March 23, 2016
Go to all tests
Repeat of 1627, all birch.
Sixth of a six matched load series to investigate PM repeatability with cordwood
Log cabin stacking, side ignition,
63.0 lbs including 3.0 lbs kindling
Baffle supports opened up. See MHA-1625Original test + 6th of 6 matched loads.
Air slots in firebox unblocked, except for top course (25% restriction)
See MHA-1623 for other modifications.
Coefficient of Variation (CV, or COV) is the standard deviation divided by the average.
25% COV means that on average a PM test number is accurate within +/- 25%
Cumulative coefficient of variation. After the first 3 tests, it looked like the COV was very tight, around 3%.
However, with more tests it settled to its realistic value, around 25%.
There are a few spare test slots left, and we will add some more runs to confirm the 25% value.
Compare with previous Lopez Labs crib repeatability testing, which yielded a COV of 10.2% with 4 sets of matched cribs (8 test runs)
in the crib testing that the CO repeatability was much tigher, within
1.5%, although the CO value itself was about 8X higher than the
In both cases, the efficiency was extremely tight.
Recently, MHA ran the Condar against Method M5G-3 on 3 pellets stoves (18 tests total) and achieved a COV of 3.1%
Compare with crib reproducibility (which is not the same as repeatability with the same operator within the same lab)
among different EPA accredited laboratories in round robin proficiency testing, running the official EPA M-28 test method
on the same stove, shipped to different laboratories, all using the M-28 specified Douglas Fir cribs at 4 different burn rates.
In an 18 year period, there were a total of 5 different stoves used.
(table is excerpted from HPBA's NSPS submission)
some of the laboratories note the large variations in different
years, running exactly the same test on exactly the same stove.
Note the super dirty non-cat, in 1989, with very good reproducibility
Raw Testo Data
Gas graph. CO low was 52 ppm
Testo gas graph.
This page was updated on March 28, 2016
This page was created on March 23, 2016