about seal a shatf under 30 meters sea water

[Oliveros]

i'm a mechanical engineer who are developing a inner pipe inspection robot, it will work under 30 meters, in sea water. i need to make the shaft of the wheel waterproof, so the sea water don't let enter to the electronical system. the shaft is near 1 cm diamenter.

please help me

 

[sreid]

There have been posts similar to this before. I'm not an expert in this area but a leak proof rotary seal at this kind of pressure is nearly impossible (there are, however, presure compensated rotary joins such as those used in "Newt Suit" diving suits). I think there are two approaches used for external drive wheels.

1) The entire internal space is nonconductive fluid filled and maintained at outside pressure.

2) The wheels are on external shafts and magnetically coupled to the inside motor shafts using permanent magnets across a non-magnetic barrier..

 

[Oliveros]

thanks,

please tell me more about ...:

"The wheels are on external shafts and magnetically coupled to the inside motor shafts using permanent magnets across a non-magnetic barrier.."

can you tell me about those kind of motors?

 

[sreid]

Magnetic couplings are widly used commercially in pump drives. To visulize, put two bar magnets on opposite sides of a piece of plastic. The noth and south poles will attract each other. If you turn one magnet the magnet on the other side will be pulled along. If the magnets are on shafts and bearings, this works very well. High energy magnets are used (Alnico or Neodymium-Boron Iron magnets-a good source for Neo magnets is Magnetic Component Engineering).

Under water at high pressure the plastic barrier might need to be domed so it won't deform under pressure and the magnets appropiatly shaped (the closer the magnets are to each other the stronger they couple). The external shaft might simply be a stainless steel shaft in carbon bearings fully exposed to the sea water (or simple seals to exclude grit).

To get a feel for the magnet couplings you could buy a commercial pump and disassemble it (Graingers-McMasters, etc.)

 

[notnats]

30 metres is only 3 bar pressure. I would have thought that commercial seals are available for this. The o ring seal on the spindle on the washing machine supply tap in my laundry at home doesn't leak (when new) and mains water pressure can be up to 10 bar (100 metres.) The o rings on gas struts do not leak gas for years at 6-10 bar.

Jeff

 

[Tofflemire]

Hi All,

I would like to add one thing to this topic. If you have a oring shaft seal, the weight of the robot onto the shaft may make your oring oval which may cause a leak path. I would try to ensure that the loads are not transferred to the oring sealing area.

Tofflemire

 

[notnats]

Well I wasn't going to use the o rings as bearings.

On further thought, if the shaft has 2 o rings with the space between them flooded with oil, this would both lubricate the o rings and provide a barrier to any water leaks.

Or, if the robot has an umbilical cord, incorporate an air line to pressurise the housing. This would keep the water out if the air pressure was greater than the water pressure.

Jeff

 

[sreid]

Shaft seals will work well at first. The duty cycle of this machine is likely to be higher than a water tap or a washing machine so a leak might be expected to occur earlier. Also the results of a leak is probably much worse. Sea water in electronics, motors, etc. is not pretty.

 

[Beggar]

Seems to me that you'd go with a standard rotary seal and then determine the required maintenance interval at which the O-rings need to be inspected or replaced.

I do think you'd be wise to go with a pair of O-rings on the shaft rather than a single. I'm not that the oil between the two O-rings will provide much benefit. More important, I think, would be to lube the O-ring/shaft at the proper mainenance intervals.

Take a gander at the Parker literature for guidance or contact one of their application engineers.

 

[zeusfaber]

If you can't make the seal leakproof, why not just take control of the leak?

Suggest you consider pressurising the inside to "ambient + a little bit" from a small supply of compressed air. With a reduced pressure differential across the seal, there should be a ower tendency to leak, and what leakage there is will be in the right direction (air out, not water in).

A.

(b.t.w., absolute pessure at 30m is 4 bar, not 3)

 

[unclesyd]

Here are two people that maybe of help.

Look at rotary seals

http://www.kalsi.com/index.htm

These people also have bearing materials as well as seals

http://www.busakshamban.us/

 

[Cockroach]

This is a very easy problem to solve. However Oliveros, there is not enough information in your original statement.

I will go with the assumptions presented to date and assume a rotary seal. Please note there are other seal types, static, dynamic (reciprocating, pneumatic floating and rotary) as well as face seal alternatives. Each of these introduces the variance of elastomer compression, that is, squeeze. The squeeze also depends on the diameter of the o-ring specified. Commercial available types are 0.070, 0.103, 0.139, 0.210 and 0.275 diameters which coincide to 000, 100, 200, 300 and 400 series o-rings respectively.

For the shaft size specified, 1.0 cm or 0.394 inches, I would use a 000-045 Series o-ring because machining the male o-ring groove would weaken the shaft somewhat. It is more cost effective to machine a male groove rather than female because the machinist has much more control on the external profile. In the elastomer series selected for rotary application, the squeeze is no greater than 11% for 200-1500 RPMs. Thus the gland specifications are 0.262/0.260 diameter X 0.075/0.079 inch wide; this gives a mid range squeeze, approximately 5.5%. Your mating piece must therefore be drilled and reamed 0.396/0.398 diamter with 63 Ra finish. A number 2-010 o-ring (0.239 ID X 0.070 CS diameter) should be used.

Now check the elastomer stretch, typically 2% to 3%. I get 9.6% which is really, really high for the application. Expect o-ring necking to lessen your squeeze, but you are in no-mans land since the a 2-011 o-ring is too large for the groove specified. A quick review for a female groove cut into the cylinder of the mating piece gives no additional advantages. Similarly, going to a 100 Series o-ring would only weaken the shaft in male groove design or present manufacturing issues with female groove design.

Finally look to lessen elastomer stretch by lessing the sqeeze. You're talking 30m of seawater, I would argue not much pressure and the o-ring selected is good for 1500 psi without the necessity to resort to backup rings. Going to a 2% squeeze would imply a male groove specification of 0.259/0.257 dia X 0.075/0.079 inch wide, the stretch drops to 8.4%. This is still on the high side, but professionally speaking, I've used specified worse with experiencing problems. For the sake of discussion go with it using 2-010 o-ring.

Now select material, dependent on fluid application and temperature only. Nitrile is acceptable, I'm a big fan of a flurocarbon alternative like Viton. The typical durometer or rubber hardness is 70.

In summary, cut a 0.259/0.257 diamter male groove, 0.075/0.79 inches long into the shaft and give 0.02R fillets in the groove, 0.005/0.007R shoulder fillets. The surface finish is 63 Ra. Drill and ream the cylinder or mating piece to the shaft, 0.396/0.398 inch diamter with interior finish 63 Ra. Specifiy a 2-010 o-ring, Parker-Hannifin Corporation made of Viton, durometer 70.

...and there you have it...

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[sreid]

A question; are o-rings generally accepted for use in rotary seal applications?

 

[Cockroach]

Yes SReid, absolutely. The Parker-Hannifin specification calls out explicitly for gland geometry and suggested squeeze, stretch on o-ring.

As an update to the excessive stretch issue, I would hot splice a bastard o-ring from 0.070 inch primer cord. That way, you get exactly the right stretch, between 2% and 3% while maintaining squeeze. Unfortunately for the o-ring gland specified, there is no imperial commercially available o-ring that will meet the recommended specifications.

Not a big deal, I have hot spliced many specialty o-ring(s). Use a certified manufacturer and be sure to vulcanize the primer cord at the joint! I haven't worked out the exact o-ring ID yet, but it is a straight forward computation now that the gland has been dimensionalized.

Hope this helps you out somewhat.

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[Dennyd]

Oliveros,

As I understand your issue you want to prevent seawater at a depth of 30m from causing a problem with your electronics. Most of the discussion above deals with ways to prevent seawater from entering the shaft seal for a wheel. Have you also considered making your electronics sealed from seawater? How about encapsulating the electronics in some potting compound. At 30m you are likely to get some water in your system. I'd suggest you encapsulate as much of you electronics as possible.

- - -Dennyd

 

[Cockroach]

I would try to resist potting compound of any sort with this application.

One obvious problem will be with seal cracks developed from relative motion between mating parts. I guess a silicone rubber could be used, but given the environment and usage over time, I doubt it would last long. An alternative could be usage of a gasket, but the draw back here is with the compression requirement to make a seal. I would expect temperature variations, this means thermal expansion and contraction at the joint, hence variation in gasket compression. Perhaps the risk is too great for this type of application.

O-rings are used predominently across multiple industries without significant problems. That is why they are so popular: ease of use, simplicity in design, inexpensive and readly available commercial supply. Of couse the rubber compound itself exhibits memory, the mating part movement argument is immediately addressed. So too is the thermal issue since the rubber itself does not typically change characteristics until 125F or so.

I think either an o-ring is acceptable, perhaps a polypak would be the backup plan.

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[turboengineer]

There is no such thing as a leak proof seal. All seals have a leak rate, so you need to find a rate that is acceptable or rethink your approach. I read in one of the posts a little about magnetic couplings. I use them often in hermetically sealed pumps that I design. If you really need to be leak tight then that’s the best way for you to go. They are easy to design, there are a number of manufacturers out there that make them, and they don’t really cost that much.