Come on; you know that's not what he or anyone else is saying.
Redrives for most commercial engines.
- Thread starter John Penry-Evans
- Start date
Help Support Homebuilt Aircraft & Kit Plane Forum:
challenger_II
Well-Known Member
I am curious as to why no one has attempted to use a Chrysler Fluid Drive as a coupling between the engine and reduction drive.
Dillpickle
Well-Known Member
I belive The Infamous Molt Taylor did that on the Imp. It was not without problems. Haven't thought about that for decades...was it the Dodge Flexidyne?
Are you thinking of the Dodge Flexidyne?
Dillpickle
Well-Known Member
Spare us the sarcasm . I'm pretty sure some readers appreciate the knowledge.
As noted previously, the engineered selection of a torsional soft element can work very well.
He is just trolling us.
challenger_II
Well-Known Member
Actually, no I am not. I am introducing an alternative idea into the discussion.
after reading this and other threads on TV,a guibo
is a device that transmits mechanical rotational
energy and if it was joining a dc electric motor and
a fan or pump,would do nothing,in the case of piston engine and gear reduction driving a fan,the soft rubber elements,compress and decompress with each pulse of power from each combustion event,
the durometer or spring rate of the soft elements
determines where the power pulses are placed in time,always delayed and somewhat spread out,
with the plan bieng to move the resonant frequency
out of the operational range
given that there is a good selection of interchangable
soft elements for guibos,and realy awsome free frequency apps for your phone ,its not that far a reach to just go for it,little math and a sweet build will help
is a device that transmits mechanical rotational
energy and if it was joining a dc electric motor and
a fan or pump,would do nothing,in the case of piston engine and gear reduction driving a fan,the soft rubber elements,compress and decompress with each pulse of power from each combustion event,
the durometer or spring rate of the soft elements
determines where the power pulses are placed in time,always delayed and somewhat spread out,
with the plan bieng to move the resonant frequency
out of the operational range
given that there is a good selection of interchangable
soft elements for guibos,and realy awsome free frequency apps for your phone ,its not that far a reach to just go for it,little math and a sweet build will help
PMD
Well-Known Member
The thing on your coil spring automobile suspension that we call a "shock absorber" is actually a damper - as you are aware that the coil spring returns a significant amount of its energy without a lot of damping. However, if we go into the trailer world, you will find a HUGE number of rubber suspended trailers (usually "torsion" axles where a square shaft squishes four round rubber rods lodged in the corner of a square tube, 45 degrees out of phase). Those ubiquitous axles do NOT use any "shock absorbers" (actually called dampers in most chassis engineering circles) and they do not bounce. Their rubber is indeed a very effective damper.
That said: there are rubber and rubber-like formulations of very low through very high hysteresis but the problem is when you walk into your rubber store (as I frequently do) and ask for specs, you will pretty much get dimensions and durometer values. When you ask for resilience and hysteresis values you will likely get a "huh?????" response. Yeah, the formulator will likely have that, just takes a lot of digging to get to it. If you are designing and series producing guibos, you are probably having the exact properties molded to the exact net dimensions to get what you want. Where things get a bit dicey is when making one-off couplings where it is only practical to work with materials that are readily available. That is when one plays with the radius, dimensions, number of drive interfaces, cooling, etc. to get something one can live with (and their drive system can literally live with). It is often as dog proposed just as easy to build a part and test, varying material if you miss-guess to change resilience/hysteresis to suit.
Oh: BTW - I have soup sandwiches quite regularly (grilled cheese floating in cream of tomato soup).
Last edited:
Giubo, giubo, giubo. Never guibo.
Just saying.
Just saying.
No one said rubber had no damping. I even posted a Centaflex data table with the damping values. Doesn't matter. If you run the typical engine-propeller system in resonance for very long, the byproduct of that damping (heat) will destroy the coupler. It's pointless.
Frequency for those axles is what, 1 hz? They don't heat much.
WonderousMountain
Well-Known Member
The app looks okay to me.
Continuing the theme...all "giubos" are not the same...
Take a look at the three examples pictured below. One is a OEM automotive drivetrain part. The other two are primarily intended to tailor torsional behavior. What is the significant difference?
Take a look at the three examples pictured below. One is a OEM automotive drivetrain part. The other two are primarily intended to tailor torsional behavior. What is the significant difference?
Attachments
PMD
Well-Known Member
Yes, caught your table, just pointing out the blanket statement you made about rubber not being a damper. Yes, the axles have a low frequency by massive amplitude/deformation as each 6" or so lever can be operating with a base load of 2,000 lb/ft. The DO have more than adequate cooling as the axle housings are usually in a clear airstream.
As it happens, I am just designing an element to replace a factory guibo in a diesel air compressor. The screw is driving by a gear train and the coupling is burried inside of a bellhousing - so cooling is a really big issue. The original ones (regular failure item) are a very high hardness rubber (not sure what compound) bonded onto a central drive "gear" (literally) but the high temps mean the rubber becomes brittle and breaks not only the bond to the gear but cracks apart and comes out in chunks. Tempted to use P/U as I can cast net shape I want (will reduce load by more than tripling radius of the drive interface but it is not very good a dissipating heat. Not likely going to go EPDM but probably NBR as I can get variety of hardness. IF I can get it at 70A or more might well go Silicone as it easily lives at those temps, but usually too much resilience, too little damping....unless I can get load down far enough.
I could calculate the daylights out of it, but as you might derive for one-off stuff just have to work around what is actually available.
Which brings up a question I've been pondering - and admittedly too lazy right now to put in the study to find the answer myself.
The 14 watts of heat mentioned above: Is that per cycle as I suspect rather than 14 watts per some time unit? 14 watts is a pretty small amount of heat to dissipate. However 14 watts X 1500 rpm (25Hz) adds up to 21000 watts/minute = 1.26 Kwhr = 4300 BTU/minute*. That is a significant amount of heat to dissipate.
* if my pre morning tea button poking is right
Centaflex is a bit fuzzy on that point. I've attached their explanation below.
A long time ago (December 1998 to be precise), a Centaflex application engineer was kind enough to forward an English translation of DIN740. Permissible power loss isn't defined in there.
I just send an information request to Centa.
Attachments
The reply will be interesting to see. I just figured that, because "watt" is a unit of power and the time unit is built in, there wasn't much else left to say. (As with a 100 watt bulb or a 1 HP engine). Watts is just watts, it measures power, and power is what we want to quantify if we're looking at the ability of this coupling to handle a particular amount of energy per unit of time.
Last edited:
Now that I have time to brew a pot of tea - a rather interesting blend of Hairy Crab and artificially aged Pu'er - and time to ponder this heat question:
I'm now thinking that the 14 watts mentioned above is based on some standard for air temp and circulation rate around the device. 14W being how much energy can be dissipated under those conditions to maintain some static internal temperature and is probably limited by the heat transfer within the rubber mass.
My guessing about this at this point in my quest for understanding:
To figure heat generated we would have to know the load on the rubber, how much it deforms under that load and then by looking at the hysteresis graph figure how much work got lost in the process during each cycle.
From that we then would have too look at the frequency to determine amount of work lost per time unit. All pretty straightforward for a simple block of rubber under equally distributed load, but once a point(s) load and non regular shapes are introduced the only practical way I see to calculate this would be with FEA?
Lacking a hysteresis graph for the whole unit (Giubo in this case) some conservative assumptions to make hand math practical might get us close enough to cut down on test time significantly?
All I really learned from this paper Prediction of Heat Generation in Rubber is that the frequency doesn't alter the basic hysteresis graph of the rubber.
