Brise Starter Current Draw / Transients

[revilla]

Been investigating starting problems on a friend's car. We have been probing this in depth for quite some time. The car is fitted with a Brise starter and standard Banner 53034 battery.

One thing which we noticed was that the battery voltage (measured directly across the battery terminals) dropped from a health 12.66V to less than 10V when cranking, which seemed on the low side. However swapping the battery for one which I'm pretty sure was good and fully charged (has been on a CTEK conditioner for the last few weeks) made absolutely no difference. The actual cranking performance is strong and normal in both cases.

This cannot be a wiring fault as the voltage was measured directly across the battery terminals; any voltage drops in the wiring would reduce the voltage measured at the starter but not at the battery.

The only conclusion I could draw was that the Brise was pulling a cranking current in excess of the capability of the battery.

Another observation was that any digital electronic device powered off the car misbehaves during cranking. Laptops connected to the ECU drop the connection and my OBDII scanner actually reboots every time the engine is cranked, suggesting large voltage transients on the supply lines beyond the voltage drop measured.

Unfortunately this strange behaviour seems to extend to the ECU which resets, recycling the IACV and requiring throttle to keep the engine idling for the first few seconds. If the IACV is unplugged whilst the engine is idling and the engine is then shut down, it will start fine and idle again as the ECU is prevented from closing it.

We have completely replaced the engine loom with a new build one and have also bypassed the wiring between the battery and the engine loom, putting in a direct fused wire of large cross-sectional area.

All of this leads me to the conclusion that the Brise starter is drawing a current which the battery cannot support and is generating unacceptable inductive transients. Unfortunately we were not able to try swapping in a standard Magneton starter as I didn't have a spare spacer ring.

Has anyone else experienced similar issues?

[elie boone]

Could it be that the Brise is running dry ? it would need more current to overcome the extra friction.

[Jonathan Kay]

Have you asked Brise for the spec figures?

I'm not sure why you're suggesting transients. Is that behaviour by the electronics any worse than you would expect when the feed drops below 10V? You could test this with an adjustable power supply (or a big variable resistor) separately from the voltage drop on cranking.

Jonathan

[revilla]

@Jonathan - Morning! I haven't had a chance to speak to Brise yet, however I have had a look at their website and it doesn't seem to give any electrical specifications - there is a downloadable data sheet but it is really only mechanical data. One thing they do claim for their starter however is a reduced current draw when compared to a standard starter. Pete's car does however seem to be loading the battery very heavily.

Yes it is possible that it is just the drop to 10V that causes the issues, although I would have though that most lightweight digital equipment probably regulates down to 5V internally and a 10V supply would still be sufficient to prevent the regulator from dropping out. My digital multimeter was showing around 10V whilst cranking but occasionally flashed up a silly value briefly during the transition, which suggested transients to me, however with a refresh rate of 2Hz it was hard to see much. Switching heavy inductive loads such as the solenoid and the motor windings will generate transients - although why they should be any worse in this car I have no idea. Being an ex-racer it does have an FIA switch which means long cable runs where ideally you would like to keep them short, although all of the cabling seemed to have been done appropriately.

As Pete was only up for the day from Southampton, a six hour round trip, we had our work cut out putting in a complete new engine loom and repairing / replacing some of the chassis loom wiring in the time available. I would have loved to have got an oscilloscope on it to see exactly what was happening at a much higher time resolution. I have lent him my spare Magneton starter as my spare engine is currently in pieces following the "finding-a-hole-in-the-block" incident; unfortunately the block is still with Ed White who was going to try to weld it for me and I left the bellhousing and the starter spacer ring on it so he check for clearance after welding, so we couldn't try fitting it yesterday. I will post the spacer down to him when I get it back and he can see what effect a change of starter has. If he is still having problems I guess we will just have to have a good 'scope session at some point.

 

[revilla]

@elie - I'm not quite sure what you mean by "running dry". Does the Brise have any special lubrication requirements?

Thanks.

[Myles]

Given the trouble I've had with my first Brise, I'd take this to refer to the possibility (certainty in my case) of a 'dry' high-resistance connection somewhere in there. In my case, it manifested itself by way of slow rotation - too slow on many occasions for the ecu to r cognise it and fire the engine up.

[revilla]

@Myles - In which case it couldn't really explain the issue I'm having. Anything with a higher than normal resistance would reduce the current draw on the battery and therefore give less of a voltage drop measured at the battery. In this case it was cranking strongly / normally but appeared to drawing excessive current.

[elie boone]

Revilla, what i mean is that maybe there is more than normal mechanical resistance to rotate the rotor.

[Mavic82]

If the voltage drops to 10v when cranking you will certainly have electrical issues. Most ECU's will not operate correctly below around 11v.

what causes the drop is another issue. Can you try both batteries in parallel to increase capacity?

 

[Titanium7]

A new Brise starter on my previous K Series and upon cranking immobilised the car (5AS). Only way I could stop it was to disable the immobiliser in the Emerald ECU.

Tried a new battery, even batteries in parallel with no effect. Had long conversations with Mike Molloy who developed the immobiliser but in the end put it down to the voltage drop caused by the Brise.

Eventually put an original starter back on and the problem disappeared!

[OliverSedlacek]

10V should be enough to keep ECUs and the like happy, and from memory some of the automotive standards call for systems to keep working down to 6V during cranking. Having said that, the infamous K click is clear evidence of a lack of adherence to that standard.

If the battery is holding up to at least 9V during cranking, this really should be enough and the fault is elsewhere. Assuming also that the engine is spinning briskly, then the starter motor gets a clean bill of health. My next port of call would be to measure the supply voltage at the ECU during cranking. If it's less than the battery voltage, this would need remedying.

[revilla]

@Mavic82 - I think my friend has tried paralleling another jump battery with little effect, although I guess the quality of the jump leads and the other battery is an unknown. We cannot at the moment try paralleling the known-good battery I have as he is now back in Southampton while I am near you!

 

@Titanium7 - This at least gives me hope that swapping starters may fix things. It seems that something at least similar to our problem has been seen before.

 

@OliverSedlacek - We have already made the measurements you suggest. Indeed this is why we replaced the engine loom. As well as the battery drop, there is around 0.75V dropped through the FIA switch and the extended battery wiring to it. This is with a brand new switch, which was changed as the voltage drop with the original switch whilst cranking was more like 1.5V. On the original loom, which was in a pretty tatty state when we removed it, there was a further massive drop in voltage through the engine loom to the ECU. To measure the voltage as seen by the ECU we were disconnecting the lambda sensor and measuring the voltage on the supply pin of the lambda connector, as within the loom this is commoned directly with the ECU supply pin and when drawing no current through it, it gave us a pretty good direct read of what the ECU would be seeing. Measured between the spade connector where the purple wire from the 30A "ECU"1 fuse feeds the engine loom and this point, there was a further 4V drop whilst cranking. We determined that much of this was in the Main Relay of the MFRU and the remainder due to resistance in poor spliced joints within the engine loom leading to voltage drops when carrying the starter solenoid current.

 

To get around all of these I made him a complete new engine loom from scratch, with all brand new connectors, wires and spliced joints, including a bypass of the Main Relay. Everything was brand new from the ground up working directly off my modifed / corrected wiring diagram. We installed this on Sunday, hopeful that the problem would be cured. The voltage drop through the loom, previously measured at 4V, was now only 30mV whilst cranking (so 133x better and no longer significant).

 

However, rather disappointingly the ECU still reset when cranking. Using a stethoscope on the IACV it can clearly be heard going through its reset cycle (going to fully closed then opening a fixed number of steps) when the starter button is pressed and the engine refuses to idle without throttle until the cycle is complete and the IACV is back in a suitable idle position. We tried unplugging the IACV whilst the engine was idling to "freeze" it in a resonable idle position, then stopping and restarting the engine. With the IACV disconnected it started perfectly every time. This is clearly not a satisfactory solution though for the long term as the ECU no longer has any control over idling conditions.

 

So the voltage drops we are left with when cranking are i) approximately 2.6V lost at the battery itself under load, bringing it down to 10V and ii) approximately 0.75V through the FIA switch wiring (which is carrying the full starter current for at least part of its length), bringing it down to around 9.25V. This is what is seen at the ECU with no further significant drops in the loom.

 

Personally I would be surprised if this alone were enough to reset the ECU as I would have thought it would be more robust than that to cope with low battery situations, however I note the differing opinions expressed here. Given the odd response that Titanium7 had from his 5AS, and given that other electronic equipment plugged into the vehicle also seemed to be badly affected whilst cranking (my CanScan 900 rebooted every time we cranked and I know it will run at that voltage because I've recently been using on my Touran which has developed, amongst other things, battery and alternator problems which cause the voltage to drop lower than that whilst cranking but still do not reset the scanner, Mark Stacey at Z&F found his laptop misbehaving etc.) I strongly suspect the Brise is throwing out nasty inductive spikes onto the power lines.

 

In a car without an FIA switch these will be dumped across a low internal impedance battery through a short cable which will act as a buffer, and the supply for the looms will be taken directly from the battery where the spikes will be shunted. With the FIA switch these spikes will propagate along the length the of the extended battery cables like a transmission line, and the take-off for the looms comes at the mid-point which is the switch itself. My strong suspicion is that this is allowing the high-frequency components of the spikes from the Brise to appear strongly in the main looms.

 

@Titanium7 - Do you by any chance also have an FIA switch?

 

Only a good examination with an oscilloscope will tell for sure, but in the meantime he will try a standard Magneton starter and see if the problem goes away.

 

Thanks as always for all your comments and input.

 

[revilla]

PS: Sorry guys for the complete lack of spacing which makes my previous post rather unreadable, I did try putting multiple line breaks in between paragrahs but it seems the new forum is still determined to beat me.

[Jonathan Kay]

Ways around the spacing problem: bulletted list, numbered list, change text style between paragraphs...

Jonathan

[revilla]

@Jonathan - Thanks. I've put bullets in, it makes it a little more readable but it still refuses to space it!

[Jonathan Kay]

How about using a variable resistor to turn down the feed until you know the voltage at which the ECU etc trip, regardless of that starter?

(I can't see how to study the current drawn without a test bench and a big DC ammeter, and I wouldn't know what was normal.)

Jonathan

[revilla]

I've just caught up with you on the bulleted list ... put an extra bullet in ...

 

A variable resistor would probably not be good as the voltage seen by the ECU would then fluctuate with the instantaneous current it was drawing - it would probably be all over the place, however a decent variable voltage bench power supply wired to the supply point to the engine loom from the body loom ... that would probably do the trick.

 

I don't have one, but I know a man who might!

 

[Mavic82]

From experience 10.8v and below stops ECU's working. 

How about running a direct earth to the starter?

[Titanium7]

Hi Revilla

No FIA switch on the car when the Brise used to activate the immobiliser. As I said the problem completely disappeared when I replaced the Brise back with a Magnetron Starter.

regatds

Martyn

[revilla]

@Mavic82 - There is already a direct earth to the starter. The main earth lead from the battery is atttached directly to the starter mounting bolt. the car has a shortened passenger footwell and the battery is low down very near to the starter so the earth cable run is very short. We could try a short direct positive lead but this would render the FIA switch useless / invalid.

 

@Titanium7 - Thanks, all good info. All being well I should get my old block complete with bellhousing and starter spacer ring back tonight, so I will post the spacer down to Pete and he should be able to try it with my spare Magenton starter. We will see what happens.

 

[revilla]

ISO 7637-2:2011 "Road vehicles -- Electrical disturbances from conduction and coupling -- Part 2: Electrical transient conduction along supply lines only" defines specifications under "Pulse 4" for the supply disturbance which devices designed to meet the standard should tolerate.

Although I have no idea whether MEMS3 and 5AS were designed to meet any similar standard (they obviously predate the 2011 strandard), the standard at least gives some idea of what kind of supply disturbance may reasonably be expected.

Pulse 4 specifies a voltage drop across the the supply of 6-7V for 15-40ms followed by a lesser voltage drop of 2.5-6V for 0.5-20s.

A voltage drop of approx. 3.35V at the ECU during the cranking period as we are seeing is toward the lower end of this specification and is therefore not to be considered unreasonable.

This further suggests to me that the starter is generating transient pulses outside of this envelope which are the source of the EMC problems. I just wish I could get my 'scope on it right now and see what was going on!

 

[revilla]

OK an update - Pete has replaced his Brise with my spare Magneton and all the electronics gremlins have disappeared, the ECU no longer resets and no longer closes the IACV on cranking and the car starts on the button and idles, without throttle, reliably every time.

The outstanding question is still whether it was down to pure current loading or electrical noise spikes. I have plans for some sensible ways to investigate further and would love to get to the bottom of it properly. If Pete can be persuaded that it is worthwhile to pursue, I will report back. If I were in his position I may however be tempted just to stick with the Magneton and call it job done!

In the meantime it is worth noting for the record that this is at least the second time that unexplained electronic malfunctions whilst cranking a Brise starter have resolved by reverting to a Magneton unit.

If anyone else suffers from strange ECU behaviour during starting on a Brise starter I would recommend that they regard the starter motor as a possible source of some kind of electromagnetic incompatibility, precise details unknown at this time.

[Mark Durrant]

I suggest you give Tim at Brise a call as I have always found him very helpful

[kart]

Had the same problem myself with the immobiliser activating upon cranking.  Spoke to Brise and they said they've seen this issue a couple of times now but they don't have an explanation.  Anyway they exchanged the problematc (Hitachi?) unit for a Nippon-Denso based one (which I believe used to be supplied by Caterham as their "race starter") and all is well with the world.

For the record my car is a 2006 S3 Roadsport running a 1.6K, I wonder if it's something to do with EU3 cars?

[revilla]

I did say I would report back if I managed to get to the bottom of this one ...

Since my last post I have tried fitting Pete's Brise starter motor to both my car and my spare engine. In both cases the engine developed identical starting problems to those Pete was experiencing. Given that Pete's starting problems had also disappeared when switching to a Magneton I think this conclusively identifies the Brise starter motor as the source of the problems.

I finally managed to get a Digital Storage Oscilloscope on an engine with both starter motors (i.e. with the Brise that was causing the problem and a Magneton). The results are interesting. Neither motor seemed to be generating any measurable nasty high frequency spikes, it purely seems to come down to the large inrush currents drawn by the motors before they spin up to speed. The Brise is just transiently drawing too much current for the battery to cope with and the voltage is dropping away. We were seeing that with a multimeter but it was very hard to get accurate measurements of what was going on as the meter only sampled once a second or so, so we got hints that there was a very deep dip in voltage but couldn't quatify anything reliably.

Here are the raw trace screen captures and the numbers separately. I took a number of captures from each motor and the exact shapes of the graph were remarkably consistent from one run to the next, so I've just included one from each. In each case I cranked the engine with the crankshaft position sensor disconnected so it wouldn't fire. I measured the supply voltage as seen by the ECU (using the technique of disconnecting the lambda sensor and measuring the supply voltage at the connector - with the sensor disconnected this is effectively a direct unloaded connection to the ECU supply pin). I set the scope to trigger on the voltage dropping through a 7V threshold but to record the voltage both before and after the trigger point. In the graphs below the small dips before the trigger line are due to the solenoid engaging (the solenoid didn't drop the voltage enought to trigger the scope), the much larger drops after the trigger line are the solenoid contacts closing and the motor engaging. In both cases the battery voltage (after a number of previous runs) was reading 12.3V.

I actually did the Brise first so if anything the battery was in a slightly better state of charge when testing the Brise than it was when testing the Magneton, but I only cranked for a second or two each time so I don't think it really made any difference.

Brise

The ECU supply voltage fell to 5.2V and was well below cranking voltage for about 100ms.

[brise]

Magneton

The ECU supply voltage only fell to 7.2V and was well below cranking voltage for about only 50ms.

Combined

I used Paint.NET to combine to above images to make it easier to compare the traces.



So overall the inrush transient from the Brise is deeper and wider; it is drawing a higher initial current off the battery for a longer period of time, causing the ECU voltage to fall further and for a longer period of time, which is obviously beyond what it's internal power supply circuitry can smooth out.

To be honest when I saw the trace from the Brise I wasn't suprised the ECU was getting upset as 5.2V is very low; the thing that surprised me was how badly it was still dropping away with the Magneton. At 7.2V the voltage was transiently way below the traditionally measured "Battery Voltage Whilst Cranking" which it settles to and which is measured on a multimeter. All of this was measured on a spare engine with a brand-new hand-made wiring loom with no significant voltage drops through the loom (previous testing showed that the ECU supply voltage was no lower than 200mV below the battery voltage even whilst cranking).

There really isn't as much difference between them as I expected and this must mean that the Magneton arrangement is still pretty marginal, which makes sense of the reports of people having almost identical starting issues (i.e. engine still cranks strongly but won't fire without throttle) when the battery is a bit down. I guess with a slightly weaker battery the Magenton trace would drop down to around where the Brise curve is.

It does appear however that the Brise solenoid draws less current as the smaller solenoid spike is much deeper with the Magneton; this is probably related to it's relative immunity to the K-Click.