A Slightly Geekish Look At FSI Coolant Warm-up

[Andrew]

Only read if interested!

Currently I am involved in constructing a label file for the FSI ECU, a small portion of which is the coolant groups 130-137 and along with my contributions to Evros's current coolant thread I have have become intrigued with FSI coolant warm-up, how it behaves and what exactly these groups are telling us.

Below is a detailed look at the VCDS log of groups 130 -132 during a test run from cold in my FSI, that I undertook to gain more data on FSI coolant warm-up.

Background

Ambient temperature (from dash) = 8.5℃. Cabin climate off. The route was a fairly brisk run along country roads with intermittent slows spells through three villages and the return journey. The car was stationary with engine off for a few minutes in the middle while I briefly did some shopping!

The log is of this VCDS screen and is 1760 lines long.

The start of the log.

The same view but tied up with duplicated columns hidden and the elapsed time changed from seconds formatted to mins:secs. ( Sorry noticed later I missed a duplicated column.)

Note also hidden is column F, the 'Fan overrun' field that displays 'OFF' throughout. I know @PlasticMac has an interesting point about this and I will let Mac comment.
'

With sampling rate of about 1 second not a lot happens quickly, in fact it takes 27 seconds for the coolant temperature to rise 1℃ but warm-up speeds up a little.

This slow increase in coolant temperature continues, painfully slowly when looking for a change over several hundred lines of log. After eight and half minutes the magical 90℃ is reached. Note the dash temperature gauge is in agreement to within a couple of degrees at this point but as said many times stays there from this point on.

The climb in coolant temperature continues towards the setpoint of 110℃ and hopefully more interesting points to look at and when it arrives it is not what I expected at all! Bizarre the ECU starts to drop the engine and radiator outlet setpoints???? What is the point of that? It knows the coolant temperature is nowhere near the normal 110℃ why is it doing things? if the ECU is worried about the rate of rise of the engine coolant temperature then why it does it not just open the thermostat. Somebody please explain.

Andy

 

[Andrew]

Continuation.

This unexpected lowering of setpoints leads to a brief event. Appears this 4 seconds was a precursor, cue action, the ECU opens the thermostat,

I assume that is what column F is telling me, actually confirmed by column P on the right. The column on the right is an 8 bit field the meaning of each bit is defined by

I have little idea what '1000' actually means, but up until now that's what shows over the previous hundreds of lines of log. I cannot get my head round ' Control deviation (0 > setpoint/1 <setpoint)'. Can someone please explain.
I am happy with '10', and means the ECU is applying voltage through the wires embedded in the thermostat wax and melting it, resulting in column F not showing zero, the thermostat is open. Or is it - read on.

Th action is over in a heartbeat.

In under 2 seconds the ECU kills the wax heating and the thermostat shuts. Seems odd to me, I cannot envisage that would be long enough to melt the wax. However the ECU then promptly resets the setpoints.

What was the point of that 13 second flurry of ECU intervention. It did not achieve anything and back to square one? Almost seems like the ECU is flexing its coolant temperature control muscles! The ECU god moves in mysterious ways.

The ECU is happy and lets the coolant temperature rise towards its setpoint during the next two and half minutes without intervention, not worth a screen shot

Nearly there and the ECU gets nervous, drops the setpoints again and opens the thermostat and the temperature begins to plummet.

With roughly 10 seconds of thermostat open the radiator outlet temperature has not increased, stuck on ambient. A testimony to the efficiency of the radiator. The hot coolant will be hitting cold metal for the first time so perhaps not surprising. It does not last the flow through the radiator return hose to thermostat begins to show. I am surprised how fast the coolant temperature drops about 10℃.

The cycle starts again, the coolant temperature begins its climb but at one point it drops of is own accord by a few degrees, probably driving fast here, but not worth a screen shot but the rise continues until the setpoint is reached and exceeded!!!

Note the zero in the right hand column I assume in response to the 111℃.

The cycle continues, but the coolant is generally running hot with longer periods of the thermostat open.

And home.

One last comment. I really wish I had a passenger sat in the back seat with the laptop adding marker points to the log and making notes, e.g. marker point 4, queuing at roundabout, marker point 7, three minutes at 60 mph, marker point 10, 25 mph through village etc.. It would be so useful. I drive with the laptop on the front passenger seat and the lid down otherwise it's so distracting.

Another last comment. Does anybody know about the 'Heating flow potentiometer'? Its column always shows 100% and never changes?

Andy

 

[Evripidis]

Hi Andy, in my tests the setpoints where always constant. This is peculiar!

Great effort, thank you for your input. I hope that this will help me resolve my issues and will benefit others as well.

@Andrew for the setpoint: Have a look at an example from automatic control, called PID controller and its flavour called setpoint weighting or setpoint ramping.

Edit 2: I have read on page 17 of this http://www.volkspage.net/technik/ssp/ssp/SSP_222.pdf that the setpoints are indeed mapped, i.e, they are variable based on other conditions such as load. This is the reason why I have never seen them change, my test were conducted with the vehicle being stationary. You did a live test. I am limited in that respect that my laptop will not function without being plugged in the wall socket.

 

[PlasticMac]

I'm not a morning person, so sorry for the slow response.
First, the term "Fan Overrun" OFF, refers to the need to run the fan with the ignition OFF. If the Engine outlet temperature is still high when the ignition is turned OFF, * the fan will run until the temperature falls to a safe level. I don't know what the safe temperature level is, but my guess it around 85C. The fan will also run, if either of the coolant temperature sensors fail. I don't know if this is also a Fan Overrun, but I think it may well be, because with the ignition OFF, you can't scan, so if Fan Overrun was only when ignition is OFF, there seems no point in monitoring it.
*(Worth bearing in mind when investigating battery drain when parked).

I'll study the logs a bit more, and come back later, but for now, as @Evripidis says, SSP 222 explain things a bit. My first thought is that the thermostat, which I thought opens at 110C (via the wax capsule), or 85C, onlv via the ECU override. But, I'm thinking now that if 110C is not reached, and in current Load, (MAF), Road Speed, and Air Inlet Temperature, it is unlikely to be, the ECU will lower the set point, so that the thermostat opens at the exact point required to achieve best combination of MPG, and Emissions.
Mac.

 

[PlasticMac]

I'm not a morning person, so sorry for the slow response.
First, the term "Fan Overrun" OFF, refers to the need to run the fan with the ignition OFF. If the Engine outlet temperature is still high when the ignition is turned OFF, * the fan will run until the temperature falls to a safe level. I don't know what the safe temperature level is, but my guess it around 85C. The fan will also run, if either of the coolant temperature sensors fail. I don't know if this is also a Fan Overrun, but I think it may well be, because with the ignition OFF, you can't scan, so if Fan Overrun was only when ignition is OFF, there seems no point in monitoring it.
*(Worth bearing in mind when investigating battery drain when parked).

I'll study the logs a bit more, and come back later, but for now, as @Evripidis says, SSP 222 explain things a bit. My first thought is that the thermostat, which I thought opens at 110C (via the wax capsule), or 85C, onlv via the ECU override. But, I'm thinking now that if 110C is not reached, and in current Load, (MAF), Road Speed, and Air Inlet Temperature, it is unlikely to be, the ECU will lower the set point, so that the thermostat opens at the exact point required to achieve best combination of MPG, and Emissions.
Mac.

I'd think the engine temperature Set Point is likely to be the maximum, stable temperature that can be achieved, while still allowing the thermostat to cycle open and close.
So, I'd expect to see the Set Point temperature and the actual Engine Coolant temperature, to Converge.
Mac.

 

[Evripidis]

What worries me the most is that the cabin heat function appears to be affecting DIRECTLY the mapping, i.e., you get worse MPG if you turn on the heating on full blast. This agrees with other cars that do not have such a mapped thermostat function, i.e., you drop the coolant temp by running a fan in front of the heater matrix, thus the engine will run colder, thus worse mileage. But in the VAG case you can alter the mapped coolant temp directly by the heating function by the means of a setpoint (page 16). The refresh rate of the control state machine according to the comic book is 1 Hz.

I am experiencing a drop of 4 MPG compared to the summer months and this might as well explain it! Even my partner who is fed up with it noticed the drop in MPG.

 

[PlasticMac]

What worries me the most is that the cabin heat function appears to be affecting DIRECTLY the mapping, i.e., you get worse MPG if you turn on the heating on full blast. This agrees with other cars that do not have such a mapped thermostat function, i.e., you drop the coolant temp by running a fan in front of the heater matrix, thus the engine will run colder, thus worse mileage. But in the VAG case you can alter the mapped coolant temp directly by the heating function by the means of a setpoint (page 16). The refresh rate of the control state machine according to the comic book is 1 Hz.

I am experiencing a drop of 4 MPG compared to the summer months and this might as well explain it! Even my partner who is fed up with it noticed the drop in MPG.

So, maybe Audi forgot the effect of the heater matrix, when developing the cooling maps.
If you switch the climate OFF, the warm up time, (according to the dash gauge), is drastically reduced. Maybe they used an existing climate software, rather than developing a custom FSI one.
Mac.

 

[Evripidis]

So, maybe Audi forgot the effect of the heater matrix, when developing the cooling maps.
If you switch the climate OFF, the warm up time, (according to the dash gauge), is drastically reduced. Maybe they used an existing climate software, rather than developing a custom FSI one.
Mac.

I have just verified this. Within 5 minutes of driving the gauge was dead on 90. I set the climate to low and then turned it off.

 

[PlasticMac]

I have just verified this. Within 5 minutes of driving the gauge was dead on 90. I set the climate to low and then turned it off.

Seems like a very basic error. There's a valve in the cylinder head circuit, that's closed when the climate is OFF. Audi could have kept it closed, regardless of climate status, until the cylinder head circuit was up to temperature. That valve is not there on non FSIs.
Wonder if a remap of the climate software is possible ...
Only joking. Just turn it off, before starting the car, then on, as soon as the gauge reaches 90C.
Mac.

 

[PlasticMac]

Continuation.

This unexpected lowering of setpoints leads to a brief event. Appears this 4 seconds was a precursor, cue action, the ECU opens the thermostat,

View attachment 105057

I assume that is what column F is telling me, actually confirmed by column P on the right. The column on the right is an 8 bit field the meaning of each bit is defined by

View attachment 105059

I have little idea what '1000' actually means, but up until now that's what shows over the previous hundreds of lines of log. I cannot get my head round ' Control deviation (0 > setpoint/1 <setpoint)'. Can someone please explain.
I am happy with '10', and means the ECU is applying voltage through the wires embedded in the thermostat wax and melting it, resulting in column F not showing zero, the thermostat is open. Or is it - read on.

Th action is over in a heartbeat.

View attachment 105061

In under 2 seconds the ECU kills the wax heating and the thermostat shuts. Seems odd to me, I cannot envisage that would be long enough to melt the wax. However the ECU then promptly resets the setpoints.

What was the point of that 13 second flurry of ECU intervention. It did not achieve anything and back to square one? Almost seems like the ECU is flexing its coolant temperature control muscles! The ECU god moves in mysterious ways.

The ECU is happy and lets the coolant temperature rise towards its setpoint during the next two and half minutes without intervention, not worth a screen shot

Nearly there and the ECU gets nervous, drops the setpoints again and opens the thermostat and the temperature begins to plummet.

View attachment 105062

With roughly 10 seconds of thermostat open the radiator outlet temperature has not increased, stuck on ambient. A testimony to the efficiency of the radiator. The hot coolant will be hitting cold metal for the first time so perhaps not surprising. It t does not last the flow through the radiator return hose to thermostat begins to show. I am surprised how fast the coolant temperature drops about 10℃.

The cycle starts again, the coolant temperature begins its climb but at one point it drops of is own accord by a few degrees, probably driving fast here, but not worth a screen shot but the rise continues until the setpoint is reached and exceeded!!!

View attachment 105072

Note the zero in the right hand column I assume in response to the 111℃.

The cycle continues, but the coolant is generally running hot with longer periods of the thermostat open.

View attachment 105073

And home.

One last comment. I really wish I had a passenger sat in the back seat with the laptop adding marker points to the log and making notes, e.g. marker point 4, queuing at roundabout, marker point 7, three minutes at 60 mph, marker point 10, 25 mph through village etc.. It would be so useful. I drive with the laptop on the front passenger seat and the lid down otherwise it's so distracting.

Another last comment. Does anybody know about the 'Heating flow potentiometer'? Its column always shows 100% and never changes?

Andy

Andy,
I'll have a think before commenting on your findings. But I do think the ECU is trying to use it's "levers" to achieve the highest engine temp, consistent with the normal thermostat cycle, possibly so it is keeping an eye on the active elements of the cooling circuit. Maybe, because of the elevated temperature, and the consequences of sensor failure if it did not check very regularly.
I wondered about the coolant flow pot, never seen it mentioned in the mapped cooling SSP.

I can't help with the laptop, as my hand is not steady enough to make notes on the move, but I'd be happy to be the pilot, to your observer.
Mac.

 

[Evripidis]

Hi Andy,

Excellent write-up as always. My comments:

1. The ecu appears to be doing what the comic booc (SSP ok) says.
2. The heating flow potentiometer (load % on my screenshots) must be the heater control "knob", i.e., temperature setting. The SSP mentions the term load on page 16 in the small graph.
3. During my static tests the control deviation bit would flip every time the setpoint was exceeded or subceed by the engine coolant temp.

Evros

 

[PlasticMac]

Andy,
I'll have a think before commenting on your findings. But I do think the ECU is trying to use it's "levers" to achieve the highest engine temp, consistent with the normal thermostat cycle, possibly so it is keeping an eye on the active elements of the cooling circuit. Maybe, because of the elevated temperature, and the consequences of sensor failure if it did not check very regularly.
I wondered about the coolant flow pot, never seen it mentioned in the mapped cooling SSP.

I can't help with the laptop, as my hand is not steady enough to make notes on the move, but I'd be happy to be the pilot, to your observer.
Mac.

Andy,
Maybe the coolant potentiometer was omitted late in the engine's development, and monitoring of flow was done by the thermostat open - close - open cycle you've seen. Where small temperature changes, associated thermostat movements, confirm flow, without a moving part and associated sensor.
Mac.

 

[PlasticMac]

Andy,
I'll have a think before commenting on your findings. But I do think the ECU is trying to use it's "levers" to achieve the highest engine temp, consistent with the normal thermostat cycle, possibly so it is keeping an eye on the active elements of the cooling circuit. Maybe, because of the elevated temperature, and the consequences of sensor failure if it did not check very regularly.
I wondered about the coolant flow pot, never seen it mentioned in the mapped cooling SSP.

I can't help with the laptop, as my hand is not steady enough to make notes on the move, but I'd be happy to be the pilot, to your observer.
Mac.

I mean I'll drive mine, in case you wondered!
6.

 

[Andrew]

Seems like a very basic error. There's a valve in the cylinder head circuit, that's closed when the climate is OFF. Audi could have kept it closed, regardless of climate status, until the cylinder head circuit was up to temperature. That valve is not there on non FSIs.
Wonder if a remap of the climate software is possible ...
Only joking. Just turn it off, before starting the car, then on, as soon as the gauge reaches 90C.
Mac.

Tell me more about this valve in the cylinder head, I've never heard of it, parts diagram?

I doubt the dual coolant circuit design. As far as I know the coolant goes into the block after the pump and comes out to discharge into thermostat housing at one point from the block.

Andy

 

[PlasticMac]

The valve, like all things FSI, it's a bit mystical.
The Mapped Cooling SSP shows it, and the electrical schematic shows the BAD only switch in the climate panel.
@Evripidis has found the valve, in pictorial form, but so far, it's physical manifestation has eluded us.
That it's working is obvious, with the climate switched OFF, the warm up time is a few minutes, compared to that when it's ON.

The * indicates it's a BAD only fitment.

Mac.

 

[PlasticMac]

Tell me more about this valve in the cylinder head, I've never heard of it, parts diagram?

I doubt the dual coolant circuit design. As far as I know the coolant goes into the block after the pump and comes out to discharge into thermostat housing at one point from the block.

Andy

I've based my belief on the descriptions in SSP222. I know that's for Lupo FSI and Golf FSI, but wouldn't think the A2's FSI implementation of mapped cooling would be any different, (but could be wrong).
It's intention is to get the combustion area warmed up as fast as possible, to maximise thermal efficiency, and minimise fuel consumption.
Being able to Isolate the heater matrix, (with the mythical valve), during warm up, fits this idea.
Mac.
Edit: I believe it's designation is N279.
Mac.

 

[PlasticMac]

@Andrew Just found this listing, allegedly the thermostat housing for A2, Lupo and Golf FSIs.
If it is for all three cars, then the A2's coolant system is the same as the Lupo and Golf, which in turn, is described in SSP222, with two coolant circuits, and heater matrix shut off valve. I've got an idea you have a spare? If so, does it match, P/N is 032121111CE?
Mac.

 

[Evripidis]

Well, going a bit off topic here but; had I purchased this it would set me back an additional 46 quid + (95+46)*0.2 = 46+28 quid extra to get over here! Roughly double the price! I blame Brexit!

 

[Andrew]

@Andrew Just found this listing, allegedly the thermostat housing for A2, Lupo and Golf FSIs.
If it is for all three cars, then the A2's coolant system is the same as the Lupo and Golf, which in turn, is described in SSP222, with two coolant circuits, and heater matrix shut off valve. I've got an idea you have a spare? If so, does it match, P/N is 032121111CE?
Mac.View attachment 105144

Good Evening Mac,

Yes, the picture does indeed look like the correct thermostat housing.

I was going to reply to reply to your previous post this evening (now), ( Been busy with the label file earlier). I was going to suggest any shut off valve for the cabin heating circuit, rather than be in the engine be in the climate unit. In a way for an A2 equipped with climate it effectively has a valve by just keeping its flaps shut but it would be needed for manual heating controls so might as well equip all A2s with a means of blocking flow through the cabin heater matrix to improve engine coolant warmup time, (and let freezing cabin occupants suffer as they cannot be trusted to turn on/up the heating with manual controls).

In practice I would suspect any valve would initially limit the amount of circulation through the heater matrix but allow some as a a fob to keep the cabin occupants happy and then when the engine coolant is sufficiently hot open fully. You can certainly experience this when the climate jumps from two bars to six and hear it in terms of fan noise. Also evidence in the graphs I have made of coolant warmup when the the rate of increase suddenly dips around 60℃ when the climate is on.

Sorry, I feel this post is very disjointed.

Andy

 

[PlasticMac]

Good Evening Mac,

Yes, the picture does indeed look like the correct thermostat housing.

I was going to reply to reply to your previous post this evening (now), ( Been busy with the label file earlier). I was going to suggest any shut off valve for the cabin heating circuit, rather than be in the engine be in the climate unit. In a way for an A2 equipped with climate it effectively has a valve by just keeping its flaps shut but it would be needed for manual heating controls so might as well equip all A2s with a means of blocking flow through the cabin heater matrix to improve engine coolant warmup time, (and let freezing cabin occupants suffer as they cannot be trusted to turn on/up the heating with manual controls).

In practice I would suspect any valve would initially limit the amount of circulation through the heater matrix but allow some as a a fob to keep the cabin occupants happy and then when the engine coolant is sufficiently hot open fully. You can certainly experience this when the climate jumps from two bars to six and hear it in terms of fan noise. Also evidence in the graphs I have made of coolant warmup when the the rate of increase suddenly dips around 60℃ when the climate is on.

Andy

When the OFF button on the climate panel is pressed, the cabin remains very cold, everything is off, blower display etc. The N279 is called "coolant shut off valve"
I suspect you're right about position, I'd say on the cabin side of the bulkhead.
The really odd thing is that the switch that operates it is shown on the schematic, (but is not identified), but the actual valve is nowhere to be found.
How us the label going?
Mac.