Aims

This is an attempt to produce an experimental comparison of the output of two commercial "twin-gear" fruit and vegetable juicers: the "Twin Health" and "Samson Ultra" machines.

Test equipment

The Samson Ultra model tested was about two years old. The Twin Health machine was a reveiw sample - with little visible wear on it.

The tests were made over a ten-day period.

Methodology

Input produce, output fibre, and output juice were all measured using highly accurate electronic scales. "Lost material" left inside the juicer was calculated by subtracting the final output figures from the input ones.

Times given record the time to the last piece of prodcue was fed into the machine. At the end of this time, the machines were left running for 30 seconds before the outputs were weighed.

Times given record the time to the last piece of prodcue was fed into the machine. At the end of this time, the machines were left running for 30 seconds before the outputs were weighed. The machines were fed at a rate that seemed appropriate.

Jams were recorded. After a jam, the machine was dismantled, cleaned, and the test was resumed with any remaining produce.

Mainly because the machines had radically different sized gears, I measured juice output and fibre output and calculated how much produce remained stuck in the machine at the end of each run.

These figures were used to calclate a "juicing rate" - that factored out the "initial loss" that arose during the early stages of each run.

To determine if the "lost" material was still accumulating over the course of the run, I divided the produce in many of the runs in half - and measured the output half way through the run.

Results

Executive summary

In my tests, the Samson Ultra produced slightly more juice than the Twin Health machine:

Machine	Total Juice/Kg
Twin Health	8.359
Samson Ultra	8.686

A more detailed breakdown shows some variation based on product type:

[Here, the "Twin Health" figures were divided by the "Samson Ultra" ones. Figures bigger than 1.00 are good for the Twin Health machine]

Produce type	Total juice
Wheatgrass	0.77
Collard greens	1.00
Celery	1.04
Rhubarb	0.89
Sprouting broccoli	0.85
Cabbage	1.01
Beetroot	0.89
Carrots	1.03
Sweet potatoes	0.90
Red grapes	1.06
Strawberries	0.80
Grapefruit (peeled)	0.82
Melon (gala)	0.97
Melon (honeydew)	0.97
Apples (GS)	0.90
Oranges	0.83
Pineapple	1.07

Brief interpretation:

Figures for juice extraction for the two machines are broadly similar - both machines are reasonably competent at extracting juice from produce.

To summarise the differences:

• Juice quanity
  • Overall, the Samson Ultra machine extracted 4% more juice from all the combined produce. It produced a bit more juice the majority of the tests.

• Juice quality
  • The Twin Health machine shipped with finer fruit and vegetable screens and - according to my subjective impressions - typically did a much better job of avoiding dumping fibre into the resulting juice. The result was more palatable, smoother, less lumpy juice.

    Also the machine typically produced less froth - and in many cases, a lot less froth - a positive sign.

• Initial loss
  • "Initial loss" is a measure of how much produce is lost inside the machine - without being output as juice or fibre.

    This can be a relevant factor when juicing relatively small volumes of material.

    Fibre left inside the machine isn't necessarily undesirable - it's only unextracted juice that might be a cause for concern.

    The Twin Health machine usually had a larger "initial loss" of produce - a fact I attribute primarily more produce being lost in its larger gears.

    The Samson Ultra experienced greater initial loss sometimes - typically in cases where the output was sticky and accumulated around the outside of the screens.

• Eventual juicing rate
  • The "eventual juicing rate" is an attempt to measure efficiency when juicing large volumes of produce.

    The idea is to produce a figure for the rate of output once the machine has been left running long enough to reach a steady state).

    Ideally, this should be determined by running some large- scale tests - but since I had collected some relevant data I was able to make an estimate of this.

    Overall, the Twin Health machine appeared to exhibit a slightly greater efficiency in its eventual rate of juicing green vegetables - but seemed slightly less efficient at juicing fruit.

Twin Health vs Samson Ultra

"Initial loss"

• This estimates how much produce remains in the juicer. These figures in this column are derived from data which is a bit noisy. Figures bigger than 1.0 could be undesirable for the Twin Health machine.

[Here, the "Twin Health" figures were divided by the "Samson Ultra" ones. Figures bigger than 1.00 are undesirable for the Twin Health machine]

Produce type	Initial loss
Wheatgrass	1.67
Collard greens	7.35
Celery	1.22
Rhubarb	3.17
Sprouting broccoli	7.58
Cabbage	1.57
Beetroot	2.03
Carrots	2.26
Sweet potatoes	2.63
Red grapes	0.37
Strawberries	2.78
Grapefruit (peeled)	1.67
Melon (gala)	1.59
Melon (honeydew)	0.76
Apples (GS)	3.08
Oranges	1.94
Pineapple	1.29

Juicing rate

• ...is an attempt to measure of the rate of juice extraction once the machine has been juicing produce for a while. This figure is the relevant one if a lot of produce is likely to be juiced. Figures bigger than 1.00 are good for the Twin Health machine.

Here is a more detailed table - from which the above summaries were largely derived:

The percentage figures here represent the proportion of the input mass which was ejected as juice.

The "Total juice yield" column gives the figures for the test produce I used - while the "Juicing rate" column is the result of factoring out the "inital loss" - giving a crude estimate of the expected output of the machines if they are fed very large volumes of produce.

Produce type	Juicing rate		Total juice yield
Produce type	Twin Health	Samson Ultra	Twin Health	Samson Ultra
Wheatgrass	92%	99%	60%	78%
Collard greens	74%	71%	64%	64%
Celery	92%	85%	79%	76%
Rhubarb	98%	89%	73%	82%
Sprouting broccoli	80%	70%	58%	68%
Cabbage	78%	75%	73%	72%
Beetroot	61%	63%	51%	57%
Carrots	66%	60%	60%	58%
Sweet potatoes	50%	50%	43%	48%
Red grapes	95%	96%	86%	81%
Strawberries	92%	96%	70%	87%
Grapefruit (peeled)	80%	86%	58%	71%
Melon (gala)	94%	89%	76%	78%
Melon (honeydew)	82%	85%	75%	76%
Apples (GS)	85%	86%	73%	81%
Oranges	77%	82%	60%	72%
Pineapple	92%	83%	76%	71%

Raw test data

These tables intended mostly for reference purposes - but see the appended notes.

Twin Health juicer yields

Produce	Screen	Total	OK?	T/s	Juice	Fibre	Lost	T/s	Juice	Fibre	Lost	F
Wheatgrass	Fine	0.116	Y	-	-	-	-	160	0.070	0.006	0.040	0
Collard greens	Fine	0.472	Y	-	-	-	-	580	0.304	0.109	0.059	1
Celery	Fine	0.257	Y	25	0.075	0.004	0.049	50	0.193	0.007	0.057	-
Celery	Fine	0.570	Y	88	0.210	0.020	0.055	180	0.462	0.048	0.060	-
Rhubarb	Fine	0.690	N	72	0.259	0.006	0.080	n/a	0.507	0.012	0.171	-
Sprouting broccoli	Fine	0.330	N	121	0.087	0.020	0.058	n/a	0.193	0.046	0.091	9
Cabbage	Fine	1.000	Y	240	0.344	0.096	0.060	538	0.728	0.206	0.066	8
Beetroot	Coarse	0.334	Y	75	0.075	0.047	0.045	154	0.171	0.108	0.055	-
Carrots	Fine	0.500	Y	67	0.132	0.070	0.048	147	0.299	0.151	0.050	-
Carrots	Fine	0.750	Y	99	0.215	0.108	0.052	199	0.452	0.239	0.059	-
Sweet potatoes	Fine	0.360	Y	-	-	-	-	140	0.155	0.155	0.050	1
Red grapes	Medium	0.600	Y	90	0.250	0.020	0.030	170	0.538	0.026	0.036	-
Red grapes	Medium	0.500	Y	57	0.191	0.007	0.052	121	0.409	0.028	0.063	1
Strawberries	Coarse	0.796	Y	91	0.240	0.021	0.137	179	0.556	0.051	0.189	-
Grapefruit (peeled)	Coarse	0.500	Y	34	0.137	0.020	0.093	85	0.290	0.071	0.139	0
Melon (gala)	Coarse	0.680	Y	105	0.250	0.013	0.077	156	0.520	0.034	0.126	1
Melon (honeydew)	Medium	1.000	Y	128	0.369	0.055	0.076	243	0.764	0.136	0.100	2
Melon (honeydew)	Medium	0.822	Y	-	-	-	-	201	0.603	0.101	0.118	1
Apples (GS)	Medium	0.444	Y	71	0.145	0.020	0.057	143	0.325	0.058	0.061	12
Oranges	Medium	0.430	Y	30	0.121	0.018	0.076	65	0.262	0.077	0.091	0
Pineapple	Coarse	0.736	Y	-	-	-	-	185	0.558	0.046	0.132	1
Totals								4082	8.359		1.704	37

Samson Ultra juicer yields

Produce	Screen	Total	OK?	T/s	Juice	Fibre	Lost	T/s	Juice	Fibre	Lost	F
Wheatgrass	Fine	0.116	Y	-	-	-	-	151	0.091	0.001	0.024	6
Collard greens	Fine	0.472	Y	-	-	-	-	330	0.301	0.163	0.008	17
Celery	Fine	0.257	Y	26	0.076	0.019	0.033	51	0.180	0.036	0.041	-
Celery	Fine	0.570	Y	65	0.230	0.031	0.024	132	0.449	0.072	0.049	-
Rhubarb	Fine	0.690	Y	62	0.309	0.034	0.002	127	0.568	0.068	0.054	-
Sprouting Broccoli	Fine	0.330	Y	70	0.101	0.040	0.024	210	0.225	0.093	0.012	18
Cabbage	Fine	1.000	Y	291	0.348	0.121	0.031	571	0.715	0.243	0.042	13
Beetroot	Fine	0.334	Y	58	0.085	0.059	0.023	148	0.192	0.115	0.027	-
Carrots	Fine	0.500	Y	150	0.140	0.070	0.040	230	0.296	0.167	0.037	-
Carrots	Fine	0.750	Y	110	0.192	0.156	0.027	210	0.428	0.296	0.026	-
Sweet potatoes	Fine	0.360	Y	-	-	-	-	143	0.172	0.169	0.019	3
Red grapes	Coarse	0.600	Y	45	0.222	0.008	0.070	91	0.483	0.020	0.097	-
Red grapes	Coarse	0.500	Y	59	0.192	0.005	0.053	120	0.404	0.017	0.079	1
Strawberries	Coarse	0.796	Y	85	0.326	0.012	0.060	165	0.698	0.030	0.068	-
Grapefruit (peeled)	Coarse	0.500	Y	41	0.171	0.023	0.056	86	0.358	0.059	0.083	0
Melon (gala)	Coarse	0.680	Y	105	0.251	0.034	0.055	154	0.532	0.069	0.079	1
Melon (honeydew)	Coarse	1.000	Y	119	0.374	0.059	0.067	235	0.775	0.093	0.132	1
Melon (honeydew)	Coarse	0.822	Y	-	-	-	-	180	0.619	0.085	0.118	1
Apples (GS)	Coarse	0.444	Y	51	0.182	0.020	0.020	100	0.361	0.058	0.025	10
Oranges	Coarse	0.430	Y	31	0.143	0.030	0.042	65	0.313	0.070	0.047	0
Pineapple	Coarse	0.736	Y	-	-	-	-	125	0.526	0.108	0.102	4
Totals								3624	8.686		1.169	75

Notes about the data:

• Regrettably, the wheatgrass test used a relatively small volume of material. As a result, little fibre was output by the end of the test - and the "initial loss" figures for this test are practically worthless.

• I used a coarse screen in the Twin Health beetroot test by mistake. The results were included anyway.

• The Twin Health machine jammed twice during the testing. In the case of the rhubarb test, the machine failed to output any fibre. The fibre accumulated inside the machine until it jammed solid - and not even reverse gear would work. In the case of the purple sprouting broccoli the cause of the jam was feeding in a particularly thick piece of stem. While such jams are not encouraging, I do not think they are very significant to normal usage - and would not recommend avoiding the Twin Health machine on the grounds that it's more likely to jam up.

• Grapefruits, oranges and pineapples were the only things that were peeled before juicing.

• Produce was chopped into small peices before the tests started.

• The last column "F" (sometimes) records the volume of foam produced.

Twin Health juice screen comparison

This was a test to look at the difference between the two "Twin Health" juice screens.

The screens had different sized holes - and different strength pressure regulators.

The tentative conclusion was that effect (on juice yield) of the larger holes was roughly cancelled out by the reduced pressure regulation - at least in the case of these apples.

Produce	Screen	Total	T/s	Juice	Fibre	Lost	T/s	Juice	Fibre	Lost	F
Apples (GS)	Medium	0.444	71	0.145	0.020	0.057	143	0.325	0.058	0.061	12
Apples (GS)	Coarse	0.444	82	0.155	0.012	0.055	132	0.322	0.045	0.077	8

On the significance of juice yields

How important is juice yield?

Produce is not that expensive - is it possible to make up for low yields by simply pressing a little extra produce?

Maybe - but maybe not.

The argument has been made that the easiest juice to extract is of the worst quality - and that the very last bit of the juice that can be pressed from produce is the most laden with nutrients.

Naturally this argument is made my the manufacturers of the most efficient and expensive equipment - and by those who advocate the use of blenders.

It appears to be generally true that certain nutrients tend to be more strongly bound to the fibre - and that discarding the fibre alters the micronutrients which are present in the produce.

Simple-minded models where the produce consists of a variety of structures (e.g. pulp, skin and seeds) which require different thresholds of force to disrupt them would suggest that the last bit of juice extracted from produce may well contain a different set of nutrients from the first bits of juice extracted.

That seeds and skin often contain interesting phytonutrients - and are also harder to crush - supports the idea that the last bit of juice wrung from produce may well be the best quality.

So - though I am not aware of any proper scientific investigation of the issue - my opinion is that the idea that squeezing produce well may a bit be superior to squeezing more produce less efficiently may well have something to it - though I don't feel I am in a position to offer much useful advice about how heavily the issue should be weighted.

Figures cited in "Twin Health" publicity materials

Produce type	Mass	Twin Health	Samson Ultra
Granny Smiths	1kg	780ml	700ml
Celery	1kg	850ml	710ml
Spinach	1kg	730ml	695ml
Beetroot	1kg	670ml	610ml
Red Grapes	1kg	860ml	700ml

The "Twin Health" publicity materials claim:

The above chart is an extract from an extensive juice study conducted in May 2003, to measure the actual output of juice excluding any foam content. The figures shown are the averages of 9 x 1kg test samples of the same produce for each machine, and took over 500kg of produce and 160 hours to complete.

[This test can be performed by anyone, and similar results will be achieved.]

Juicing large volumes of produce may give accurate results - but it glosses over the issue of initial produce loss - and that is not one of the Twin Health machine's stronger points.

I also got relatively good results from the Twin Health machine using red grapes - but grapes seemed to be one of the very few fruit that it could beat the Samson machine on.

Generally, my test results suggest that the Twin Health machine typically produces slightly less juice - except when juicing lots of leafy vegetables.

However, my subjective impression is that the finer screens available can often mean that the resulting output is of a better quality.

Why does the Twin Health machine not produce quite so much juice?

One hypothesis would be that the Samson machine is doing a poorer job of separating the juice and the fibre - and is dumping fibre into the juice.

However, my impression is that this isn't the whole story - the yields were still worse even when (closely comparable in terms of hole size) coarse screens were used on both models.

Other possible factors that might explain the difference include lower pressures, not such close tolerances inside the machine - and perhaps somewhat greater retention of unjuiced material inside the machine at the end of the tests.

If more time had been available, for testing, the hypothesis of lower pressure being responsible for the somewhat-lower yields could have been examined - by using greater pressure restriction with the coarse fruit screen on the Twin Health machine.

Test loads

Photos of the actual produce used to test with:

Wheatgrass test load

rhubarb test load

Purple sprouting broccoli test load