock?) It seems that knock is poorly understood and even now after decades of spark ignition engines there is little agreement on what types of detonation are damaging and which are not? Single event on a display or a noise that a human reacts to and a microprocessor retarding the timing on an 8 ms timeslice could be two different things entirely?

Don't worry, John at J&S knows I am only going to use it as a monitor to start with, but apparently you can put one on a NA Miata and advance the factory timing from 10 to 16 degrees base timing, then add a turbo at 10 PSI (with intercooler and extra fuel) and it will not det with a J&S on it setup correctly.

The problem I see in mapping a high output engine is how to get realistic ignition timing when you do it open loop. If you map it on a dyno you end up frying the car and getting lame timing, engine dynos are expensive, and then perhaps not realistic in replicating road conditions. With a road NA engine, a lot of it is probably at MBT, and if it does ping the damage potential is far less. Subarus do seem crazy, on the same tank of fuel, running apparently the same boost and fuelling one minute and being happy then detting the next, sometimes you go on a mad blast and explore all the zones, think you are happy, then you go on cruise, EGTs are not excessive, then you wack the throttle open and you get a splash and a tinkle. On my own car I end up reflashing a few times after a brand new map through the next week or so as I rediscover knock in some areas, and that is pulling back a little safety margin every time but you end up like the old Subaru ECUs that always retarded and never advanced and can get lame timing... partly this is due to my greed for boost and not stopping with any component until I break it or try hard to (at least on my own car ;) )

Bruntingthorpe sounds like a good venue for high speed testing where you can fully load up at say 6000 RPM with a very slow ramp rate and get the timing right, but you might be able to run say 8 degrees more timing on cylinder #1 at 6000 RPM in 2nd gear on a cold day? More compensation maps in the better ECUs are nice, but on the other hand, they add another set of variables to manipulate. What should be so difficult about the knock phenomenon that we cannot trust a microprocessor to control it and instead map for a worst case scenario for temp/humidity/load/octane/component wear etc?

Any and all input welcome, let's debate this. I'm keen to learn, I'll share all my findings with this little box as and when I can.

[Edited by john banks - 9/4/2003 10:15:19 PM]

sorry, but.. say a 10mhz microprocessor in the j&s for example??

you map with det cans yes?

How quick would you have to react, ie listen think and retard to match the J&s??

money is on the J&S over your ears all day long..

10Mhz is 10 million cycles per second.

hmm

its back the 8 bits in a byte argument.

[Edited by David_Wallis - 9/5/2003 2:09:12 AM]

I've got one ordered. Was happy with the knocklink but on a rebuilt engine it is not so useful.

I will initially try it in knock display mode only and correlate with det cans. Will update with any results.

See http://forums.nasioc.com/forums/show...hreadid=402721

Let me know how you get on john..

also interested in cost.. (via mail :D)

David

Q?

How can it retard on individuel cylinders with only one knocksensor??

And can it be used with a LinkECU

Skassa

Time gated acquisition by intercepting the ECU signal (and retarding it on the next spark as required) to the coil. So it knows which cylinder to retard next time.

Yes,

I found out...

Looks like a must have part for safety...

Skassa

I seen that mentioned in Julian Edgars 21st Century Performance book some time ago. Wasnt too much info on it.
If it works the way it should, then it would be a great performance and safety aid.

John,

Wich version did you go for??


Skassa

It could be interesting if it's capable to work until the rev limiter and is not disturbed by engine noise different to knock.

Mark.

Went for 2 channel version as discussed with J&S. EMS I hope it does. It is supposed to.

I really look forward to hear the result of this.

Is it supposed to be so fast that You can give it way to much advance, and then it will regulate it down without damageing knock??

Skassa

Apparently so, you can put a turbo on a NA high compression engine safely enough as long as there is fuel etc and it will do the timing. However, if it is that far out it also has a midrange retard feature say if you throw a turbo on a Ford Focus.

There is also a nitrous retard feature too.

It will be a while until I test is as it may be a few weeks until I get it, install it, run my engine in, then test it in display mode whilst listening with det cans etc etc.

The hope would be that you could safely run a more advanced map as the claim is that the next spark event on a very slightly knocking cylinder will be retarded. It does not learn, because it is supposed to be quick enough to not need to.

I aim to do comparative runs to do a full shakedown of the product and will report results as usual (read long and boring with lots of graphs). It will be a while though.

There may be good gains from the individual cylinder control, plus gains that you can run say 5 degrees more on the first WOT on a cold day, but the unit would theoretically control timing with the same map from a cold day to a hot track day.

Safest way to do open loop timing is on a hot day with the load locked on the brakes or a dyno. The engine should then never knock. You can however get far more timing at 6000 RPM in 2nd gear than you can at 6000 RPM in 5th gear.

It could make a more responsive car that makes the most of ignition in all circumstances.

I suppose the Denso ECUs are supposed to do this, which a lot of the time they do quite well.

We'll see if this unit lives up to the promises on my engine, but I am excited by the possibilities.

I can´t hardly wait for the results!

Mark.

Why not try it yourself?

There is one concern that with a very advanced base map that constantly needs to be retarded that there would still be a mild knock each time you go through that zone at least on the first spark event. You would not hear this through det cans, but could the damage be cumulative? This is my lingering doubt... comments?

John,

If it doesn't work on your car I can keep my money in the pocket! ;)

Mark.

[Edited by EMS - 8/18/2003 9:05:21 PM]

John,

I know some people here in Denmark wich mappes alot of 1,8turbo´ed VW/Audi and BMW wich i think uses a simlier system(OE BOSCH) to retard advance(just 2 knocksensors)and it´s constandly retarding/upping the advance.....They seem to live happely..

Skassa

[Edited by Denmark - 8/18/2003 10:17:12 PM]

If it learns though (a la Bosch and Denso) that is good. If it does not learn you could get a single mild detonation episode in each cylinder before it retards. This would be too short lived to ever hear through det cans I think. I hear mild det is not supposed to be damaging, but this may be a unique situation if the device only lets it happen for one spark event... with open loop timing if it dets it would probably do it for many engine revolutions and so would be heard?

Maybe I am being paranoid, but would appreciate other views - if it does not learn and has a single mild det episode every time before retarding is this important?

If it is I could still map it so it is never intended to need the J&S to retard, then it would be a (useful) safety device rather than a performance device.

John,

Is there any update on this??.
Did you get the safegaurd yet????

i have just spoke to the guys i know who map a lot of the VW 1,8turbo and the bosch ecu is also running closedloop knockcontrol without learning....

When the cars are driven it´s constantly retading the advance(based on knock from the two knocksensors)...

Thanks,
Skassa

Not arrived yet. Money takes a while to clear the way you pay for it with a bank transfer.

Okay,

It´s close then:D

Skassa

Detailed results will be a while, since the engine is just complete and I expect will arrive next week. Then I am away on holiday, again :D ;)

Yippee - so mine is finally being built then!

Have a good holiday John!

:D

A motor free zone with lots of canals :D

However it doesn't neccesarily wait for det to do it. An ecu can be set to react to almost any level of knock sensor output.

John you would probably come unstuck if you rely on a device reacting to real det, it would only take an "event" to lose the engine and no amount of ignition retard would save it.

Map it to just under the det line and then it would be a useful tool, keep a sharp eye on the knocklink though.

Not against the system, if it works then great, the Scoob engine is not prone to consistent behaviour and, as mentioned, an "event" could trigger engine failure ... event being something that starts a plasma torch in the cylinder for example.

Think on ... Harvey did in a plug and damaged a valve, and nothing showed on his knocklink at all. So the system would not have reacted. Yet without Harvey feeling that something had happened and getting off the gas further damage would have resulted. Just an illustration of why you should still take great care even with this system.

Is the trip point programmable ?

cheers

bob