Construction:

The unit consists of a thermal grid (approximately 5x4.5 cm), made of 12 heating elements (approximately 5cm long and 1/3cm wide). Each element has its own thermocouple and will have its own temperature control circuit.

Individual Element Construction: Grill Element #2.

Assembly:

Bracket construction:

Instead of affixing the 12 (or 14) elements directly to the Peltier device as was previously done, an aluminum plate/bracket, was used to facilitate disassembly and avoid damage to the Peltier device.

Assembling the element array:

The elements are held in place by means of a second plate while the epoxy is curing ( see drawing). Notice that the elements are placed in an alternating pattern to avoid contact between the brass end caps that cover the heater wires weld points.

Heater and driver board:

Proposed assembly.

Test:

  [Image: 5cm x 5cm gradient mapping.]
   
  [Image: Same zoomed image.]

Thermocouple Check:

The thermocouples of the 12 elements were checked by connecting them one at a time to the Element Temperature input of our setup. These measurements were performed at room temperature with power applied to neither the Peltier nor the heaters.
#1	23.6		#2	24.5
#3	23.8		#4	24.6
#5	23.4		#6	24.5
#7(P)	23.9		#8	24.6
#9	24.0		#10	24.5
#11	24.*		#12	24.5

#7(P) is the element connected to the Peltier controlling circuit used as the Peltier reference temperature.
#11 was found to be unstable: oscillating between 24.4 and 24.9.	
After #12, #1 was measured again and found to have climbed to 24.6C. This, we assume, was caused by a rise in the room temperature, indicating that the 12 thermocouples, except for #11, were satisfactory.

An Omega 5800 precision thermometer probe was placed at the center of elements #5 and #6 and held in place by means of a rubber band. No thermal grease was used to avoid causing a mess between the elements. No power was applied to the heaters but the Peltier temperature was raised and the resulting temperature readings recorded:
Peltier 	Omega 5800

  20C		  19.8C
  30C		  29.4C
  40C		  38.3C

We will consider these results close enough for now since barely half the surface of the Omega 5800 probe was in contact with one of the 2 elements.

Heaters Preliminary Check:

The heaters of even elements were connected in series to a DC power supply. The voltage was increased until the current reached 1.2Amp. The temperature was then 31.3C on the heated reference element #6.
It seemed like the "Grill effect" was too obvious to be true. We used a hand held thermocouple to spot-check the temperature of each grill element at both ends and center (and sometime in between). The results are shown in the table below.
We realize that this is a crude, inaccurate way to map the grill temperature, but still gives a fairly good idea of the temperature gradient.

element#	1	2	3	4	5	6	7	8	9	10	11	12
								**			*	
Imbedded    	13.4	45.9	15.7	34.8	15.4	31.3	15	30.5	17	37.9	18.7	50
												
right		19	40	20.5	36	19	37	19	38	20	32	23	41
												
center		17	41	18.5	33	18	30	18.5	29	19	33	21	44
			39										38
left		18	35	19	36	19	30	19	31	20	36	21	34

Range			6		3		7		9		4		10


* = unstable (quiets down when Peltier current is turned off).												
** = Peltier controller "Peltier temperature".												

Temperature maps - Peltier @ 15C

Heaters OFF:

element #	1	2	3	4	5	6	7	8	9	10	11	12
									**			*	
Imbedded	15.5	15.7	15.7	15.7	15.5	15.6	15.5	15.0	15.7	15.6	15.1-16	15.7
													
right		18.0	17.5	18.0	17.5	18.0	18.0	18.0	17.5	18.0	17.5	18.0	18.0
													
center		17.5	18.0	18.0	18.0	18.0	18.0	17.5	18.0	18.5	18.0	18.0	17.5
													
left		18.0	18.0	18.0	19.0	18.5	18.5	18.0	18.0	18.0	18.5	18.0	18.0

Odd Heaters ON - 1.2Amp:

element #	1	2	3	4	5	6	7	8	9	10	11	12
									**				
Imbedded	27.2	16.6	39.4	16.2	35.8	15.7	27.5	15.0	49.0	16.2	42.1	15.4
													
right		27.0	21.4	34.5	21.0	32.3	20.1	29.5	20.0	35.0	19.6	42.0	21.5
													
center		27.0	21.0	35.0	20.4	33.4	19.9	27.0	20.0	41.0	20.0	33.0	19.0
													
left		26.7	19.2	36.6	21.0	32.0	20.0	28.8	19.0	37.5	20.1	27.0	19.5
		
Range		0.3		2.1		1.4		2.5		6		15

Even Heaters ON - 1.2Amp:

Same as "Heaters Preliminary Check" above.
element#	1	2	3	4	5	6	7	8	9	10	11	12
								**			*	
Imbedded    	13.4	45.9	15.7	34.8	15.4	31.3	15	30.5	17	37.9	18.7	50
												
right		19	40	20.5	36	19	37	19	38	20	32	23	41
												
center		17	41	18.5	33	18	30	18.5	29	19	33	21	44
			39										38
left		18	35	19	36	19	30	19	31	20	36	21	34

Range			6		3		7		9		4		10

#4, #6, #8 & #10 Heaters ON - 1.2Amp:

This test was performed to avoid the two heaters #2 and #12 that had exhibited higher temperatures than most.
element #	1	2	3	4	5	6	7	8	9	10	11	12
								***					
Imbedded	12.5	13.0	14.2	34.0	15.6	31.0	15.0	30.0	16.9	36.8	15.5	14.0
													
right		15.0	16.5	18.5	26.2	19.0	27.2	19.0	28.0	19.0	31.0	19.5	17.0
													
center		15.0	16.0	18.0	31.1	20.0	28.5	20.0	28.0	20.0	33.0	18.0	17.0
													
left		15.0	18.0	19.5	33.7	20.0	29.0	19.0	29.0	21.0	32.0	18.0	17.0


Range					7.5		1.8		1		2
This case did not give any sensation of high heat. However, when the current was turned up to 1.55Amp causing an increase of temperature at element #7 to 41C, a very hot sensation was perceived but more investigation is required to ascertain that it was not due to hot spots as previously.

#4, #6, #8 & #10 Heaters ON - 1.55Amp:

element#	1	2	3	4	5	6	7	8	9	10	11	12
												
												
right		13.3	14.1	15.6	33.5	17.4	34.5	17.5	36.5	19.0	42.8	18.7	15.4
												
center		13.5	14.0	15.8	41.3	18.5	37.2	18.2	36.3	20.0	45.0	18.0	15.6
												
left		14.2	15.0	17.1	48.8	19.5	39.0	19.2	40.2	20.9	46.5	17.3	16.0
					[48.5]		[39.2]						
												
Range					15.3		4.7		3.9		3.7		
This shows that there was obviously a hot spot (48.8C) and very warm areas (45.0 and 46.5C).

Peltier Temperature Gradient

4cm x 4cm Peltier

A 4cm x 4cm Peltier device temperature gradient over its surface.
x	38.2	x	x	x	38.5	x
38.8	x	x	38.8	x	x	x
x	39.2	x	x	x	39.0	x
38.8	x	x	39.2	x	x	38.8
x	39.0	x	x	x	39.5	x
39.3	x	x	39.5	x	x	39.4
x	39.2	x	x	x	39.8	x
39.2	x	x	39.6	x	x	38.8
x	38.9	x	x	x	39.7	x
38.4	x	x	39.5	x	x	39.4
x	38.6	x	x	x	39.6	x
37.6	x	x	39.6	x	x	39.1
x	38.5	x	x	x	39.4	x
x	x	x	39.1	x	x	R
x	38.2	x	x	x	38.8	R
x	x	x	38.2	x	x	x

  [Image: 4cm x 4cm gradient mapping.]
The measurements were made using the Bismuth Telluride pellets composing the Peltier as a grid. A tiny bead of thermal grease was placed at every other pellet location, and a thermocouple wire was placed, in turn, at each bead and held in position by means of an X-ACTO PC board holder.
The x's represent the pellet locations that were not measured.
The R's represent the pellet locations that were used as temperature reference for the Peltier controller.

The Peltier heat sink water temperature was held at a constant 10 to 11C, the other side of the Peltier (points marked as "R") was set at 39.5C.
The reason for choosing 39.5 rather than 40 was that it is easier to observe fast changing readings around 39.5(39.3, 39.6 etc) than around 40 (39.8 40.2 etc.) The overall temperature gradient was 2.1C, which is not ideal, but much better than the gradient experienced in the case of the heated elements, even with the bad ones removed.

More Temperature Mapping

The weld of the thermocouple used during the mapping of the elements had failed sometime during the test exhibiting intermittent readings. Since there was no way to know when the failure had started two more thermocouples imbedded in epoxy were made and more random mapping done to check the validity of the previous tests.

No Heat:

This first test was performed to check the quality of the thermocouple.
2008 April 18
element #	1	2	3	4	5	6	7	8	9	10	11	12
													
Imbedded													
													
right		16.2			15.6		15.5		15.7		16.1		16.4
													
center		16.1			15.8		15.3		15.5		15.9		16.2
													
left		16.0			16.0		15.3		15.8		16.0		16.2

range		0.2			0.4		0.2		0.3		0.2		0.2

Odd Heaters ON - 1.20Amp:

2008 April 18												
element#	1	2	3	4	5	6	7	8	9	10	11	12
												
												
right		27.2		35.7		31.6		28.9		36.4		46.1	
													
center		26.9		35.3		32.6		26.4		41.7		36.1	
											
left		26.7		35.8		31.7		28.4		37.3		27.1	

range		0.5		0.5		1.0		2.5		5.3		19.0??		

Odd Heaters ON - 1.29Amp:

2008 April 18
element #	1	2	3	4	5	6	7	8	9	10	11	12
													
													
right				38.8		31.9		29.5		39.7			
													
center				37.8		33.6		26.9		*			
													
left				38.4		32.8		29.6		41.5			
													
range				1.0		1.7		2.7		??			
* looked good at first but drifted: the thermoplastic was sliding off the sensor tip. Re-adjusted, but temperature jumped 5C????

Odd Heaters ON - 1.5Amp last check:

2008 April 24												
Odd heaters @ 1.5amp												
Water bath @ 3.0C												
Peltier 15C												
dots of thermal grease were used rather than an epoxy embeded thermocouple.
												
element#	1	2	3	4	5	6	7	8	9	10	11	12
												
												
right				49.0		42.3		35.1		45.5		66.0	
													
center				49.8		44.0		32.2		58.5		52.0	
												
left				51.5		42.5		33.8		54.0		32.8	

range				2.5		1.7		2.9					
Elements #9 and #11 obviously have huge problems.

Heat Sink Brass Tank Temperature Map:

Because we doubted the procedure we used for temperature mapping, we turned over the grill system and temperature-mapped the brass tank used for heatsinking the Peltier. We traced grid lines 1cm apart, placed a dot of thermal paste at each intersection and used the same thermocouple as used above.
2008 April 25												
Heat sink brass tank temperature map with water @ ~40C												

	1	2	3	4	5							
												
												
1	41.4		41.2		41.2							
2					41.5							
3	41.2		41.1		41.7							
4					41.5							
5	41.2		41.1		41.5							

No problem seen with the technique.

Conclusion

  1. A better way to measure temperature gradients is needed.
  2. A maximum of 2C gradient will be difficult to reach even with careful selection of the elements.
  3. Wider than 5cm grill would probably require split elements: with multiple heater thermocouple pairs.
  4. Using custom narrow Peltier devices may not give much better results. See the Peltier element temperature map.

Started 2008 April 7
Last revision: 2008 April 25