[Stoves] Stirling to drive fan

[AJH]

On Sun, 24 Sep 2006 08:48:34 GMT, steve wrote:

>I was thinking of the heater/light first of all. At 150mm
>you could fabricate the "washers" from sheetcut into 'C'
>shapes and rolled/welded

OK but see below it's a large number of washers. The reason I
suggested clamping is because of the difficulty in welding all these
joints. I did think building a pile, washer by washer, and friction
welding them one at a time may be possible.

>> I think you may be disappointed, the nickel-iron one I
>> chose was the best I  could see from common metals, the
>> others were fairly esoteric nickel alloys.
>
>I haven't tried it, I have no feel for the problem yet,
>Incoloy perhaps ? 

Here's an old post on the subject:

On Sat, 06 Nov 2004 15:42:27 +0000, list at sylva.icuklive.co.uk wrote:

>
>Is there a simple way for an artisan crafstman to make a usable teg?
>
>As I said before the most suitable commonly available metals would
>seem to be nickel and iron.

I left out some initial calculations from a google:

Material  Seebeck  Coeff. *
Iron      19 micro Volts/deg C @ 0C
Nickel    -15 

Assuming this is somewhere near linear in the range 0 to 400C:

So if the water in the pot is the cold sink and boiling and we have a
flame of 800C but the boundary layer halves what the thermo couple see
then we have a delta t of 300C to play with. If my interpretation is
correct we should have 34micro Volts times 300 output per junction, we
are looking for 1.25V , suggesting 123 couples or 246 junctions,
fiddly.

*******

And another:

To: "Stoves List" <STOVES at LISTSERV.REPP.ORG>
Subject: Re: [Stoves] RE: The dream power plant for the tiny fan
From: list at sylva.icuklive.co.uk
Date: Sat, 06 Nov 2004 19:42:51 +0000
Cc: 

On Sat, 6 Nov 2004 21:27:42 +0200, Crispin Pemberton-Pigott wrote:

>I need basic info on this.  I understand that the junction produces
>electricity without a thermal gradient.  Is this correct?
 
Our posts to the list were in the ether simultaneously.

No, as with any heat engine there has to be a hot source and a cold
sink, heat flowing between them does the work (creating electricity)
>
>That K-Band probe should give you about 75mV I think.  Not much but it is
>only one junction.

You need to specify the difference in temperature, are you sure you
mean milli rather than micro Volts per degree C?

http://www.fact-index.com/t/th/thermocouple.html

suggest a K type to be 41 micro Volts/deg C, this is chrome-alumel.
both alloys with high percentages of nickel.
>
>I have not heard of Belleville washers. Perhaps we know them by a different
>name.

It's a spring washer formed by stamping the washer into a flat
truncated cone.

see
http://www.mech.uwa.edu.au/DANotes/springs/intro/discSpringsBIG.gif

but alternate nickel and iron, fill the spaces with insulation without
ruining thermal contact at the edges and have the flue gas come up the
central hole without the tensioning device interfering with flow.
>
>How about a stack of stainless steel and iron?  I can afford to put together
>quite a stack of these, perhaps 0.5mm thick.  They could just get coated
>with soot though I suppose it will conduct.

You will have to try differing metals and see, if you have a spot
welder to make test junctions and a high impedance Voltmeter able to
discriminate a few micro Volts you may find something, otoh chromel
and alumel may be available to you.
>
>I  am still stuck at the stage of trying to work out what gets hot.  Do I
>have to cool one end of the wire?

Yes, ambient air can be the cool side and the furnace the hot side,
though I am advocating furnace one side and boiling water the other to
make use of all the heat.
>
>I was thinking that I make U-shaped wires and stack them up in a pile of U's
>with the ends spotwelded together so there is a circuit going back and forth
>and upwards from U to U.  Then heat the ends and keep the U body cool.
>Right or not?

Sort of but I think you have a series of Vees, with the legs of the
dissimilar metals, and keep one of the joints hot and the other cold.
See my later calculations for the number of junctions, 250 odd unless
I am mistaken.
>
>If I stack pennies and nickles and heat it (no cooling at any point) will I
>get a voltage top to bottom? 

yes if your pennies are copper and nickels nickel, alloying seems to
produce different Seebeck properties.


>I understand that the power comes from heat
>absorbed making electrons move.

Yes and in one bar they would just move from hot to cold and stop once
the potential was established, it is the change in metal between hot
and cold junctions that allows this potential to drive a current.


>  Perhaps there is no cooling required.

Yes there must always be a difference in temperature and the heat must
flow from hot to cold.


********


It all makes this a rather long post but you will see that the
electricity produced is only a fraction of the heat that passes
through the junctions, and this is recirculated. The total heat of
combustion and electrcity still passes out the top of the stack unless
the heat dissipated in the fan plus LEDs doesn't get captured, even so
it's fairly insignificant.

So the temperature of the cold side at the top of the stack is near
ambient at the bottom probably getting on for 400C, the inside will be
some 800+C, depending on fuel cv, fuel moisture content and excess
air. Any higher and the metals won't survive. The differential between
the stack inside top and bottom   will be dependant on the
conductivity of the pile and how tall it is.

AJH