MG4 Mysteries

[Martinonline]

Why is 29.8 arbitrary? It may be result of experimentation or analysis. If you use Watt-hours then is 29800Wh/100km (or simplifying 298Wh/km) arbitrary?

 

[decrep]

Well it just seems that way, from tests done, where energy out put is increasing but indication doesn't.
It doesn't appear to be a real figure of energy consumed after that point. Only the max amount Saic wants to publicize

Have a look at the video of the Norwegian guy in Thailand doing a stress test of an MG4 up a mountain.

EDIT,
After thinking about this overnight, I'm thinking, the energy being consumed number, may be actual. It would make sense to put a limit on battery power drain.

[ Edit moderator: battery drain → battery power drain ]

 

[Blindin]

So does that mean (possibly) your expectations are a bit to high. ?

Well, I personally don't think my expectations are a bit high of buying a new car where every feature works as it should, but being used to a lower quality car where you may already be used to it may make it more acceptable.

 

[Coulomb]

Why (for us metric types) does the energy consumption number cap at 29.8 kWh/100km.

I bet it's a limitation of how they store a relevant number, likely in 16, 24, or 32 bits. If you don't cap it, it will "wrap around" to either 0 (unsigned integer) or to the largest representable negative number (signed integer).

I've done a search for Chinese energy units,

China is a 100% SI (metric) country. The Brits could learn from that (at least they're partly metric). That's why all MG bolts have thankfully metric threads.

China does tend to keep an eye on the American market though, even though SAIC aren't selling in the USA as yet. So many Chinese products use imperial units and American spelling for anything customer facing, though I've not noticed any of that with MG, apart from miles and PSI for British models.

I read recently that the first Chinese company (a Geely brand, Zeekr) are making a big push into the US auto market, despite the tariffs. I guess they're banking on Trump not being president in 2025.

 

[decrep]

I bet it's a limitation of how they store a relevant number, likely in 16, 24, or 32 bits. If you don't cap it, it will "wrap around" to either 0 (unsigned integer) or to the largest representable negative number (signed integer).

That makes sense, lets see where has my binary gone? 16bits should easily count to 30.
In fact if my doodling is correct it can be done with 5 bits, I think 01011 is 30. But it's a very long time ago that I played with this stuff.

 

[Tuareg]

Binary 01011 is 11 Decimal, ....... 0+8+0+2+1

Binary 11110 is Decimal 30, ........16+8+4+2+0

16 bit can store an unsigned value of 65535 and a signed value of +/- 32767

 

[decrep]

Binary 01011 is 11 Decimal, ....... 0+8+0+2+1

Thanks Tuareg, I see where I went wrong, left the first 1 then got mixed up.
How about 11110, 16+8+4+2+0

 

[Coulomb]

16bits should easily count to 30.

Certainly it can represent the final answer easily, but there has to be internal calculation (at least one divide, for example), and that's where I expect the limitation. Probably the energy used has to be stored down to the mWh or so, otherwise the accumulation of energy used will be very inaccurate. They probably need a circular buffer of energy use, so that they always have a running average of usage.

 

[decrep]

I can see why they need a power limit, but why an energy limit?

This is so complicated, it's doing this old head in.

 

[AlC]

Certainly it can represent the final answer easily, but there has to be internal calculation (at least one divide, for example), and that's where I expect the limitation. Probably the energy used has to be stored down to the mWh or so, otherwise the accumulation of energy used will be very inaccurate. They probably need a circular buffer of energy use, so that they always have a running average of usage.

Yes I agree, I'm pretty sure it's the result of a calculation from some internal representation that brings it up to 29.8. I just can't believe that any engineer, given the design requirement to "cap the value of the worst value displayed on the efficiency meter" would decide "OK, I'm going to stop counting at 29.8 and not 30"

I am left wondering what that internal representation could be though.

One thing I'm sure of, the car is actually calculating the efficiency even when it's over 29.8 kWh/100 (or 2 miles/kWh for UK cars). It's not just clipping the data. My reason for thinking that is, you can get deep into 29.8+ territory (e.g. when you're just starting out, driving uphill, with the heater blasting) and it will take a while for it to begin to come back down once you're driving more efficiently. If it was just refusing to calculate once your economy hit 29.8, you'd expect that number to begin falling immediately once you started downhill and were on regen, for example. But it doesn't. It'll stay up there at 29.8 for a while, presumably until the actual calculated value drops below 29.8 and then you'll see the display updating again.

And, assuming that it is calculating the efficiency but refusing to display it when it gets too high, I have to wonder - why? Was someone worried that higher numbers would scare off consumers? Were they worried about people posting shocking screenshots on social media of the display reading out "99+"?

I'm pretty sure when Bjorn tested the MG4 out in Thailand he also made the comment that other Chinese cars do it too, although I've not been able to independently verify. Could there be some legislative requirement or industry body standard in China?

So many mysteries. Lucky this is the MG4 mystery thread I guess!

 

[JeanLeFlambeur]

1. When the front wipers are going, if you put it into reverse, it'll run the rear wiper a couple of times for you to clear the back window. Nice touch.

I mean it is a nice function but my 2007 Focus Ztech also did this.

 

[AlC]

I'm adding another mystery to the list:

The car locks all doors at 15km/h. When you stop and open the driver's side door, it's supposed to unlock all the doors, and usually does but very occasionally it doesn't. Often when there's some drama with the kids and we need to pull over and sort things out. I can't work out why, but I have a suspicion if my wife opens the passenger side door just very slightly before me, it doesn't trigger the rear doors to unlock and I have to manually press the button.

I'd like to test this theory but my wife's not keen on sitting in the car with me as we go around the block multiple times to test various door opening sequences, funnily enough.

 

[T0M]

I have one as well.

Few days ago, I've opened drivers and passenger windows. After that, I was not able to close it any more.
I had to lock the car for ~20min and then it worked again.

 

[decrep]

Just came across this yesterday. As I adjusted LKA, the home page suddenly changed to something I'd never seen before and no pressing of the Home button would bring it back. Can't say I actually panicked, but I was a tad worried, one of the mysterious faults had appeared.

Then I noticed that little white stripe at the bottom of the page, and sure enough, swiping the panel sideways restored the original home page.

I wonder how many more surprises there are in store for us.

 

[OriginalBigAl]

I have one as well.

Few days ago, I've opened drivers and passenger windows. After that, I was not able to close it any more.
I had to lock the car for ~20min and then it worked again.

Watched a YT video on a 2year old MGHS which did something similar , They put this down to the BCM getting to warm ( was it a hot day ) and having a hissy fit. Once it had cooled down it was fine.????

 

[siteguru]

Just came across this yesterday. As I adjusted LKA, the home page suddenly changed to something I'd never seen before and no pressing of the Home button would bring it back. Can't say I actually panicked, but I was a tad worried, one of the mysterious faults had appeared.

View attachment 28873

Then I noticed that little white stripe at the bottom of the page, and sure enough, swiping the panel sideways restored the original home page.

I wonder how many more surprises there are in store for us.

That's standard functionality ... most of us have known about that for a long time.

 

[decrep]

I suspected that, but it was a complete mystery to me

 

[T0M]

Watched a YT video on a 2year old MGHS which did something similar , They put this down to the BCM getting to warm ( was it a hot day ) and having a hissy fit. Once it had cooled down it was fine.????

Yes, it was hot day. I think it just needed reset which happened when car was locked for 20min.

 

[NickHeynes]

Well, I personally don't think my expectations are a bit high of buying a new car where every feature works as it should, but being used to a lower quality car where you may already be used to it may make it more acceptable.

Personally I would say that, regardless of the price , if a feature is there it should work . Otherwise it's not a feature and certainly of no benefit..

 

[AlC]

OK well I worked out the mystery of the self locking doors not automatically unlocking sometimes.

If the front passenger side door is opened first, and then the drivers side door is open within ~1 second or so of the passenger door, the rear doors are not automatically unlocked.

If you open the driver's side door first or allow ~2-3 seconds after the passenger side door is opened, all doors unlock automatically. Weird.

Another mystery I haven't been able to solve is: How long after stopping the car do I need to wait before the "after start" trip meter will reset? Sometimes I'll jump back into the car a after 10+ minutes, and having shut down the car by locking the doors, and the "after start" trip meter still thinks I'm on the same trip.