Image from kawasakionlineparts.com.au.

Cooling system for Dummies – my explanation for those who have NO IDEA how it works!

Note on de-icing.  Some bikes have an additional circuit that comes from the outlet pip (32033) to run coolant to the carbs to prevent icing in cold climates (not discussed here).

Cooling is not the best description – because we generally think of cooler as being a temperature LESS than something else.  So a cooling system makes things colder right?  Well, no.  The role of the cooling system in an engine is actually to hold the engine at the ideal operating temperature.  And for a performance bike that means when it is going pretty fast.

But cooling is the common phrase so I’ve used it here.

Air cooling (used for many older bikes) is far, far simpler – no seals, no water, no hoses, pumps & components, etc.  A major reason reason why the VW Beetle was so successful in it’s era of pre-emission legislation!  Suzuki even developed a system where the oil was used even more creatively.

This article nicely explains the Suzuki system, hilariously also the ‘chip-on-the-shoulder’ other marquee devotees obviously suffered from with the public impact of the GPZ900R.

At the time, many people called sportbikes “Ninjas” no matter the brand, but the GSX-R750 was a real sportbike while the Ninja was still just a water-cooled GPz with body work.

But water absorbs heat better than air, so water cooled engines can run hotter and be made smaller – which is more efficient.  But ultimately all engines are air cooled!  This is because even liquid cooled engines move heat from the engine to water and then from water to air.

For the GPz900r a major advantage of water cooling was size.  As the engine is cooled by internal water channels the pistons are able to be arranged closer together, and along with moving the cam chain this reduced the width of the engine significantly by 5″.

How does it work?

Think of an electric kettle. Yes we think of it as heating the water for a cup of tea, but another way of looking at this is that the water is cooling the hot electric element.  This is a fluid heat exchanger – heat is exchanged from the electric element to the water.  One goes down (or stays the same) whilst the other goes up.  The larger the water source means greater ability to absorb heat, why large kettles (or ig pots on the stove) take longer to boil than small ones!

If we let the kettle boil then water is converted as steam and dissipates into the atmosphere, which reduces the volume of the water and further reduces it’s ability to absorb heat.  This is the start of a downward spiral until all the water has disappeared.  With no fluid for a heat exchange the element will just get hotter & hotter until it destroys itself – hopefully not burning down the house in the process.

Water-cooled engines use exactly the same principle for the first part of their system – the engine is the hot element and water is used to absorb the heat. But that is only half of the engine cooling system, because the water would just get hotter and hotter until it also boils away and the engine destroys itself (like the kettle).

So the engine liquid cooling system has TWO parts – cooling the engine and then cooling the water.

Water runs from bottom pipe of the radiator (30960) to the engine-driven water-pump (49044).  The pump creates the water path that circulates through the different parts – connecting them all with pipes & hoses and hose clamps..  Water from the pump is pushed to the water pipe (32033A) that splits the single water flow into 4 ports into the front of the engine.

This is the first part of the cooling system – the engine temperature is controlled by transferring heat to the water (heat exchange #1) which then exits from the back of the engine into another water pipe.

From here heated water hits a temperature sensitive valve – the thermostat (49054).  The role of this is simply to minimise the time to warm up the engine and if operating properly it has no effect on the normal cooling system processes.  With the engine warm and the thermostat open the water flows to the top of the radiator, and then moves through lots of internal channels within the radiator before reaching the bottom. (heat exchange #2).

An issue when the engine is cold is water pressure.  Just like the garden hose when the tap is on (so engine running and waterpump spinning) but the nozzle off pressure builds up inside.  When speaking to an engineer he was expecting a discrete bypass system, however the GPz900R system does this simply with a small hole on top of the thermostat.

The radiator is the second part of the cooling system, A basic understanding of how it works is demonstrated by the household fan. When air blows onto your face we feel cool, even though the air is exactly the same temperature.  It achieves this by evaporation, so removing heat = feeling cooler.  Although a slightly different process, an engine radiator also needs moving air to exchange heat from the water to the air and thus operate effectively.

This water cycle then repeats, and as long as heat exchange #1 is LESS than heat exchange #2 the system the water will not boil and prevents the engine from overheating.

Temperatures above 100°C

Unfortunately the optimum engine operating temperature is typically above boiling point of water – so higher than 100°C or 212°F.  Life would be so much simpler if it were lower!   So we want the coolant to be over 100°C but not boil……

We achieve this in a couple of ways.

Coolant – we mix other compounds into water (like glycol) to make coolant, which increases the boiling point as high as 109°C.  But there is a limit, the coolant still needs to flow easily (viscosity) and not be corrosive to the engine nor be unduly poisonous to human beings!

And this still isn’t high enough for the GPz900R system.

Pressure – an interesting real-world observation is that if you boil water on Mt Everest the temperature your cuppa will be just 70°C or 158°F!  Still pretty warm considering, but a lot lower than normal.  This is because temperature is directly proportional to pressure, the lower boiling point on the mountain is due to lower atmospheric pressure at high altitudes.

And the reverse is true, so if we increase pressure we can increase the boiling point of the coolant.  Technically you can superheatwater to an incredible 374°C or 705°F! 

So a closed system would theoretically allow the pressure to build up indefinitely, in reality this would eventually burst through a join or seal.  So the system controls this with pressure valves inside the radiator cap that allows excess coolant to overflow into a bottle, and then draw it back again as the system cools.

Overflow vs Expansion Tank

The GPz900R has an overflow tank (43078). 

Sometimes the two are confused because they do both relate to the expansion of coolant as it heats up, this is a nice quick video that explains the difference. One of the most obvious visual differences is because an expansion tank is pressurised it will typically be made of aluminium (though can be plastic), to save weight a modern overflow tank will nearly always typically be plastic (no pressure) which also means it can be moulded to fit efficiently into tight areas (like a bike).

IMHO an additional expansion tank would be a good upgrade, as this increases the actual volume of coolant in the system.  There is even a bit of space above the radiator where an additional tank could fit.  So an option for racing or bigger engines, however not sure it’s really necessary for a standard bike.

With the ignition switched ON the GPz900R radiator fan only switches on when the coolant temperature at the thermostat housing exceeds 110°C.  If it cools down below that – then the fan switches off.  According to an engineer the average operating temp is usually +10-15°C above the thermostat (82°C), obviously being at the top of the circuit and just out of the engine the thermostat location is a hot spot.

This all reinforces the performance heritage of the bike.

According to a 1984 article the coolant temp when the gauge is at normal (12 o’clock) was 220°F (104°C) and nudging the red zone at 235°F (112°C).  So because the fan only switches on at 110°C this was a bit concerning to owners, hence the addition of a resistor (PN: 28018-1052) into the circuit that simply moved the gauge needle *left* by about 5°C (10°F) thus making it appear like the fan is switching on earlier!

When on the move the cooling system is perfect, the radiator size & coolant volume are all, unsurprisingly, designed to minimise size & weight and to perform effectively when moving.  In the cooler climates of Japan & Europe I’m sure she’s fine at low speeds as well, in fact in many locations carb icing due to low temperatures was far more of a concern for early bikes.  Where she gets a bit-hot-under-the-collar is when the ambient temperature is hotter, the engine (and water pump) are operating at low speed, and there isn’t the natural airflow to:

• optimise heat exchange through the radiator and;
• carry radiated heat away from header pipes, exhaust and the engine itself

This is the role of the electric fan, however as it’s size is limited there are clearly times when this isn’t enough cooling for many owners.

Of course when brand new it may have all been fine.  All systems lose efficiency over time, corrosion or gunk can impede water flow through the radiator and the engine, the water pump might lose some efficiency, the thermostat might not fully open impeding water flow, and as it wears the engine itself might run a little hotter.

So I suspect that in today’s warmer world even a pristine 30-40 year old cooling system could experience over-heating issues in the city. And on the stinking hot days (+38°C/100°F) after a good ride (so engine hot) it probably only takes a few red lights before you begin creeping up into the danger zone.

.

When I’m on the move my temperature gauge now sits steady at 12 o’clock.

But for the previous 5 years I’ve owned the bike the normal gauge position was the red mark.  It was only after it all began to go haywire (see temp gauge glitch) and I had to investigate cooling system components that the needle now sits here.  Now it could be reasoned that the default gauge position is 12 o’clock and I agree that this is normally true.  However my big Ford BA Fairmont temperature gauge sits steady at 1/4, the only time I’ve ever seen it get to 1/2 is a 45°C day in the middle of Australia with the air-conditioning running almost full blast…

My point is unless you buy something new then you won’t know what WAS the original normal.

Apr 2021 update: reading an article from the Performance Portfolio suggests that 12 o’clock WAS the normal operating position and measured 200°F (93°C) at just before the red zone 235°F (112°C).  So you can see the OEM fan only kicked in VERY late at 110°C!  However Kawasaki soon modified the circuit (added a resistor) so the gauge shows the bike running ‘cooler’ to hopefully alleviate owner concerns with a rising temp gauge.  So 12 o’clock now becomes 210°F (99°C), and assuming later bikes kept this same temp display range then the normal operating temperature position should be BEFORE 12 o’clock.  

The gauge measures the thermostat mounted sensor’s resistance to ground – so corrosion on my wiring terminal increased resistance at that join gradually reducing my gauge reaction to the point where it only showed 1/2 yet would boil over.  Nothing major (leaking hose clamp) but that was just good luck, I never would have seen the big one.  Interestingly the sensor (PN:92066-1183) is the same for all bikes, so the Kawasaki changes to the gauge’s reading position must happen in the circuit (resistor) or perhaps at the actual gauge itself for later bikes.

Engine Guard

I’d considered replacing the gauge sensor with a proper temperature sensor and gauge that gives actual coolant temp, but then this won’t work with the original gauge, so maybe just a second system is a better option.  I was looking at VDO etc. and have just found this (Dec 2020). This is a local Aussie supplier engineguard.com.au with a flexible multi-sensor system (inc low oil pressure) for a very fair and reasonable price – just AU$99.  If anyone has used it let me know.

The  complexity with the GPz900r is that some or all parts of the cooling system could be the cause of over-heating. As mentioned I doubt it was really an issue with a new bike, but none of us have new ones anymore!.  So problems are rarely easy to identify as they can be hidden (eg hose clamps)  or just not obvious (e, temp sender wire corrosion)

So IMHO the only way to approach over-heating issues is to first confirm the basics – make sure the system is pressurising and there are no leaks from any joins when the system is hot.  As I don’t have a pressure gauge the only way is to eye-ball every single join and there are several hidden deep under the fairings and tank.

Make sure she is properly filled (and bled) with coolant (or even just distilled water for testing) and confirm the water pump is working during the bleed process.  How much is normal?  Well for my bike at idle (and thermostat closed) nice strong steady stream comes out when you break the bleed nut seal.  If it’s only dribbling out with the nut clearly open or surging then I would suspect a faulty waterpump.

Then its fix by replacement.  What mean by this is just take a best guess as to which component is faulty and check it is operating at OEM spec.  If you can’t check it then I  simply replace it.  There really aren’t that many components in the system, and even if it doesn’t fix the problem the you know it’s not the fault of this part, and the new part will probably work better than the old one anyway.

If you end up replacing every cooling system component and still have over-heating issues – that sucks.  However it likely means that you have bigger problems – which are going to be way more expensive than the cooling system.  And when those bigger problems are fixed you have a new cooling system will not only work better but should help prevent the bigger problems from happening again.

This is a great explanation of the electrics inplay with the cooling system.  Because I don’t want this to be lost I have cut and pasted from kawasakimotorcycle.org, please visit forum to read in context.

GPZzone also looks like it ahs the exact same information page.

GPZ 900R COOLING FAN SYSTEM
Send the Missus and kids on a day trip to the Outer Hebrides lock yourself in a room and have an ice-cold beer at your fingertips – you’ll need it! The cooling system on the 900R is fairly complex and cleverly designed, but with a little patience it can be understood to help with those all too common overheating problems – I’m sure that we’ve all had the temperature gauge in the red at some stage, and wondered why the fan hasn’t come on !
The following text hopefully will help you grasp the logic and operation of the system to enable you to tackle any possible faults with gusto and confidence ! Good luck !

The cooling fan circuit consists of three temperature sensors, two relays, and of course the fan. The system automatically provides three modes of operation and to differentiate between them we will consider these as ignition off, ignition on (normal) and ignition on (standby). Whereas each mode utilises different sensors and power supplies, we shall see that the two relays are powered and in use constantly. Let’s now look at the modes of operation:

Ignition off Mode

With the ignition off, the first temperature sensor, the 97 C switch (Motoshoot note: on the LH side of the radiator) is armed and if the coolant is at 97 C or above, the fan will be commanded to run until the temperature falls to below 97 C. (Note – this sensor is not active with the ignition on during normal operation.)  It is therefore normal for the fan to run when the ignition is turned off with the engine hot.

Ignition on (normal) Mode

Throughout the ignition on (normal) mode, the 97 C switch is disabled and the second sensor (110C) on the thermostat assembly is armed. This will signal the fan to run if the coolant temperature increases to 110 C. Once the temperature decreases to below this threshold, the fan will cease.

Ignition on (standby) Mode

For the last mode, the third sensor actually measures oil temperature and switches at 120 C. This appears to be a fail-safe feature which switches temperature sensing from the 110 C switch to the 97 C switch if:
1. The 110 C switch is defective (we have to assume that if the oil reaches 120 C, the coolant temperature is above the 110 C threshold, therefore the 110 C sensor has failed).
2. The oil temperature rises due to a problem in the oil cooling system.

Under either of these conditions, the 97 C sensor will be activated to prevent cooking your engine. This will command the fan to run until the coolant temperature decreases to below 97 C.

Relays

The first of the two relays, the Fan Switch Relay is situated adjacent to the headlight and determines which sensor is used under the prevailing operating conditions. It routes an earth signal (if temperature threshold exceeded) from either the 97 C or the 110 C sensor to the Fan Relay (situated adjacent to the fuse box) to bring on the fan. Failure of either of these relays will render your cooling fan system inoperative.

Sensors

Both of the engine coolant temperature sensors function in exactly the same manner, in that the resistance is inversely proportional to temperature – in other words the higher the temperature the lower the resistance. This is in contrast to the oil temperature sensor, which works in the opposite way – the resistance is conversely proportional to temperature, the resistance increases with temperature. The coolant sensors provide an earth when hot, the oil sensor provides an earth when cold.

Remember this – it is invaluable when troubleshooting.

Troubleshooting the system

Before you go tearing the bike apart, first determine if there is a fault – does the fan run when it is supposed to ? If you have read this far you should a fair idea of what is supposed to happen and when. Let us start with a very simple check which does not require fairing removal etc, but it will prove about 90% of the cooling fan system system.

Turn the ignition on, (no need to start the engine) disconnect the black wire from the 110 C sensor on the right-hand side of the thermostat and ground that connector to earth. The fan should run. Refit the wire back on the sensor. Be careful not to snap the terminal off the sensor, they’re delicate!)

If a fault persists, and you suspect component failure, then the place to start is the fuse box, check, the main fuse, the horn fuse, and the fan fuse, each of these has a role to play in the cooling fan circuit.

Remember that overheating can be caused by problems other than component failure, such as, chafed wiring, air locks, sticking thermostat or a particularly bizarre electrical problem.

Two 3ft lengths of wire with crocodile clips at either end is invaluable when troubleshooting, as they enable you to provide power from the battery, to virtually any component on the bike.

COOLING SYSTEM COMPONENT BASIC FAULT-FINDING

Remember that the status of the ignition switch is critical during most of these checks.

To prove the fan. and To Prove the Fan Relay (Situated adjacent to the fuse box)

With the ignition off, disconnect the fan switch relay. (adjacent to the headlight). from the 6 pin connector, earth the Red & White wire. The fan should operate. If it doesn’t the fault lies either in the fan, or the fan relay. Disconnect the two pin connector from the fan, apply 12 Volts directly to the fan. If the fan fails to spin the fan is faulty. If the fan spins, then the fault lies in the Fan relay, or, there is a solitary blue and white wire supplying the fan relay from the rear of the fuse box, this connection often snaps as it is particularly prone to corrosion, remove the fuse box and check this wire.

To prove The Fan Switch relay (Adjacent to the headlight)

With the ignition switched off, disconnect the yellow wire from the 97C switch (rear LH side of the radiator and touch it to earth. The fan should operate. Reconnect the wire.
With the ignition switched on, disconnect the black wire from the 110C switch (side of thermostat housing) and touch it to earth. The fan should operate. Reconnect the wire.
If the fan fails to operate when checking the 97C sensor (yellow wire) it must be assumed that the Fan switch relay is at fault. This can either be replaced, or checked further using the diagnostic checks given in the manual.
If the fan fails to operate when checking the 110C sensor, the fan switch relay, or the oil temperature sensor may be at fault.

To prove the Oil temperature sensor

With the ignition either on or off, disconnect the multiplug connector from the fan switch relay and using an ammeter, check for continuity between the Green & yellow wire, and earth. If there is no continuity, it must be assumed that the oil temperature sensor or the wiring to it is faulty.

To Prove the 97C & 110C sensors
It is a very clumsy process to prove that these sensors are working properly, the procedure for this is featured in all of the manuals.  But let common sense prevail. If the system is not working with the ignition on, and you have proven all of the other components by following the above checks then the 110C switch must be at fault. If the fan does not come on when the engine is very hot, and the ignition is off, then the 97C sensor must be faulty.

Nb:- Remember that if the oil temperature sensor is faulty, it will switch the reliance of the cooling system (ignition on) mode from the 110C sensor across to the 97C sensor. In this case the 97C sensor will bring the fan on if the ignition is on, wheras normally it only operates when the ignition is off.

BTW I’m pretty sure Japanese caps are measured in kg/cm². This is because although the common definition of ‘Bar’ can be used reasonably interchangeably, it isn’t actually an official SI unit and I suspect the Japanese are pretty picky on these things. It doesn’t really make any practical difference!

1.1kg/cm² = 1.07 Bar

As coolant heat rises its volume expands increasing pressure inside the system (see discussion on expansion tanks). At a certain point the valve inside the cap opens allowing coolant to pass through to the overflow tank.  As it cools the volume of coolant reduces in the system lowering pressure, so the valve opens again allowing the coolant to flow back from the overflow tank.

So you don’t potentially draw air into the system it’s important to not let the overflow tank get empty!

So yes it’s only a radiator cap, but it’s also a pretty clever dual-action valve.  And how much nicer is the original?  IMHO just goes to show how more thought went into design in those days – even for this type of functional component!

As discussed the early models had a 0.9kg/cm² (12.8psi) cap (see plenty of these on eBay) – later ones 1.1 kg/cm²  (15.6psi).  Higher pressure has two advantages:

1. it maintains the maximum coolant volume for longer.  As mentioned as coolant volume reduces so does the cooling capacity of the system.
2. it gives a higher coolant boiling point,

The boiling point of water at atmospheric pressure is 212°F (100°C). Using coolant (higher boiling point) and 15PSI of pressure the boiilng point significantly increases to  265°F (129°C).

So I would surmise that the higher pressure cap was introduced because the bike was prone to over-heating in traffic.

As it is the same cooling system components across ALL models Kawasaki obviously considered it fine to run the slightly higher pressure for the later versions, so even if I owned an early bike with the 0.9 cap as OEM I’d probably replace with the 1.1.  

As we know every little bit helps.

Note: although you can purchase 21psi (1.47 kg/cm² ) radiator caps I strongly believe that the GPz900R system IS NOT designed for this pressure and would leak profusely and  possibly damage major components like the water pump.

New original Nippon Denso caps are no longer available, here is the replacement – used on many Kawasaki’s.  A notable difference is the smaller diameter of the rubber pressure seal.  I wondered about this but it seems to work OK, and because the seal isn’t contacting the sides of the pipe it’s actually easier to install & remove.  Being a modern cheap-skate world the seal thickness is also thinner, hopefully it’s better quality.

FYI the original information is actually printed onto thin aluminium plate – so thin it looks like a sticker.  If you use a knife and carefully work it under from the edges you can gently remove the plate – I then just double-sided taped it onto the new one!

Another change (I asked the kids who studied this at school!) is that the Japanese characters have changed from hirigana to katakana. No idea why!

But is a ‘bland’ looking genuine Kawasaki (Nippon Denso) cap worth +AU$60 when you can buy an generic equivalent (maybe also Nippon Denso!) from Repco (local auto store in Oz) for maybe AU$15 bucks?

Well since they both work the same it really comes down to visuals!

Nice article from japan.webike.net discussing the role of a thermostat with images from the GPz900r.  I wish it was simple to replace  but it isn’t, you really do need to remove the tank first. 🙂

According to a 1984 article (Performance Portfolio) the thermostat begins to open at 176°F (80°C) to 183°F (84°C) and is fully open at 203°F (95°C).

As noted elsewhere the only system to allow coolant bypass when the thermosta is clsoed is a small hole

As described above the electric cooling fan system is complex but very clever, however at low-speed operation the bike may still over-heat because the  fan simply kicks in too late.

As such many owners disconnect the sensors and relays and just use a manual switch to turn the fan on or off.

The previous owners(s) had made this modification to my bike, taking the yellow earth wire from the 97°C sensor (on the radiator) and running this to a manual switch. The lead to the 110° sensor was removed and the fan never operated with the ignition OFF. As soon as temps start to climb you just switch on the fan (closing the earth circuit) and this slows the rate of heat increase to get you through slow traffic on most hot days – though not always.

But there are negatives.

Firstly you have to remember to switch the fan on – and then remember to switch it off.  On my bike the switch is placed down in the side fairing and not illuminated, so you often simply forget it is on (uses power).  Secondly the new wiring uses an inline fuse that is hidden behind the fairing.  If this power is fed from an exiting line (assume it is) then there is another fuse somewhere in the circuit, potentially making fault checking annoying.

Lastly disconnecting the sensors & relays removes any automatic protection in the event of a temperature sender or gauge failure – or me just forgetting to switch it on.

I personally think the better solution is to keep the existing OEM system but wire in a second circuit that lets you bypass the entire OEM circuit and force the fan on.  The switch should be illuminated and located in easy view (dash surround). Another electrical option could be to just keep the manual switch as is but still use the sensors to illuminate warning lights.

Jan 2021 Update : As part of a new radiator install I have rewired this switch with the following EARTHING circuits.

• The BLACK/RED wire running straight to the frame is simply for the switch frame allowing ON illumination.
• The lower BLACK/RED wire runs to a spare earth plug connected under the tank.
• The RED wire is an extension of the original lead to the 97°C radiator sensor (no longer used)
• The BLACK lead connects to the 110°C thermostat sensor

How it works.  The fan will operate if the RED lead is earthed, so manually closing the switch circuit turns the fan on. .If I forget and the thermostat coolant temp exceeds 110°C the fan will come on automatically even with the switch circuit open.  

FYI I have left the fan on for over 30min (forget) after turning the bike ignition off to help cool down, and this hasn’t drained the battery enough to stop the bike from restarting.

Radiator Misting

Well this idea has some merit.  Misting water in front of the radiator would have a massive increase in heat transfer, a technique used to increase the performance of machinery like the Subaru STI intercooler.  Generally over-heating on the bike is it’s a rare occurrence, and a system like this might just be enough to get you through the really hot days stuck in traffic.

This idea is already used to give greater cooling when towing caravans, but you would need a bike rack to mount the Rapid Cool Radiator Mist Kit!

OK, obviously this product is over-kill, but it does show that the concept & logic is sound, and there’s plenty of space on the bike to mount a small reservoir.

Booster Pump

Whilst discussing possible fans with a Davies-Craig engineer he said forget the fan, there simply isn’t going to be enough space to install anything that is going to be significantly better than the OEM one.  He recommended this product – a booster pump.

The idea is logical.  When sitting at traffic light the bike is at idle, so the water pump isn’t spinning very fast and the coolant flow is low. Adding this into the circuit means the coolant flow can be increased at the flick of a switch which could also be really useful in reducing heatsoak.  He explained that despite the faster flow (so less time to pass through the radiator) the heat transfer is much better.

The obvious question is how do you plug it in.

He suggested plugging it directly into the cooling circuit just below the radiator and didn’t think the static resistance of the pump would be significant.  When I asked about potential damage to the OEM pump he said the bypass circuit would protect it. Huh?  What bypass circuit?

Well, it was his experience that every car or bike he had seen had a coolant bypass system from the thermostat back to the pump.  Makes sense, so that if the thermostat is closed the bypass system doesn’t damage the pump.  Well of course the GPz900R (no any of its close successors) has a bypass circuit, only the little hole on top of the thermostat. 

Whether that is enough with the extra flow of a booster pump I simply do not know.  So if I were to install one I would be connecting it to a temp switch so it ONLY operated when the thermostat was open. 

Anyway I was interested enough to investigate further, note they are significantly cheaper at the store ($160) than at the site.  The issue I see is the connections are only 19mm, quite a bit smaller than the radiator hoses themselves. I don’t like the idea of that sort of restriction so my best guess is to have a second 19mm line (maybe from the radiator) that runs to the pump and then directly to the engine intake pipe.

Which means tracking down a spare waterpipe and getting it modified to allow another pipe. Nothing is ever simple!  However I might be able to use the Mishimoto Y-adapators, it looks like they are only available in the US at the moment.  My guess is even with these I would  need a one-way valve to close this line when the pump isn’t operating.

Expansion Tank

Like most 80’s vehicles the GPz900R has a coolant overflow/recovery system.  Modern automotive systems used an expansion tank, which gives more efficiency at high temperatures as there is no loss of effective coolant volume.  Now there isn’t a lot of space but you could install an expansion tank above the radiator – which is good because the pressure cap need to be higher.  The expansion tank would be plumbed to the existing radiator cap overflow tube, the return would go to a barb replacing the unused 97°C temp sensor on the radiator, and the expnsion tnk overflow would go to the OEM overflow tank under the side cover.

This is just a generic mini-expansion tank, if you were doing this you would get the largest one that could fit.  For high-end racing they are likely custom welded out of aluminium.   A quick google tells me the required size is around 6% of the cooling system, I don’t think draw-down is really an issue here.  There’s plenty of negatives like more clamps, hoses (leaks) and you would have to remove the tank to get to it.

Feb 2021 Update : What I did was cut off two legs of the larger RX fan mounting frame and then manufacture new mounts so that it fits onto the smaller GPz900R radiator,

I made sure I checked it cleared the exhaust headers FIRST!

The fan plugs directly into the GPz loom but does have an additional earth lead (not shown in Haynes wiring loom BTW) that I simply attached to the frame.  The volume of air that the fan moves is significantly more than OEM but as mentioned elsewhere there is a risk of the fan blades melting due to the proximity to headers (single-skin after market) and lack of a shroud pushing air downwards.

Detailed Discussion

GPz900R radiator

GPZ1000RX Radiator – same width but 21 rows vs 15 and much larger fan.

Here are some interesting comparisons with the radiators from the GPz900R and the GPZ1000RX – in Australia released as the ZX1000A.  As you can see it is a much, much larger 21-row (vs 15 row) unit with a matching larger fan.

The increase in radiator size is much more than the increase in engine capacity, so despite the company dismissing early concerns I’d suggest that Kawasaki DID take owner concerns & feedback on the GPz900R’s cooling system into consideration when designing this!

I had read *somewhere* that you can fit this radiator into a GPz900R so picked up one from a wrecker in Melbourne, but when it arrived was reminded why you don’t believe everything you read on the web!   The front supporting frame does not allow for the higher RX radiator without modification.  Modifying the frame to would mean the radiator still hits the header pipes so I’m not convinced this is a solution.

But what is interesting is the larger fan.

This fan is significantly larger, which means significantly greater area as well as blade tip speed, so potentially this fan moves a lot more air across the radiator.  I’ve shown it here positioned on the front of the radiator but I’ve also confirmed it will fit in the operating (pulls air) location behind the radiator and above the header pipes.  This might be all I need to sort city traffic issues.  I have a new aluminum radiator on order from China, so when this arrives I will try and work out how to get it all to fit as the fan frame mounting points are different.

The images below also show the heavily shrouded 7-blade GPz900r fan.  The 6-blade is an aluminum fan – which was a previously available aftermarket (Muzzy) option, but I haven’t been able to source a current supplier.  Theoretically less movement in the AL blades means better efficiency, but of course it also means more damage to the radiator if they accidently meet!  My radiator has light scuff markets in one quadrant where the blade tip has brushed against the core so that’s definitely a possibility.

Fan Shroud

Well the consensus is that the heat from the headers, especially if single-skin after market (OEM double-skin) will destroy the fan blade, and I have seen a photo of a damaged one (without shroud).  So the shroud is acting as an airflow deflector pushing the air downwards and across the header pipes, hence dispersing the heat affecting the blades.

Obviously this will only work if the bike is moving or the fan operating. What is also a bit confusing is that the fan from the RX1000 doesn’t have a shroud, but the blade material looks very similar so I’m going to take this into consideration.

Aftermarket Fans

With everything apart I know I can fit a 8″ RX fan and there are modern aftermarket 8″ fans available with a decent perimeter cowlings.  the question is are they better.  The three important factors are CFM, current draw and heat.

7.5″ Plastic	8″ Plastic	8″ Plastic	8″ Plastic	10″ Chrome
Image: schnitzracing.com		Image: daviescraig.com.au	Image: maradynehp.com	Image: enginemaster.com.au
1100CFM@3A*		400CFM@5A	450CFM@8.1A	1150CFM@na
2.5″ (64mm)		56mm	2.48″ (63mm)	na

*DESPITE THE QUOTED CFM I have to take this fan’s performance figures with a grain of salt, It is not expensive nor specialised race tech so it seems improbable that it has more than TWICE the airflow of comparable 8″ fan designs with LESS current draw.  There are other brands like Perma-Cool that also claim huge CFM numbers,  I suspect that this is a ‘questionable’ marketing ploy where the quoted figures are with the fan installed in a full shroud – which makes a huge difference to the draw of a ‘puller’ fan as per this example from www.spalusa.com. Unfortunately SPAL don’t offer smaller fans suitable for the bike.

The cowl on the Davies Craig is glass-filled (GF) polypropylene (max ~130°C) and the fan GF polyamide (max ~130-150°C). The Maradyne is all GF polyamide, I’ve emailed both companies directly to ask if they think these temps are sufficient – several months later no reply from either.  IMHO the temp rating for use here seems low, but confusingly just from pure observation it also seems highly likely that GF nylon (brand name for polyamides!) would be more heat resistant than the OEM blades…..

Whilst a brushless motor would be peachy, the smallest size I could find for these motors was 11″. I have found some metal fans but again the smallest of these was 10″, too big for the GPz900r. Shown to highlight the significant difference in airflow if you can increase the blade diameter.

Summary

After consideration (not helped by useless customer support….)  I’m thinking that the non-rotaing shround sitting right next to the headers needs to be metal as per OEM – the heat is just going to be too high here for any commercial thermoplastic. Unfortunately I have only found a 10″ chrome fan which simply won’t fit.

Feb 2021 Review:  firstly this seller is no longer operating on eBay – I guess no chance of a refund then!  But I looks *exactly* the same product being sold by many other online sellers.  I had no issue with the delivery time nor really the quality of the product, but it is a reminder you need to check your online items within 30 days (I think?) to ensure your refund can get processed by PayPal or eBay.

• Radiator installed (with fitment issues) and is holding pressure fine.  However despite more rows (19 v 15) and claiming increased capacity I do not believe this radiator is quite as effective as OEM, so installing JUST this part might actually make your over-heating issues worse.
• In my case  IN COMBINATION with the larger RX1000 fan the cooling system is working excellently even under the worst-case scenario – traffic-jams on a +38°C day.
• Note that filling the radiator was quite different to OEM, when on the side-stand as you get towards the top there is an air block and it simply doesn’t fill.  You either need to either this on the centre-stand or go through the process of ‘top-up + straighten’ the bike, I think it was around 6-7 times..

You know it’s full when the bike is straightened without gurgling! Here is my notes on how it all works on the road that were originally posted on Kawasaki GPz900R Owners

Detailed Discussion

I  bought an aftermarket aluminium radiator off eBay for the bargain price of AU$191.  Of course they are now selling for AU$138……

Yes they are from China, but you don’t have a lot of options.  Below is the ONLY Apr 2021 replacement radiator available on japan.webike.net so at ~AU$900 plus shipping so there is such a massive difference in price you really can’t compare them.  Bit like comparing the GPZ900R to a modern motorcycle.  If you need a replacement I would expect this is because your OEM is cactus so you probably have no choice but to pick Chinese or Japanese.

It arrived very well packaged and externally the quality looked pretty good.  Welding would be as I would expect from a local shop, the only obvious flaw is the cheap thread inserts, one of these wasn’t crimped properly and is already loose.

Fitment

The most obvious issue is that it doesn’t quite fit properly.  This is mainly due to the lower mounting lug (top in this photo) being slightly off position (~8mm) and a design that makes the whole radiator sit around 6mm higher.  Not a lot but this means the radiator didn’t fit without trimming the upper fairing (update: not required see below).  But more annoyingly it meant that the centres (ie the bolts) of the top mounting holes don’t match the frame.  Grrr, however a three-pointed fit the GPz900R wants this to be extremely precise, unfortunately the Chinese reverse-engineering simply isn’t good enough to match.

I modified the lower lug and this let me get close enough to butcher up a fit but it’s still annoying, especially as I now DIDN’T have to trim the top fairing.  Sure the trimmed part is hidden by mid-faring but still……..grr.

Boss lifts radiator too high.  Would help if boss sat flush with the U-channel.  I used my wood vice to gently hold radiator only on the sides.

Angle grinder, hacksaw and filing. Material is quite soft so it’s pretty easy.  A balance between removing material and leaving enough weld strength – I think I got 4-5mm.

Tightening the loose thread insert by bashing with a hammer  Rough as guts….

Before and after.  Note it’s *really* close but still no banana, I assuming this is becuase the centreline of the boss itself is out by around 8mm and I can’t adjust this easily.  I have used the OEM bolt for the top LH mount and just used two quality cable-ties for the top RH mount, Rough but un-seen and more than sufficient holding strength.

You could use a smaller diameter nut & bolt that fits through the welded OEM nut (on the frame) but I’d argue the cable ties are just as strong.  Radiator still uses OEM rubber isolation washers on both sides.

A perfect fit – well looks can be deceiving!

Summary

Although some have claimed they are rubbish, the feedback on these Chinese radiators from other owners is definitely more good than bad .Curiously other owners say the fit is perfect except for the radiator being thicker and required some minor trimming of the center fairing to fit. Maybe I just got a slightly bad one, or *maybe* there are different manufacturers & suppliers.

At the end of the day for the price I paid en if it needs modifications to fit still I’m totally satisfied with what arrived.

1l > 20l = $1.34/l	500ml > 10l = $3.40/l	1.89l = $18/l
2-3 years life	na	Replace yearly

I’m loathe to recommend particular products however I have just started to use this Penrite product after years of using quality glycol (min 33% concentration) based green coolant. For me the most relevant factor is that this red product does not have the slipperiness of glycol products – it feels pretty well the same consistency & viscosity as the base distilled water.  The idea here is that lower viscosity coolant is not only more efficient (it’s still a Type B anti-boil product), but in the event of a cooling system failure then a wet track or tyre is still preferable to a *really* slippery one.

Chris R. races his 168hp@rw and recommends the Motul equivalent for the same reason.

The Penrite I can buy from my local auto store, the Motul is probably available from many motorcycle dealers (I’ll check next time) and Engine Ice mostly available here in Australia online – although your local sote may have it. (price based on a free shipping site).  The Apr 2021 prices are from Australian suppliers  and *only* give a value-4-money comparison – I simply haven’t used the others.

I’ve included Engine Ice because it is regularly recommend by many owners to help with overheating and it’s popular, one online store claims it is their 3rd highest selling product.  Given it is significantly more expensive I would hope it works!  My other modifications solved my over-heating issue, however if I was still ‘getting hot under the collar’ then I would probably accept the high purchase price and annual replacement cost as an option.