This article was first published in 2011.
So far in this series we’ve built a new intake…
… fitted a new (second-hand!) rear muffler…
...and have custom remapped the engine management.
Now it’s time for a new intercooler.
Like pretty well all turbo diesels, the 1.9 TDI PD (a VW engine design) fitted to the Skoda Roomster runs factory air/air intercooling.
The small intercooler, with a core 200 x 200 x 60mm with 50mm inlets and outlets, is fitted in the front right guard and is fed air from a dedicated front-facing vent. In fact, a nicely designed scoop takes air from this vent and directs it to the intercooler. But what is less impressive is the lack of a dedicated exit: it appears that the air that has passed through the intercooler escapes by some tortuous paths – eg back up into the engine bay and out past the undertray.
One approach to improving intercooler efficiency would have been to add some vents in the wheel-arch liner, so allowing air to escape into the (usually) lower pressure area of the wheel well. That’s the approach we (unsuccessfully) tried at Getting Air In & Out.
Another intercooler efficiency booster would have been to add a high pressure intercooler water spray – see World's Best Intercooler Water Spray, Part 1 and World's Best Intercooler Water Spray, Part 2.
However, in this case we decided the best approach was to replace the intercooler with a larger design mounted in front of the radiator.
But to take a step back, why was an upgrade intercooler needed? As described in previous parts in this series, even with the car dead standard, the intake air temp peaked at 67 degrees C – and that’s on a relatively cool day of about 20 degrees C ambient. Following the reflash, which allowed boost to be held all the way to the redline, max recorded intake air temp rose to 82 degrees C!
And yes, that’s after the intercooler…
Monitoring the intake air temp on a continuous basis showed that even with modest power use, intake air temp rose very quickly. It also fell quickly after a boost event, with these two characteristics indicating two things: the intercooler had little thermal mass (it didn’t have much ability to absorb heat) and its heat exchange properties were poor (it couldn’t get rid of enough heat in ‘real time’).
(If you have an intercooler that can absorb lots of heat, peak intake temps will be decreased as the temp spike is absorbed into the metalwork of the intercooler. Alternatively, if you have an intercooler with great real-time heat exchange abilities, the temp peaks won’t occur and so the intercooler won’t have to absorb the excess heat. To have neither a high thermal mass nor good real-time heat exchange abilities means that the intercooler simply doesn’t work very well.)
As described in some detail at
Diesel Intercooling, the requirements for a diesel car intercooler vary from those for a petrol car intercooler. These differences are primarily because a diesel passenger car engine:
is on boost a lot more, therefore the intercooler needs to be able to get rid of a lot of heat ‘real time’
has an average airflow demand that is much higher than for an equivalent petrol engine
The latter’s the case because without a throttle butterfly, at a given boost pressure the airflow is the same (high) amount at all loads. Therefore, despite the 1.9 litre turbo diesel being relatively modest in power output (91kW at the wheels – given the use of the Dyno Dynamics dyno, say 110 or 115kW at the flywheel), it would be very easy to select a new intercooler that was undersized in either airflow or heat-exchange capabilities.
So how to make the intercooler selection? I decided to use a standard intercooler from another car – a Falcon XR6 Turbo. Taking this approach reduces the price by up to 80 per cent.
The VW 1.9 litre turbo diesel is a four-stroke, so if it breathed if it were a 100 per cent volumetric efficient naturally aspirated engine, at the redline of 4500 rpm it would inhale (4500/2 * 1.9) = 4275 litres / minute. (The ‘divide by 2’ is there because a four-stroke inhales only once every two revs.)
The diesel runs boost of about 17 psi, which means the engine sees a bit over double atmospheric pressure. So let’s double the airflow (that takes us to 8550 litres / minute) and then multiply this figure by 0.7 to take into account flow losses of the simple head design.
That gives us the very rough figure of about 6,000 litres / minute airflow.
So how does a potential donor car intercooler compare?
The Falcon XR6 Turbo is a car that, here in Australia, often has its factory intercooler upgraded. The old intercoolers are then sold cheaply – in fact I have one that was purchased for nearly nothing on eBay. The Falcon engine is 4 litres in capacity, revs to about 5250 rpm and runs 6 psi boost. Do the thumbnail figures and you get something like 16,000 litres / minute airflow. That seems plenty!
(Note: these figures are very rough – all we’re trying to do is get a feel for the differing flows. They’ve been based on airflow and not power because the diesel’s power is limited largely by fuel flow - so the power figure of the diesel doesn’t reflect airflow in the same way as it does for the petrol engine.)
So based on that back-of-envelope calculation, the XR6Turbo intercooler looked good for this application. But would it fit?
There are two aspects to fitting a new intercooler: finding space for the intercooler core and end-tanks, and getting the large diameter plumbing to and from the intercooler.
In the case of the Falcon XR6 intercooler, the core just fitted under the front bumper beam.
Rather than use the original Ford mounts that were not in the correct places, I cut most of them off and made up some steel brackets.
I covered the brackets in foam pipe insulation, allowing the intercooler to move around a little as it expands and contracts with heat and cold. (Pretty well all factory intercoolers are rubber-mounted.)
The intercooler brackets bolt to the bumper intrusion beam – a very tough piece of high-strength steel that was difficult to drill!
Here’s the rear view showing how the intercooler is attached to the bumper beam.
The feed pipe to the intercooler was routed through the space vacated by the missing factory intercooler.
The return pipe to the engine was placed on top of the bumper beam…
…and then passes back into the engine bay under the headlight. This is a tight squeeze – but the only cutting needed was to shave the folded lip on top of the bumper beam.
A factory sensor, normally mounted on the intercooler exit pipe, had to be integrated into the new plumbing. A steel mounting block (red arrow) was made with two tapped holes into which the retaining screws are inserted. The wire harness needed to be extended about 10cm. Note also the foam rubber cladding (green arrow) used around a section of the intercooler plumbing to stop any rattles.
I made the plumbing using mild steel, mandrel bent exhaust bends, fusion welded together with an oxy-acetylene kit. Adaptors to match the 2 inch plumbing to the 2¼ inch intercooler connections were made by an exhaust shop.
Said quickly it all sounds easy but in fact fitting the intercooler was a major job that took quite some days.
To ensure that the airflow passing through the front grille makes its way through the multiple front heat exchangers (intercooler, air con condenser and radiator) and doesn’t escape sideways, foam rubber side pieces were added. These foam pieces were spray painted black and were glued into place with contact adhesive.
With the bumper cover back on, there’s no indication of the new intercooler.
The first few tests were problematic. First up, the hose before the intercooler blew off under boost. I tightened the clamp only to have the same thing happen again. I then welded lips at each end of the steel pipe section and the problem was solved.
On the road, the intake temperature decreased to a maximum of 30 degrees C above ambient – that compared with the 60 degrees C above ambient previously recorded. In normal, undulating country road driving, where at 100 km/h the car can be on 6-15 psi boost continuously, the intake temp sat about 6 degrees C above ambient - an excellent result. (Note: all intake temp measurements made by the factory sensor located immediately after the intercooler, with the data accessed through the OBD port.)
That was all good, but then I started assessing the on-road performance…
On the Road
The first thing that I noticed was that the car felt no faster! It felt the same as it came on boost; it felt much the same in the midrange and the top-end.
It was time to get out the accelerometer. As with previous parts in this series, the actual on-road acceleration was measured at full throttle right through second gear.
And what it showed was that fitting the intercooler had apparently caused performance to go backwards. As this graph shows, with the intercooler fitted (black line) performance was down over the previous iteration right through the rev range, in most places back to standard! Only at 4000 rpm did it return to the previous level, but at 4500 rpm it was again down.
So what was going on?
My mind returned to the initial test drives with the new intercooler fitted, where the hose in front of the ‘cooler blew off a number of times. Despite being held in place in exactly the same way, no other hoses had come off. So perhaps the intercooler was in fact restrictive, and that explained the lack of performance improvement and the hose coming adrift?
I removed the plumbing in front of the intercooler and silver-soldered a brass hose tail to it, allowing measurement of pressure. I already had a similar fitting attached to the plumbing after the intercooler. I then hit the road, measuring boost pressures before and after the intercooler. I did multiple runs, swapping a single pressure gauge back and forth between the hoses and so cancelling-out any error that might have been introduced by using two different gauges.
These tests showed that the pressure drop across the intercooler was quite small – at max power it peaked at only 1.5 psi. On a boost pressure of 17.5 psi, that’s an absolutely acceptable drop across the intercooler.
So the intercooler and its plumbing were flowing well, and the temperature reduction it was causing in the intake air was far greater than achieved by the standard intercooler. That adds up to a denser charge, one containing more oxygen - and so with the addition of more fuel….
But perhaps just as had occurred with the new intake and exhaust, maybe the intercooler’s effectiveness was not being compensated for by the engine management – maybe it was running leaner? You’d think that with the engine management having an intake air temp sensor it would know what was going on, but perhaps those maps overseeing ‘smoke’ and ‘torque limiting’ were capping the fuel additions, so the lower intake temps were not causing any extra fuelling to occur?
It was time to go back to the ESP dyno. However, the dyno run showed that the power curves before and after the fitting the XR6 intercooler were extremely close – in fact less than 1kW different in peak power and within a few kilowatts right through the rev range. Air/fuel ratios were also the same or richer, and the boost was unchanged. (The funny peak as the engine comes on boost is not replicated on the road.)
So perhaps my road measurements had been in error? I did another accelerometer run on the road (orange line) and found that, over much of the rev range, measured acceleration was still down by about 5 per cent. But the dyno had showed the same power before and after…
So perhaps the problems were as simple as my on-road measurements being in error? Or maybe other factors had changed?
Possible variations that occurred between the on-road runs included differing levels in the fuel tank, differing ambient temps and atmospheric pressure, and differing road temps. I took care to minimise these differences but realistically the runs would not have been identical. An oil change also occurred along the way.
Also, during the modification process, the front brakes were increased in size, so increasing rotational mass - this would tend to slow acceleration.
So if there was no power increase, was the fitting of the larger intercooler a waste of time? For two reasons, I don’t think so.
Firstly, the fuel economy is clearly improved. This change is not visible in all driving: it seems more pronounced at higher speeds where loads are greater. In Part 4 of this series I said that typical fuel economy had improved to about 5.3 litres/100km. That is still the case in most circumstances, but in freeway use at a sustained 110 km/h, 5.1 litres/100km is now common.
Secondly, where previously the car could be felt in certain circumstances to get intercooler heat soak and so become quite unresponsive (a good example being climbing a winding mountain road with lots of second and third gear use), performance in these conditions now doesn’t drop away.
When you look over the modifications that have been made in this series, there is only one step that was unambiguously good in terms of on-road performance – the Powerchip reflash. Fascinatingly, the intake, exhaust and intercooler modifications all caused measured decreases in road performance! (But each of these steps improved fuel economy.)
However, it is very likely that the final outcome (in terms of performance and fuel economy) could not have been achieved without the complete suite of modifications being undertaken. What is clear, though, is that on a diesel like this, the custom reflash should always be the last step – the one done after all the other modifications have been completed.
So what have we ended up with?
Peak power at the wheels has gone from 77kW to 91kW – a gain of 18 per cent. However, in the mid-range power is typically up by about 20 per cent and in fact power has improved at every point in the rev range between 1000 rpm and the redline of 4500 rpm. Boost also arrives a little earlier in the rev range.
Very low rpm tractability is also clearly improved, allowing the car to be driven a gear higher in many urban situations.
Measured acceleration on the road has improved at every engine speed above 1000 rpm; at the redline the acceleration is stronger by an incredible 60 per cent! It now makes sense to hold the engine to the redline of 4500 rpm, rather than changing at 4000 rpm.
Fuel economy in my normal driving has improved from about 5.6 litres/100km to 5.3 litres/100km. In real-world freeway use (110 km/h, hills) the fuel economy can now be as low as 5.1 litres/100km. This spread of figures represents an improvement in fuel economy of 5 – 9 per cent.
Other than the reflash and fitting the muffler, I did all the work myself. The muffler and intercooler were cheaply sourced second-hand, I made the intercooler plumbing (a second version is shown here; I wasn’t happy with the first version), and the intake was made from cheap plastic pipe. As noted previously, the reflash was performed by Powerchip at no charge, however their retail price is AUD$1290, including GST. Therefore, at full rates, the complete modification process (new intake, new muffler, reflash and new intercooler) would have cost a total of about $1500 - $1600 (and quite some time!).
For that money the modification outcome is simply outstanding.