Tech article #2 - The HUMVEE Diesel Engine Cooling Paradox

Engine cooling systems shared a parallel development path since the very first internal combustion engines were developed. From the first in 1794 with Robers Street’s internal combustion gasoline engine and more specifically to this discussion later in 1926 with Rudolf Diesel’s (image below) compressed charge compression ignition engine commonly named after its inventor as a Diesel engine, cooling has been paramount. The two main types of cooling were liquid based and air/oil based. Liquid cooled engines were the logical evolution from purely aircooled engines which reached their peak of development in the late 90's with Porsche, the last real standout with air cooled technology. Liquid was able to cool more consistently and with less complicated engineering than that required with air cooling technology. However the relative ease of using liquid cooling led to short cuts being made for the sake of packaging and reduced costs. Engineers were asked to simplify the cooling systems and their design so that the same family of engine could be adapted more quickly and easily to multiple platforms. The end result was, and is, a common configuration among most liquid cooled engines, where coolant is fed into the lower engine block at the front and drawn out again at the front top of the heads.

To understand the process of how liquid cooling works we must understand the dynamics involved. As coolant makes its way through the engine its temperature is raised through a process called heat conduction or diffusion (image below), whereby the engine and coolant are attempting to reach a point of thermal equilibrium. In other words they want to be the same static temperature. If you leave a glass of water outside long enough it will warm to the temperature of the air around the glass. Since a running engine loses some of its combustion energy in the form of heat (to a lessor extent heat is also added through friction) conducted into the block, it has a heat source external to the coolant adding heat. The coolant also has a cooling source external to the engine when it passes through the radiator conducting its heat into the atmosphere. This process of heat conduction is what is supposed to keep your engine in a tightly controlled band of operating temperature. Additional devices in the system regulate the amount of cooling taking place in this process.

One of the most important devices in an engine cooling system is the thermostat, common in nearly all liquid cooling systems. Automotive thermostats are generally wax thermostatic elements (image below) which as their name implies function by wax melting and solidifying inside an element which when reaching a designed temperature expands and contracts actuating a valve to allow more or less coolant to flow past it. Since thermostats are a restriction to coolant flow, the higher the temperature the more they open until the full open position. During this process they regulate the flow from closed to open. One important, and often overlooked dynamic, is conductive saturation time. Ever since we started modifying cars, racers have surmised that removing thermostats would improve cooling and remove one more potential part to fail causing a racing DNF (did not finish). However depending on the complex flow dynamics in each system this in fact may reduce cooling efficiency. By removing the thermostat more coolant will flow due the the reduced restriction of having to flow through the thermostat. This increased flow sounds good but it also increases the velocity of the coolant and as such the coolant spends less time in the radiator. Less time in the radiator means the coolant can only pass a limited amount of heat to the outside air before returning to the engine coolant inlet with limited temperature reduction. Racers were overheating and have since introduced restrictors which fit into the location of the thermostat and come in various dimensions to tune the coolant saturation time in the radiator to achieve a balance of improved flow and thermal conduction.

Another important device in an engine cooling system is the cooling fan. Nearly all road driven vehicles have some sort of fan to aid in heat conduction from the radiator coolant to the outside air. These are there to supplement the normal flow of air during vehicle operation for times when passive airflow is insufficient to properly cool the coolant (high loads at low speed, or in the case of the HUMVEE also a poor radiator air flow design). Fans can be mechanical, pneumatic, or electronic and can be actuated through many different clutch designs. Most are mechanical clutches, which like a thermostat regulate, in this case from a free wheel to a full speed lock (turning at engine speed plus or minus pulley ratio). Unlike a thermostat however, using a viscous coupling with a bi-metallic strip to regulate from a free wheel state through to a locked ‘full speed’ state. Electronic fans are controlled by electronic temperature sensors and a fan controller which electronically switches the fan(s) on and off and in some more modern examples can even control the speed the fans operate at (Pulse Width Modulation or PWM technology). Many vehicles use a combination of mechanical and electronic fans. In the case of the HUMVEE, its system uses an electronic temperature sensor and electronic control to actuate a hydraulic valve (Cadillac valve) which engages or releases a hydro-mechanical fan (the HUMVEE fan moves an immense amount of air however also requires significant mechanical energy from the engine to operate). Image below is of overheating HMMWV in Afghanistan.

The ideal temperature operating range for a diesel engine is far narrower and more critical than that of a gasoline engine. Diesels use heat as part of the ignition equation along with compression and air. Diesels actually have two ignition events: an ignition delay period or pre-ignition event (not to be confused with the undesired gasoline engine pre-ignition or detonation), and a rapid ignition/combustion event. The ignition delay period is the 'spark' that ignites the rest of the combustion process. In order for the delay period to reach its combustion point, the fuel and air must have enough temperature and pressure (compression) to start a cascading process of burn in the ensuing rapid ignition/combustion event. So contrary to convention, colder is NOT better. Heat is good in a diesel engine and without it can never achieve maximum power and energy from the fuel injected into it. However with all good things moderation is required. If a diesel engine is allowed to run too hot, outside the ideal operating range, it again falls out of efficiency/power and in most cases loses reliability. If too much heat is introduced into the combustion chamber the ignition delay period can ignite too soon. Like a gasoline engine, you do not want to have the rapid expansion of a burning combustion event with its high pressures and temperatures meeting a fast moving piston too early while still on its way up in the compression stroke. The piston and expanding combustion meeting this way often end with failed pistons (example picture of failed #8 piston and cylinder below), piston rings, rods, crankshafts, head gaskets, heads and the block (example of cracked block main bearing web picture below) and cylinders as well. Additionally too much heat can lead to lubricating oil losing its viscosity and ability to keep rotating assemblies from contacting each other. Too much heat expands metals beyond their designed tolerances which leads to added friction (adding even more heat) and possible failure.

Now looking specifically at the Detroit Diesel developed and later Chevy/GMC and further still GEP (AM General) branded small V8 engines in both civilian and military configurations we see the same common cooling system packaging. This is true with all variants including 6.2 and 6.5 and in turbo or normally aspirated versions. Coolant being pushed or pulled (both are true being a closed loop system) by a water pump enters the lower forward block. Coolant is first passed through the cooling jackets around the cylinders to cool them and then up into the head(s) before being passed back out the again at the front of the engine. However to truly understand this flow pattern requires visual depiction as provided below. What must be understood is that the coolant does not pass from the front of the block to the back of the block completely and then up to the heads and forward out the front. Coolant is actually passed into the front of the block but can immediately pass vertically up into the heads right away from the front of the engine. Since the flow is exiting the front of the heads, the path of least resistance is the shortest distance, therefore the majority of coolant enters the front of the block and turns vertically up into the front of the heads and back out to the radiator. This ease of flow back out the front of the heads causes the forward most cylinders to run coolest. Subsequently, with their reduced cooling flow at the the rear of the block and heads the rearmost cylinders run hottest.

This phenomenon of overcooled front cylinders and overheated rears is common among most engines with this front in/out cooling configuration, with some being more susceptible than others. Only now some of the most recent 'next gen' engines with the help of modern computer modeling are now restricting flow at the front of the block by changes to the casting of the cooling ports to facilitate flow at the rear of the block reducing the inadequate coolant flow there. Older engines like the Detroit Diesel small V8s have no such added engineering to balance the cooling across the block and heads. In order to live with the in-balance created, manufacturers adjust the cooling system (thermostat temperature and restriction, radiator size, fan(s) size and operating temperature, water pump, etc.) to a best overall average where the center block area is the theoretical ideal and the inherent under-cooling of the rear and over-cooling of the front is an accepted compromise. Unfortunately in the case of the 6.2/6.5s this in-balance was particularly grave. The rearmost cylinders run so hot that minor imperfections in the original design of the earliest engines, the 6.2 block and 6.5, became catastrophic failures of the cylinders, cracked heads, blown head gaskets and spun rod bearings, crank failures, and cracked main bearing caps and block area. In an attempt to reduce these failures each subsequent engine redesign had strengthening added, compression ratios reduced and even piston to cylinder clearances enlarged just in the rearmost cylinders to accommodate the over expansion due to heat there. Often not discussed is the loss of power and efficiency experienced in the forward most cylinders due to over-cooling there. Since the ideal operating temps are never reached, excessive smoke and inefficient burn are experienced in those cylinders. While not as notable as a catastrophic failure to the assembly, reduced MPG, HP and higher particulate emissions are unwanted side effects of over-cooling.

Most engines draw coolant to feed the heater core from the rear of the engine to tap into the overheated coolant there. The obvious benefit is a faster cabin temperature rise and a higher overall heat output. The additional benefit is reduced temperatures in the rear cylinders during use of the heater core as it draws heat away where there was once little flow. Of course in the summer months there would be no such benefit. In the HUMVEE however the heater is fed from the front of the block and not the rear. This causes the heaters to take longer than necessary to warm and lower overall temperature.

So herein is the paradox, how can an inherent core engineering defect be rectified without a complete redesign of an engine? The solution has been utilized in the motorsports industry for decades, a brainchild of early engineers that were tuning racing engines to make maximum power. Those engineers found that they had to tune mixtures lean and at the verge of detonation to get the most from their power plants. The closer they got to peak power the more they started seeing failures at the rear of the engines. They realized the rears were running too hot to function at the lean mixtures the rest of the cylinders could run at safely. After years of testing and tuning they started to feed coolant through manifolds which had variable sized restrictors to feed each section of the purpose built racing engine separately to balance the cooling needs across the block and reach a balance so that every cylinder could operate in its ideal temperature range. While a complex and expensive solution, it worked. However this solution was not possible in all forms of motorsport where restricted classes like NASCAR, and Grand Touring, had to maintain the factory produced block. Early greats like Smokey Yunick (image below) and Bill Jenkins, two notable engineers in those early days, started externally feeding or drawing coolant to/from the rear of the block to balance out the temps. Both discuss their findings in their respective books, but to summarize, they were successful.

Enter Paradox by Design. Our background is also in the motorsports industry where we too used this ‘fix’ to run factory engine blocks at as much as 4X the original output in HP and TQ and doing so for 24 hour endurance races. While the HUMVEE is about as far as you can get from a multi-million dollar endurance racing car, we are still engineers and we don’t often sit idly by when a solution is available to fix an engineering issue. Yes, we are the guys with modified lawn mowers and little around the house is stock. So when we took a liking to the HUMVEE in 2009 we also started working on ways to improve them where we could. Much of what we have done is not available for sale and may never be. We did not start our work on HUMVEEs as a business venture but as enthusiasts which is still our focus. However we did start to help our local off-roading friends with issues that came up and as they saw things we were doing they wanted the same for their trucks. Thus Paradox by Design was born. We are not a large shop and perhaps that is why we are able to do what we do. The overhead is low and we follow a simple mantra, if we do it for us, offer to others too. So after finding the rear cylinders running 40-50 degrees hotter than the front we started to investigate solutions to fix the issue. We first sought examples already in the market and found two. Both were identical so we assume one copied the other. However we immediately saw a flaw in those designs, where they used the heater circuit to return flow from the rear of the block. The inherent design flaw is that without the heater ‘ON’, those systems never reach proper flow. While running the heater would attempt to balance the flow, it is likely that users may not want to have the heater on in the summer! Then there was the poor quality of those kits, requiring cutting into hoses and using hose clamps. Then to add to it all they were well over $300, which made our decision to design, test and build our own a forgone conclusion.

What we came up with is a simple yet effective solution, by opening up a path for coolant to flow from the rearmost area of the block directly to the coolant return in front of the engine (depicted below). The front crossover is where the pump is drawing coolant from, so now a flow pattern is created where once stagnant coolant at the rear is flowed to the front. Once this flow is established the removed coolant is replaced with coolant from the feed at the front lower entrance of the block which has to pass down the length of the block to do so. This continuous flow balances out the temperatures across the block and heads where it did not before.

We use all aerospace quality AN fittings, and in the latest kits, stainless steel braided hose that is covered in a soft outer layer to keep from chaffing any underhood parts it comes in contact with. We machine our own custom rear plates which replace the dead block off plates in these engines with an aluminum CNC machined plate with an integrated AN fitting. Everything is bolt on and is of the highest quality. No hose clamps and hoses to cut and splice here. These kits are bolt-on and can be removed if you sell the truck. Image below is a kit for a Civilian 6.2ltr or military CUCV.

One question we get asked is why the are the internal flow dimensions are relatively small? The size is a result of our testing. We started larger and worked our way up and down in orifice size to find as close a balance as possible. Too large and you can actually reverse the results and make the rear of the engine run too cool. It does not take a lot of flow here to achieve the desired results. Image below is of our civilian 6.5ltr kit.

After the extensive development and testing which led to these kits the engines now enjoy a near perfect balance of temperatures across the engine, which reduces the failures common in these engines and improves the efficiency of the engine. Some notable positive effects from these kits:

• Coolant temperatures are balanced across the block and heads
• Engines reach operating temperature faster
• Heaters reach operating temperature faster and produce higher heat levels
• Overall smoke output is reduced
• Most see fan engagement cycles reduced and duration of operation lower
• Engine noise is reduced
• Cockpit temperatures are lower
• Slight HP/TQ increase as well as MPG improvement
• Reduction of steam pockets and stagnant aerated coolant
• Constant bleeding of the cooling system to the overflow tank
• Reduced common failures of the block, heads, and head gaskets

While this article is focused on the HMMWVs, the same basic dynamics are true with CUCV trucks as well as civilian vehicles with the 6.2 and 6.5 engines.

A parting image below is a schematic of the cooling system of the GM LS based engines. These engines replaced the original GM small block V8 design which had been in production for generations. There were many improvements, too numerous to list here, but one thing that was not addressed fully in the redesign was the inherent cooling imbalance. Very soon after production started, GM realized they needed to better distribute coolant to the rear of the block and heads. Their solution is a hose assembly from the rear of the engine on both sides leading to the front of the cooling system. This system then ties into the return or suction side at the radiator so that coolant could be drawn out of the back of the block and heads. This system helps to balance the running temps across the engine. This is GM’s version of a cooling balance kit which they call a steam vent.