2KW Bearing Heater / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

2KW Bearing Heater

1 2 next

Move Thread

LAN_403

Raka

Tue Aug 25 2009, 06:36PM

Registered Member #1838 Joined: Tue Dec 02 2008, 06:01PM
Location:
Posts: 38

It surprised me that I couldnâ€™t find any diy Bearing heaters on the net. Here is my attempt:

Core cross section: 30x30mm
Coil: 600 turns AWG12
Supply: Mains 220V 50Hz

I expected it would need a shunted (with air gap) core, but the current drawn is surprisingly low.

klugesmith

Wed Aug 26 2009, 01:22AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Very nice work!
I'd love to know:
Where did you get the core laminations?
How much current does it draw, with and without the bearing?
What are typical bearing temperatures and heating times?
Thanks
Rich

Raka

Wed Aug 26 2009, 06:42PM

Registered Member #1838 Joined: Tue Dec 02 2008, 06:01PM
Location:
Posts: 38

Im rather disappointed with the results.
The specs of the bearing in the picture is:
ID: 45mm
OD: 85mm
Width: 30mm
Mass: 0.6Kg

The bearing causes a current draw of 3A from the wall. It gets heated from 24C to 100C in exactly 3min.

The core is from a transformer I found on the dump. It used to be a standard E-I type, until I laid my hacksaw on it. Unfortunately I donâ€™t have any specs on the original transformer, nor any pics.

I had the Reco.sc220 bearing heater in mind (

) when I started off. Its core is 19x19mm for 2KW. It seems as if my core saturates at less than 2KW. I noticed that the two uprights is rather magnetic during operation while the horizontal bar stays neutral. Also there is not much difference in current drawn when the bearing is fitted or not. The current is about 7A with no load or crossbar

The temperature is sensed with the pn junction of a transistor and is clipped on the bearing with a magnet.

Ps: The wire only looks charred. I donâ€™t like the looks of it, but itâ€™s brand new.

klugesmith

Wed Aug 26 2009, 08:59PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Raka,

Also nice work on the enclosure and temperature probe.

Are you looking for electromagnetic engineering help here?
Perhaps the 2 kW figure for Reco sc220 is based on 60 Hz and the thickest bearing they could fit on a 51x51mm crossbar.
You probably know that the heating power is proportional to the squares of:
* crossbar thickness
* [edit] and crossbar width (yes, power goes as square of core area)[\edit]
* crossbar magnetic flux density
* mains frequency
[edit] Oops, and based on V^2/R model the maximum power is proportional to (not squared) [\edit]
* bearing race cross-sectional area
* bearing race electrical conductivity

Have you seen the inside of a commercial bearing heater? As you mentioned, they might depend on leakage inductance (from long & thin core legs? magnetic shunt?) to limit the input current in case of, say, putting a thick copper bushing on the crossbar. But maybe just a fuse, circuit breaker, or thermal switch.

I figure you're in the right ballpark, perhaps on the long side, with the number of primary turns.
Before messing with the main winding, here are some easy measurements that would answer key questions:
* primary coil resistance
* AC voltage developed in test coil of 10 or 20 turns around middle of crossbar (need not fit closely)
* AC voltage developed in same test coil on core close to primary coil.
* AC voltage developed in test coil wound around middle of primary coil.

-Rich

[edit] I'll second what Richie said -- make sure the mating surfaces of crossbar and main core are smooth and flat; even the thickness of paint makes a significant gap. Care to try the old machinist's art of scraping to flatness?

[re-edit] In a pinch you might lap the mating surfaces to each other using sandpaper to wear down the high spots.
But on second thought, a core gap might be OK. It would reduce the inductance and increase the no-load magnetizing current, to relatively more stable & predictable levels. With air gap you could still flirt with saturation in the crossbar, at H levels where the magnetizing current penalty would be much less than in an efficient continuous-duty transformer. Test coil voltage measurements can directly indicate the magnetic fluxes in different sections of the core.
The theoretical maximum for your existing primary coil is 220/600 = 0.367 V/turn(RMS), corresponding to a peak flux of 1.8T in 30x30mm at 50 Hz (review please?) -- but I bet you knew that when you designed the coil.

GeordieBoy

Wed Aug 26 2009, 09:02PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

> I noticed that the two uprights is rather magnetic during operation while the horizontal bar stays neutral. Also there is not much difference in current drawn when the bearing is fitted or not.

You have to keep the air-gap between the two uprights and the I-piece (the bar across the top) as small as possible. You want most of the magnetic flux to flow around the complete magnetic circuit, including the laminations with the bearing around them. Make sure there is no paint of varnish or anything on the top surface of the uprights, or on the I-piece that goes across them. This should have smooth metal surfaces that mate tightly with no air gap to ensure that there is minimal flux leakage. Your comments quoted make me think this may not be the case.

The current draw with a load fitted should be considerably higher than the no load current with no bearing present. If this is not the case then it means that most of the input current is going into other losses and overall efficiency will be very poor. No load current with the I-piece securely in place should be very low if you have enough turns on the primary, the correct applied voltage/frequency and no significant gaps in the magnetic path.

-Richie,

Raka

Thu Aug 27 2009, 07:38PM

Registered Member #1838 Joined: Tue Dec 02 2008, 06:01PM
Location:
Posts: 38

> Are you looking for electromagnetic engineering help here?

Yea Id welcome any help or comments, Your inputs are always of great value sometimes even more so than practical experience.

I did polish the surfaces of the contact points last night, The time to heat the bearing now reduced to 2min flat (Thanks Richie) It doesnâ€™t make cense to me that such a lot of flux will jump the 90mm between the posts rather than travel through a couple of layers of paint, unless there is no more room in the core (saturated). The thing is the core doesnâ€™t get warm in the slightest.

> Have you seen the inside of a commercial bearing heater?

No. I can only find commercial sites on bearing heaters and I donâ€™t know someone who owns one. Id love to have a peek inside. Some sites also mention build in de-magnetizing, liked to see that too. Cant say that my test bearing is left magnetized afterwards.

>Before messing with the main winding, here are some easy measurements that would answer key >questions:
>* primary coil resistance
>* AC voltage developed in test coil of 10 or 20 turns around middle of crossbar (need not fit closely)
>* AC voltage developed in same test coil on core close to primary coil.
>* AC voltage developed in test coil wound around middle of primary coil.

The test you mentioned: A coil of 20turns develop 5.85V over the crossbar and 6.31V low down on one of the legs. It used to be 4.82V crossbar and 5.77V leg.
A four turn secondary on top of the primary yields 1.47V So that is about 80% of the flux that make it to the crossbar?
No load current is down to 1.66A

I also fitted a magnetic shunt with 2mm air gap and 30x30mm cross section, it reduces the voltage in the test coil across the crossbar to 5.41V

The crossbar and the uprights are moderately magnetic now when in operation.
Oh,yes and the dc resistance of the main coil is 0.8 ohm

>The theoretical maximum for your existing primary coil is 220/600 = 0.367 V/turn(RMS), corresponding to
>a peak flux of 1.8T in 30x30mm at 50 Hz (review please?) -- but I bet you knew that when you designed
> the coil.

Oops. I didnâ€™t check that beforehand because I knew of the commercial heater with a much thinner core for 2KW. Yes your calculation is on the dot Im trying to squeeze 1.8T through a core that is probably only good for 1.3T at most. But why isnt the core heating up then?

Tonskulus

Thu Aug 27 2009, 08:45PM

Registered Member #1223 Joined: Thu Jan 10 2008, 04:32PM
Location:
Posts: 133

Power that is delivered to a bearing can be calculated:
Uin*In(with bearing fitted) - Un*In(no load).

You said that current drawn is surprisingly low but for me, 3Amps is not very low.

So if input current wont rise with bearing fitted, it shouldnt heat at all. If input current rises from 3A to 4A, there is 1A difference so 1 x 220V = 220Watts of heating power.

It looks like saturating problem..

Raka

Thu Aug 27 2009, 09:30PM

Registered Member #1838 Joined: Tue Dec 02 2008, 06:01PM
Location:
Posts: 38

Tonskulus, that bearing is the one that fits the tightest (best coupling) you might get one that is a bit wider, but I recon that one represents a rather demanding load, about the worst load it will need to handle Therefore I would have liked to see that bearing draw about full power or at least 2/3 of full power. I expected to see about 6A. The wire is rated for 9.3A corresponding to 2KW power

klugesmith

Fri Aug 28 2009, 12:32AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Let's see now.
From stated bearing mass and heating rate, your apparent power has improved from 127 to 190 watts.
Could be substantially more if you measured the outer race temp, and inner race heats faster. Compared to the outer race, if the inner has a similar cross-section and volts/turn (i.e. neglect magnetic flux in the bearing metal), then inner has 5/8 as much circumference, resistance, and mass, with 8/5 as much current and power dissipation. Might the thermal conduction time constant between the 2 races be the better part of 1 minute?

Is it possible that you've got a stainless steel bearing with relatively high electrical resistivity? I made a SWAG at the part dimensions, to figure the resistance of both races in parallel.
Taking resistivity of 2e-7 or 7.4e-7 ohm-m (1080 carbon steel at 60C vs. a random austenitic stainless), the shorted turn R is 107 or 397 microohms. With measured 293 mV/turn it would carry 2728 or 737 amperes, thus get 799 or 216 watts. (I think the former number is the best you can hope for).

Now those secondary ampere-turns should be matched by additional primary currents of 4.5 or 1.2 amperes, roughly in phase with mains voltage. They would not add linearly to the no-load magnetizing current, which I assume is mostly reactive.
[edit] I bet magnetizing current waveform here (near saturation) is highly distorted and not amenable to traditional vector addition, even if you use a true-RMS meter. A true wattmeter or watthour meter would be handy. [\edit]

Got to run. Hope this gives you some useful ideas to explore.
Rich

Raka

Mon Aug 31 2009, 05:45PM

Registered Member #1838 Joined: Tue Dec 02 2008, 06:01PM
Location:
Posts: 38

I measured the temperature of the inner race, because itâ€™s just the temperature of the inner race that matters if you want to expand the bearing to slide over a shaft.
Im pretty sure that the bearing is made of carbon steel, judged by how easily it rusted and how strong the magnetic probe is attracted to it
I really appreciate the calculations you did, especially the maximum wattage to aim for. Thanks. I donâ€™t have the know-how to do that

I enlarged the spool today and will be winding some more windings tomorrow. I figure that if the voltage on the test coil is 0.2925V/turn at the crossbar and 0.3675V/turn across the primary than 79% of the flux goes inside the core, and that means that I have 79% of the windings needed? Doesnâ€™t really matter that much, I only have enough wire left for an estimated 150 turns, it might just make it. So Ill wind it and keep you posted with the results.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob