Oil explanation for Guzzi (&BMW) owners: 75W90 is not an alternative for 85W90!

[Bazil]

The manual for my 1961 Royal Enfield says to use engine oil "but on no account should heavy yellow grease be used in the gearbox".

It wouldn't have even occurred to me to use grease in a  gearbox ! Was this a thing for older low powered machines?

[berniebee]

Hiya. My replies below in purple in good faith (Oh no: oil thread descends into typical oil thread thread... play nicely children lol.)

Good to have the brain cells stretched, cheers 😊

 I finally looked it up and the empirical formula for liquid viscosity is indeed a two parameter exponential. I admit some laziness here, as I saw the curve and knowing that oil will freeze into a solid given low enough temperatures (Ie: infinite viscosity) I assumed it was asymptotic - a hyperbola. The point is that it's a non linear curve. ( I will not go into arguing that an exponential decline curve is a simply a special case of a hyperbolic curve.  )

As for engine temperature being linearly dependant on ambient temperature, yes, it is somewhat more complicated than that. For example, a water cooled engine with a thermostat and switchable cooling fan will not necessarily get as hot as ambient plus a fixed number.

But for something like a rear drive (Or an air cooled engine) , I believe it's essentially correct that it's operating temperature will be ambient plus X. Here's why:
I come from an electronics background where device junction temperature is calculated as : Tj = Ta + Pd x Rja
 (Pretend the lower case letters are subscripts.)
Tj is junction (ie; device) temperature
Ta is ambient temperature
Pd is power dissipated
Rja is thermal resistance

The formula tells us that the device temperature is equal to the amount of power (times a factor) PLUS the ambient temperature. In other words,  with everything else being equal, the device temperature is ambient temperature plus X. My reasoning is that it matters not whether the heat is generated electrically or mechanically nor whether we are talking about a chunk of silicon vs a chunk of steel.

The rear drive is going to see huge variations in operating temperatures (And therefore viscosity)  with different climates and for that matter different riding styles. I agree that putting sewing machine oil or grease in the rear drive is a bad idea. But otherwise, put something labelled as gear oil in there and be done with it.

There is a far more important factor for rear drive longevity.  EP oils have support additives that get depleted over time. That can lead to gear pitting and worse. To extend the life of your rear drive, do regular oil changes. If you live in a hot climate or ride like (Insert favorite Moto GP racer here) , replace it more often!

BTW, some may be surprised to read that SAE 90 gear oil has the same viscosity as SAE  50 motor oil. It's all about the additives.





   

[Old Jock]

I will not go into arguing that an exponential decline curve is a simply a special case of a hyperbolic curve. 

I thought that too bernie, but not clever enough to go into the mathematics that have long since been erased from my aging brain

I did some Googling around the relationship and the easiest of the formulas I found was the Walter Equation, it uses base 10 not e and is really limited to Newtonian fluids, whereas most multigrades are non Newtonian, but I reckon might give a decent enough interpolation as just a rough guide

Formula is

log log(V+0.7)=b1+b2*logT

There is a paper I stumbled across that's pretty good on this, it's a pdf download so if you click on the link it will download it (it also give formulas for the constants and how to use the formula)

https://www.doria.fi/bitstream/handle/10024/177814/rautelin_wolter.pdf?sequence=2&isAllowed=y

Going to make up a program for my calculator (remember them?) so as I can estimate some viscosities for different grades as I'm going to change the grade on my HiCam so it would be nice to calculate some predictive percentages

[berniebee]

I thought that too bernie, but not clever enough to go into the mathematics that have long since been erased from my aging brain

 I remembered that factoid about trig hyperbolic curves being able to be expressed as exponential equations, but the reason I didn't want to go into it is.... because it's been far too long for me too! 
Sigh, all those advanced calculus, linear equations and probabilities classes and last week I had to Google something to help tutor a high schooler.  I even had to look up the relatively simple thermal equation for device junction temperature, just to make sure.  And I used that one frequently at times in my career. 

Yep, I'm getting old. Calculators?  Why, I was in the last year of class at my school that was forced to buy and learn how to use a slide rule, though by then every student in the class had a calculator. We all whined about it while the teacher sheepishly defended the outdated school policy.  From then on it was a class wide arms race to buy the newest, greatest Texas Instruments calculator ("It's programmable? Ooooh!").  Or an HP calculator, if you were of the weirdo Reverse Polish Notation persuasion.   Remember RPN? My youngest son (a computer science grad) tells me that RPN is still used in some programming languages and he says it's more efficient. I'd threaten him with expulsion from my will if he mentions RPN again, but he's already earning far more that I ever did. I still have an almost new slide rule somewhere in the bottom of my desk, because you know, if the entire world is destroyed by nuclear hellfire and overrun with zombies, but I need to calculate an efficient trajectory to Mars...

Thanks, I'll check out that formula - but I'll still buy whatever gear oil is on sale! 

[Old Jock]

Yup RPN fanboy here (sorry) HPs were always my favourate but expensive here in the UK.

Used a sliderule in school too and then later went through various calculators at college, the ultimate being the HP41C. I've gone back to using an HP48GX that I had, I've just found there's a whole subculture out there extolling, using, collecting and modifying/repairing calculators. Google up HP41C prices, amazing!!

Suppose it's like owning that LM I that I lusted for as a boy but could never afford. I still don't have one but love the Magni based on the LM 1000 and it's pretty dammed close.

Just old men with too much money and time on their hands 

As far as the oils go, I agree with you, the HiCam is a bit of a special case and its engine oil grade I'm considering changing in a desperate attempt to keep oil pressure up as the dammed thing runs super hot

Anyway enough "Off Topic" 

[Don G]

Royal Enfield gear box's had steel gears running on steel shafts, no bronze bushings at all. What Enfield did was pack the gear box's recesses with heavy yellow grease and then added oil until it was full, at least all of the ones I owned were like that.  DonG

[Philnewbike]

I finally looked it up and the empirical formula for liquid viscosity is indeed a two parameter exponential. I admit some laziness here, as I saw the curve and knowing that oil will freeze into a solid given low enough temperatures (Ie: infinite viscosity) I assumed it was asymptotic - a hyperbola. The point is that it's a non linear curve. ( I will not go into arguing that an exponential decline curve is a simply a special case of a hyperbolic curve.  )

As for engine temperature being linearly dependant on ambient temperature, yes, it is somewhat more complicated than that. For example, a water cooled engine with a thermostat and switchable cooling fan will not necessarily get as hot as ambient plus a fixed number.

But for something like a rear drive (Or an air cooled engine) , I believe it's essentially correct that it's operating temperature will be ambient plus X. Here's why:
I come from an electronics background where device junction temperature is calculated as : Tj = Ta + Pd x Rja
 (Pretend the lower case letters are subscripts.)
Tj is junction (ie; device) temperature
Ta is ambient temperature
Pd is power dissipated
Rja is thermal resistance

The formula tells us that the device temperature is equal to the amount of power (times a factor) PLUS the ambient temperature. In other words,  with everything else being equal, the device temperature is ambient temperature plus X. My reasoning is that it matters not whether the heat is generated electrically or mechanically nor whether we are talking about a chunk of silicon vs a chunk of steel.

The rear drive is going to see huge variations in operating temperatures (And therefore viscosity)  with different climates and for that matter different riding styles. I agree that putting sewing machine oil or grease in the rear drive is a bad idea. But otherwise, put something labelled as gear oil in there and be done with it.

There is a far more important factor for rear drive longevity.  EP oils have support additives that get depleted over time. That can lead to gear pitting and worse. To extend the life of your rear drive, do regular oil changes. If you live in a hot climate or ride like (Insert favorite Moto GP racer here) , replace it more often!

BTW, some may be surprised to read that SAE 90 gear oil has the same viscosity as SAE  50 motor oil. It's all about the additives.





 

Hi. Very interesting. I agree with you. I've since had to remind myself about forced convective cooling and it is highly dependent on ambient temp. In summary any engine (air/oil/water cooled) that has a way to control it's running temperature by limiting the max cooling available via a thermostat will have an oil operating temperature relatively independent of ambient. And as you say, for one that doesn't (and for the final drive) the oil operating temperature is going to be very dependent on ambient temperature (I never ride in anything other than 15-25 degrees so might be hard for me to detect!). Would be interesting to see what the fellas with temp sensors find... So maybe final drive is one case where very cold/hot running riders may want to adjust their choice...

Gearbox (and engine) still slightly different case though:
I am under the (incorrect?) assumption that because the oil operating temperature in Guzzis is, lets say, ~80-90 deg C (?) and the heads are at ~110 C then because this is quite high a change in ambient temperature is only ever likely to be 10 to 20% ballpark, meaning the change is relatively small (i.e. 40 C shaft + 20 C change in ambient is 50% increase whereas 100 C engine + 20 C change in ambient is 20% change). And therefore because the gearbox is directly heated by the engine this too will have a steadier temp like the engine unlike the shaft drive. Hence I am going to stick to me guns and say 90 weight (as the Getrag people stated in the OP) is best for most conditions in gearbox unless you're planning to cross Antarctica or something. [edit: or possibly 75w140 looking at the graph at the end?]

Very good discussion, cheers all

[kingoffleece]

BEST THREAD EVER.  Well done, gents, well done indeed.

[Philnewbike]

Thanks.

Here's the chart I plotted based on exponential curve (log-linear) from two data points (not every one is plotted). Obviously not ideal but the data sheets generally only give 40 and 100 C. Guzzi spec of 80w90 and 85w90 is dashed line

(NB I have tried this before in another thread and I think got the numbers a bit muddled. this one should be ok)



buy flip coin

[berniebee]

Yup RPN fanboy here (sorry) HPs were always my favourate but expensive here in the UK.

You do realize I can never speak to you again. 

[berniebee]

Gearbox (and engine) still slightly different case though:

Agreed, the gearbox and transmission are a different and more complicated case.   The info you posted at the start and the oil viscosity curves are much appreciated and obviously, a great thread starter!

[SmithSwede]

This is an interesting topic, even though I suspect it ultimately makes no practical difference. 

I’m assuming there is some kind of engineer’s formula that is used to spec things like gear oil grade and viscosity.   You plug in important parameters like the size of gears, type of gear, type of bearings, rotational speed, load, expected heat, etc.   Then the formula gives you a recommended range of outputs. 

Does anyone know that formula?  Tried to apply it to a Guzzi?

My hunch is the output of that standard formula probably translates to “75 to 90 weight is fine.   80 is optimal absent special circumstances.”

[SmithSwede]

Forgot to add.  I’m under the impression that gear oil differs from engine oil in that gear oil does not have viscosity improvers.    So a motor oil may act like a 10W when cold, but the VI makes it act like a 40 weight when hot. 

In contrast, gear oil has to fall within the viscosity specs “naturally.”   

I think that is because VI additives are easily sheared, and gear boxes are shear machines. 

[Mr Revhead]

Shall we get into the importance of shear strength? Viscosity is only part of it!

[Old Jock]

Forgot to add.  I’m under the impression that gear oil differs from engine oil in that gear oil does not have viscosity improvers.    So a motor oil may act like a 10W when cold, but the VI makes it act like a 40 weight when hot. 

In contrast, gear oil has to fall within the viscosity specs “naturally.”   

I think that is because VI additives are easily sheared, and gear boxes are shear machines.

That's interesting Smithswede, I always laboured under the (mis?)apprehension that it was viscosity improvers that gave you multigrades in the first place, however I've never been done any research into them in any detail as I'm still trying to get my head around engine oil.

Shall we get into the importance of shear strength? Viscosity is only part of it!

Would that topic not be better discussed in a separate thread, this one is complex enough for my dimm brain without inserting yet another variable?

I'm going to do some comparisons of the Walther (I miss-spelled it earlier) equation vs an exponential function just to see how they compare.

I'm thinking about engine oil, as it'd be interesting to be able to have a comparison (albeit rough) of oils of different grades at different temperatures. That way I might be able to get some sort of handle on the effect changing an oil grade might make on the pressure at running temps on my toasty HiCam, all other things being equal.

Another science experiment......... ...................

You do realize I can never speak to you again. 

 

John

[Old Jock]

Question to Philnewbike

Did you use an expression like this to draw your graphs?

V=e^(b1+b2T)

Where
V is dynamic viscosity (mm^2/s or Cst)
T is temperature (K)
b1 & b2 are constants
e is Euler's constant (aka Napier's constant or base of natural logarithm)

Just wondering as I wrote a short program for my calculator for both the above and the Walther formula I posted earlier so as I can compare results

[SIR REAL ED]

It wouldn't have even occurred to me to use grease in a  gearbox ! Was this a thing for older low powered machines?

A not unheard of practice in the olden days.  I know a couple old timer types who had gearbox problems in vehicles using gear oil and solved them (so they claim) by switching to grease.

I would not recommend however switching to genuine Moto Guzzi Invisiable Bearing Grease (tm).  "Forgive me Father for I have sinned.  It been many years since my last confession....."

[SIR REAL ED]

This is an interesting topic, even though I suspect it ultimately makes no practical difference. 

As a mechanical engineer, and former mechanic on all sorts of old junk...... I'd be very comfortable betting the farm that you are right SmithSwede.

The two concerns that are probably much more a factor than actual gear face wear are 1.  Noise.  and 2.  Specifying a viscosity that makes the differential work properly (N/A for Guzzi's).

Until someone wants to perform two ride till gear failure, or ride 100k miles on two identical bikes with different oils and measure wear, it's all just fun campfire banter.

Also it is fun to spoil our babies with special oil treats.

Don't run it dry, don't fill it with cheap vegetable oil from Piggly Wiggly, don't use genuine Moto Guzzi Invisiable Bearing Grease (tm), and enjoy the ride.

[Philnewbike]

This is an interesting topic, even though I suspect it ultimately makes no practical difference.  I disagree lol 

I’m assuming there is some kind of engineer’s formula that is used to spec things like gear oil grade and viscosity.   You plug in important parameters like the size of gears, type of gear, type of bearings, rotational speed, load, expected heat, etc.   Then the formula gives you a recommended range of outputs.  Yes there are various charts used that deal with gear surface finish, clearance, load, speed, etc

Does anyone know that formula?  Tried to apply it to a Guzzi? Impossible without engineering drawings

My hunch is the output of that standard formula probably translates to “75 to 90 weight is fine.   80 is optimal absent special circumstances.”  As Margate Thatcher would say: no, no, no ! Please see my first post as to why 

[Philnewbike]

Forgot to add.  I’m under the impression that gear oil differs from engine oil in that gear oil does not have viscosity improvers.    So a motor oil may act like a 10W when cold, but the VI makes it act like a 40 weight when hot. 

In contrast, gear oil has to fall within the viscosity specs “naturally.”   

I think that is because VI additives are easily sheared, and gear boxes are shear machines.

Yes VI modifiers shear over time and the oil tends towards it's W weight. Wide viscosity gear oils do have VI improvers, but some others of high quality synthetic type naturally fulfil some W ratings, e.g. 85W90

[Philnewbike]

Question to Philnewbike

Did you use an expression like this to draw your graphs?

V=e^(b1+b2T)

Where
V is dynamic viscosity (mm^2/s or Cst)
T is temperature (K)
b1 & b2 are constants
e is Euler's constant (aka Napier's constant or base of natural logarithm)

Just wondering as I wrote a short program for my calculator for both the above and the Walther formula I posted earlier so as I can compare results

No nothing that complicated. I had two data points and know that the function of viscosity they follow is fairly exponential. I therefore plotted the variable as a log base 10 so I could read off in-between readings. Your way sounds better

[Philnewbike]

As a mechanical engineer, and former mechanic on all sorts of old junk...... I'd be very comfortable betting the farm that you are right SmithSwede.

The two concerns that are probably much more a factor than actual gear face wear are 1.  Noise.  and 2.  Specifying a viscosity that makes the differential work properly (N/A for Guzzi's).

Until someone wants to perform two ride till gear failure, or ride 100k miles on two identical bikes with different oils and measure wear, it's all just fun campfire banter.  Yes and no. It is objectively fact that using 75W90 instead of 85W90 or 90 will mean you're running on an oil much thinner than the oil engineers at Agip (and BMW/Getrag in my example) recommended. As for using 75W140 etc expensive stuff then yes, more subjective

Also it is fun to spoil our babies with special oil treats.

Don't run it dry, don't fill it with cheap vegetable oil from Piggly Wiggly, don't use genuine Moto Guzzi Invisiable Bearing Grease (tm), and enjoy the ride.

[Old Jock]

No nothing that complicated. I had two data points and know that the function of viscosity they follow is fairly exponential. I therefore plotted the variable as a log base 10 so I could read off in-between readings. Your way sounds better

Now I'm just confused, although that doesn't take much.

Are you saying you used log10^x and not e^x to plot the viscosity, I thought earlier you stated that you used the exponential function, which I understand as "e", if I'm labouring under a misapprehension please tell me

I don't know if it's a better solution or not given Walther's equation uses log10, albeit twice.

Pretty easy to do as a little bit of high school algebra will give you 2 equations with 2 unknowns.

So
b1=lnV1-T1*(lnV2-lnV1)/(T2-T1)
b2=(lnV2-lnV1)/(T2-T1)
Where
T (K)
V (mm^2/s or CSt)

I reckon overall Walther's equation will probably be the more accurate of the 2 methods (see my earlier post), but even then it will be ballpark (at best).

I'm a bit off topic here as what I'm looking at is a method to calulate engine oil viscosity at different temps

What I'm trying to do is calculate what the change in viscosity will be for a given grade and manufacturer at a specific temp, percentage and absolute.

Using that I can determine the equivalent temp of the new calculated viscosity would be with my existing oil.

I already know what the pressures will be with my existing oil at various temps so I can determine if changing the oil weight would make any difference and what that the pressure difference might be

As an example I currently run 15/50 and I know that when the sump temp gets above 110C the pressure starts to nose dive and by 120C the pressure is around or a little below 10psi at idle (oil light on and confirmed with a gauge). So if I go to a 10/60 for example I could get an idea if at 120C the additional viscosity would be as high as my exisitng 15/50 at 110C (I'm just throwing out numbers here).

I know the 10 part is irrelevant but there's another reason that a thinner grade would help when cold, which I can also now roughly calculate

However theoretical and experimental are 2 quite different animals

[Philnewbike]

Now I'm just confused, although that doesn't take much.

Disclaimer: I an an engineer but not an oil or gearbox expert (but I know a couple and have a read a lot), so though I do think I know more than the average oil threader (the reason I posted) I am by no means 'gospel' !

Are you saying you used log10^x and not e^x to plot the viscosity, I thought earlier you stated that you used the exponential function, which I understand as "e", if I'm labouring under a misapprehension please tell me

Yes. See here:https://www.engineeringtoolbox.com/dynamic-viscosity-motor-oils-d_1759.html
The plot of viscosity is approximately a straight line when done on a log10-linear graph. OK not perfect but not far off. So, given I have 2 data points (visc at 40 and 100 C is always given in data sheets) and the knowledge that log10-linear is approx straight, I fitted a line of best fit through those data points using a log10 function in Excel. I then read off the in-between numbers at eg 65 C (gearbox temp)

I don't know if it's a better solution or not given Walther's equation uses log10, albeit twice.

Pretty easy to do as a little bit of high school algebra will give you 2 equations with 2 unknowns.

So
b1=lnV1-T1*(lnV2-lnV1)/(T2-T1)
b2=(lnV2-lnV1)/(T2-T1)
Where
T (K)
V (mm^2/s or CSt)

I'm not familiar with the Walther's equation sorry

I reckon overall Walther's equation will probably be the more accurate of the 2 methods (see my earlier post), but even then it will be ballpark (at best).

I'm a bit off topic here as what I'm looking at is a method to calulate engine oil viscosity at different temps

What I'm trying to do is calculate what the change in viscosity will be for a given grade and manufacturer at a specific temp, percentage and absolute.

Good idea. I think the main thing missing from my estimate is the viscosity index (VI) variable - i.e. each oil will have it's own VI number and so will actually need a slightly different line of best fit. This goes back to what the chap above, Smithswede, was saying about gear oils and VI polymers. Actually some gear oils of wide range do use VI polymers (like 10W60 engine oil does) but in the gearbox are in a environment perfect for chopping up polymer chains and thinning the oil towards its 'W' base viscosity. E.g., if you use an 75W140 then this (according to my rough plotting) will be OK as the viscosity at the operating temp is approx same as 85W90. But, after a period of time the polymers might be chopped up and you slowly end up with a 75 weight. It is hard to say which gear oils have VI improvers and which don;t though...

Using that I can determine the equivalent temp of the new calculated viscosity would be with my existing oil.

I already know what the pressures will be with my existing oil at various temps so I can determine if changing the oil weight would make any difference and what that the pressure difference might be

As an example I currently run 15/50 and I know that when the sump temp gets above 110C the pressure starts to nose dive and by 120C the pressure is around or a little below 10psi at idle (oil light on and confirmed with a gauge). So if I go to a 10/60 for example I could get an idea if at 120C the additional viscosity would be as high as my exisitng 15/50 at 110C (I'm just throwing out numbers here).

I'd suggest 20W60

I know the 10 part is irrelevant but there's another reason that a thinner grade would help when cold, which I can also now roughly calculate

However theoretical and experimental are 2 quite different animals

[Philnewbike]

Just to throw another cat amongst the pigeons: The original CARC Agip gearbox spec was for 85W90 and in the data sheet it says it's suitable for LS differentials. Now was that done on purpose or was it just an additional but unneeded spec of that particular oil they wanted to sell?

I am wondering if the engagement of dog gears in the gearbox are actually kept better engaged under load if the oil is not as slippery... I am guessing now!  You may or may not know but oil can have high film protection AND high or low slipperiness depending on it's design. (E.g. gear oil with high slipperiness is bad for synchromesh boxes because the syncro rings then don't work as designed - they need friction.)

If your gearbox has any false neutrals or jumps out of gear then maybe (??) 1) a return to spec and 2) use a gear oil that says it's suitable for LS differentials

Food for thought !

[Ncdan]

Lots of very good technical information here in regards to gear oils. I changed the gear weight in my Stone when I first got it because I thought a had a minor seepage, to a heavier weight oil.
A question I have is has anyone actually experienced a rear end or transmission failure that the cause can actually be traced back to the oil that’s was in it at the time. The lack of oil is not a factor in the question.

[Old Jock]

Thank you for the reply Phil please don't think I'm picking at your method I'm not, I'm just trying to pick up as much data as I can.

I was an egineer before I retired although competent on a good day was about my best effort.

If you are interested go to the 3rd post down on this page from the top where I listed Walther's Equation, but more to the point is a link to what (IMHO) is a pretty good paper on testing and gives a breakdown of the various calculations that can be used to interpolate (or extrapolate) oil viscosity at given temp

Many thanks for your efforts and your replies it's appreciated

John

[krglorioso]

     EP additives and Brass synchros don't play well with each other.

       Paul B

I recently corresponded with RedLine oil and was told that their gear oils are completely safe to use in gearboxes that have brass or copper components.

Ralph

[slowmover]

What a heavy duty oil thread. Is this stuff ok?

[SmithSwede]

Two other comments.   First, I find it amusing that we are supposed to take as received wisdom the comments of some engineers at Getrag.   But if you read the cited article, they are being interviewed in the context of explaining why BMW gearboxes suck today, and have always sucked.  Phrases spring up like “who is guilty for creating this miserable transmission?”  Why am I supposed to treat their recommendation of 90 weight as gospel?

Second, what are we to make of Guzzi specifying in their owner’s manual the use of 85W-140 oil in the gear box (and final drive) for the 1987 V75?  I bet that’s basically the same gear box as on any splash drive small block. 

Again, I think we are probably reading way to much into a spec in the manual.   But I could be wrong about that and am happy to learn.