Features of good airflow meter design
Discussion
At first glance one would expect that the air flow in the inlet tract to be in one direction only, however that is not the case. At lower engine speeds before the engine comes "on cam", there is considerable reverse flow out of the inlet during the overlap period between the inlet valve opening and the exhaust valve closing.
This is a simple feature of camshaft design, since the inlet valve and exhaust valve are open at the same time at the end of the exhaust stroke for a very short period know as "overlap". If there is any doubt about this and confirmation of the fact is needed, just look inside the inlet trumpets in the plenum. There are carbon deposits inside the trumpets, which look just like exhaust deposits. These are indeed exhaust deposits, from the exhaust gas that is forced back up the inlet manifold.
For a vehicle manufacturer, this reverse flow is useful in controlling oxides of Nitrogen (NOx) present in the exhaust gas without using an Exhaust Gas Recirculation (EGR) system. The exhaust gas that is returned with the inlet charge has the effect of slowing the advance of the flame front in the cylinder, hence reducing the combustion temperature that produces NOx in the exhaust (note NOx production increases with lean mixtures and the associated increasing combustion temperatures).
Once the engine comes on cam at higher speeds, the effects of shock ramming in the inlet and inertial scavenging in the exhaust take over. When this happens, the reverse flow of exhaust gas into the inlet system is substantially reduced.
This is all-important because it has a major impact on measurement of air flow in the inlet tract. Due to the pulsing and potential reverse flow in the inlet, all normal AFMs are made using what is known as "flow rectification". This ensures that flow is measured in one direction only, by mounting the sensing element in a specially-designed housing.
Without such measures, the air mass can be measured three times (in, out, and in again) leading to very rich running at low speeds (especially on full throttle) which shows up as massive low-speed over-enrichment.
Another important AFM feature is the mesh or "chip cutter" grille, which is required to collate the airflow as it hits the bridge and even out the airflow across the section of the meter. In the TVR installation, you can almost (but not quite) get away without it due to the fact the air flows up a long straight tube before it hits the meter.
The basic principal of just dropping a bare sensor into the airflow is not a good one, and is never seen on ANY professionally manufactured air flow meter (AFM). This is part of the complex problem of low engine speed and part-load mixture control.
All this information is freely available in any one of the many reference documents on AFM construction, and is indeed borne out by any experimentation. The many years of detailed research results into this topic would be normally be considered essential reading for anyone embarking upon a new AFM design project.
It is certainly beneficial to use a larger section for the AFM than the standard Lucas 5AM for larger engines (4.5 Litre plus), and improvements of around 8-10BHP are seen for 4.5 and 5.0 Litre cars when using the 77mm Bosch Hot Film or Lucas-Sagem 20AM AFM.
Apart from the AFM construction itself however, there is another problem in the software. If you consider the structure of the map (load vs. Speed), it is apparent that the load position in the fuel map is not determined by AFM voltage alone. At low speeds the AFM will simply not produce the full output voltage even on full load. To enable full resolution of the map load scale, the ECU is programmed with the transfer characteristic of the AFM, and scales the output accordingly.
When changing the AFM, it is always necessary to reprogram the AFM transfer characteristic to suit. If this is not done, the ECU load calculation is completely wrong which means that the ECU moves through the map incorrectly. This either means that it never gets to the top line of the map, or it get to the top too quickly. In either case you lose the resolution necessary for accurate part-throttle mixture control, usually resulting in massive low-speed overfuelling.
Although I have seen other AFM conversions there are definitely no modifications to the software I have seen, and I have now seen more than four.
This is a simple feature of camshaft design, since the inlet valve and exhaust valve are open at the same time at the end of the exhaust stroke for a very short period know as "overlap". If there is any doubt about this and confirmation of the fact is needed, just look inside the inlet trumpets in the plenum. There are carbon deposits inside the trumpets, which look just like exhaust deposits. These are indeed exhaust deposits, from the exhaust gas that is forced back up the inlet manifold.
For a vehicle manufacturer, this reverse flow is useful in controlling oxides of Nitrogen (NOx) present in the exhaust gas without using an Exhaust Gas Recirculation (EGR) system. The exhaust gas that is returned with the inlet charge has the effect of slowing the advance of the flame front in the cylinder, hence reducing the combustion temperature that produces NOx in the exhaust (note NOx production increases with lean mixtures and the associated increasing combustion temperatures).
Once the engine comes on cam at higher speeds, the effects of shock ramming in the inlet and inertial scavenging in the exhaust take over. When this happens, the reverse flow of exhaust gas into the inlet system is substantially reduced.
This is all-important because it has a major impact on measurement of air flow in the inlet tract. Due to the pulsing and potential reverse flow in the inlet, all normal AFMs are made using what is known as "flow rectification". This ensures that flow is measured in one direction only, by mounting the sensing element in a specially-designed housing.
Without such measures, the air mass can be measured three times (in, out, and in again) leading to very rich running at low speeds (especially on full throttle) which shows up as massive low-speed over-enrichment.
Another important AFM feature is the mesh or "chip cutter" grille, which is required to collate the airflow as it hits the bridge and even out the airflow across the section of the meter. In the TVR installation, you can almost (but not quite) get away without it due to the fact the air flows up a long straight tube before it hits the meter.
The basic principal of just dropping a bare sensor into the airflow is not a good one, and is never seen on ANY professionally manufactured air flow meter (AFM). This is part of the complex problem of low engine speed and part-load mixture control.
All this information is freely available in any one of the many reference documents on AFM construction, and is indeed borne out by any experimentation. The many years of detailed research results into this topic would be normally be considered essential reading for anyone embarking upon a new AFM design project.
It is certainly beneficial to use a larger section for the AFM than the standard Lucas 5AM for larger engines (4.5 Litre plus), and improvements of around 8-10BHP are seen for 4.5 and 5.0 Litre cars when using the 77mm Bosch Hot Film or Lucas-Sagem 20AM AFM.
Apart from the AFM construction itself however, there is another problem in the software. If you consider the structure of the map (load vs. Speed), it is apparent that the load position in the fuel map is not determined by AFM voltage alone. At low speeds the AFM will simply not produce the full output voltage even on full load. To enable full resolution of the map load scale, the ECU is programmed with the transfer characteristic of the AFM, and scales the output accordingly.
When changing the AFM, it is always necessary to reprogram the AFM transfer characteristic to suit. If this is not done, the ECU load calculation is completely wrong which means that the ECU moves through the map incorrectly. This either means that it never gets to the top line of the map, or it get to the top too quickly. In either case you lose the resolution necessary for accurate part-throttle mixture control, usually resulting in massive low-speed overfuelling.
Although I have seen other AFM conversions there are definitely no modifications to the software I have seen, and I have now seen more than four.
That answers all my doubts about the Torqueflow thingy. Although the water proofing thing put me off big time too. Very educational post 
One thing i remember very well was just how refined my engine was after you mapped it for me Mark! This is the main reason for me not going with the TorqueFlow unit. They will only sell the whole kit and setup themselves with an enrichment chip of some sort? I have a great ECU map. Have a 72mm plenum. Just want a bigger AFM
Regarding the inlet manifold, has anyone tried to improve this major restriction and poor design?
One thing i remember very well was just how refined my engine was after you mapped it for me Mark! This is the main reason for me not going with the TorqueFlow unit. They will only sell the whole kit and setup themselves with an enrichment chip of some sort? I have a great ECU map. Have a 72mm plenum. Just want a bigger AFM
Regarding the inlet manifold, has anyone tried to improve this major restriction and poor design?
V8d,CRE... do ported inlet manifolds, you can open them even more it's just very labour intensive...debaron said:
Chimjunkie said:
I have a great ECU map. Have a 72mm plenum. Just want a bigger AFM
I think Mark 'Blitzracing' Thompson bores out the standard 5AM and recalibrates at a very reasonable cost. I'm sending mine off to him (when I remember!) so might be worth a PM...Ryan
I am so glad an expert has spoken out.
It pays never to hold discussions with a nocturnal monkey when the organ grinder is in the room. Thank you Mark for such a great technical posting.
A wise man learns by his mistakes, a wiser man by the mistakes of others. It might wipe the smile of a Cheshire Cat's face but you can't just bolt something on to a car in isolation to make it into a supercar, it needs evaluation and hollistic treatment or you risk a compromised dogs breakfast.
Mark Adams or Paul Austin are the guys and cheaper in the long run!
ps. What did TF stand for?
It pays never to hold discussions with a nocturnal monkey when the organ grinder is in the room. Thank you Mark for such a great technical posting.
A wise man learns by his mistakes, a wiser man by the mistakes of others. It might wipe the smile of a Cheshire Cat's face but you can't just bolt something on to a car in isolation to make it into a supercar, it needs evaluation and hollistic treatment or you risk a compromised dogs breakfast.
Mark Adams or Paul Austin are the guys and cheaper in the long run!
ps. What did TF stand for?
Edited by retroptvr on Tuesday 2nd June 20:26
retroptvr said:
I am so glad an expert has spoken out.
It pays never to hold discussions with a nocturnal monkey when the organ grinder is in the room. Thank you Mark for such a great technical posting.
A wise man learns by his mistakes, a wiser man by the mistakes of others. It might wipe the smile of a Cheshire Cat's face but you can't just bolt something on to a car in isolation to make it into a supercar, it needs evaluation and hollistic treatment or you risk a compromised dogs breakfast.
Mark Adams or Paul Austin are the guys and cheaper in the long run!
ps. What did TF stand for?
Please explain It pays never to hold discussions with a nocturnal monkey when the organ grinder is in the room. Thank you Mark for such a great technical posting.
A wise man learns by his mistakes, a wiser man by the mistakes of others. It might wipe the smile of a Cheshire Cat's face but you can't just bolt something on to a car in isolation to make it into a supercar, it needs evaluation and hollistic treatment or you risk a compromised dogs breakfast.
Mark Adams or Paul Austin are the guys and cheaper in the long run!
ps. What did TF stand for?
Edited by retroptvr on Tuesday 2nd June 20:26
whos the monkey here then? just curious really 
I now believe there are no issues with damp.
The Polish guys give my engine a good soaking last weekend and its running ok.
As Eann replaced my AFM but could only tune by ear on the day (forgot exhaust probe) then i need to see him for a proper tune up.
Its running a little rich at the moment but im confident this will be sorted.
Its definately got more punch when given the beans however.
The Polish guys give my engine a good soaking last weekend and its running ok.
As Eann replaced my AFM but could only tune by ear on the day (forgot exhaust probe) then i need to see him for a proper tune up.
Its running a little rich at the moment but im confident this will be sorted.
Its definately got more punch when given the beans however.
Why would any one think that they have the privilege to repost a thread that was deleted?
This is going way O/T - If any of you have any respect for the OP you should delete your own posts. Start another thread if you want to b
h about thread deletion.. & BTW you should work harder at remembering what you said MarkB the £700 comment was not called for - practically FA in TVR world 'works out of the box'
This is going way O/T - If any of you have any respect for the OP you should delete your own posts. Start another thread if you want to b
h about thread deletion.. & BTW you should work harder at remembering what you said MarkB the £700 comment was not called for - practically FA in TVR world 'works out of the box'blitzracing said:
Im sorry if I have caused offence to those of a sensitive nature, and please feel free to spend as much money as you want on possibly unproven mods. If you would you prefer I wont pass on things to the TVR community I have leaned about the ECU and fueling (mostly the hard way) and keep it all a black art?
You haven't caused offence to anyone that matters, your posts have always been informative and interesting. The more useful info on here the better IMO
Gassing Station | Chimaera | Top of Page | What's New | My Stuff