This is consistent with the problem being saturation of the L3 inductor in INSTEON dimmers as I noted earlier and with the unofficially proposed fix by SmartHome to add series inductance. (I measured L3 inductance as18uH at minimal load and 120 and 1000hz ).
In other words, a simple fix for a common problem to many wall-mounted dimmers from many manufacturers over the years.
Typical wall- mounted dimmers for residential use have space constraints that typically preclude the large chokes found in architectural and theatrical lighting panels.
By designing a physically small dimmer, INSTEON made their dimmers useable in widely-used small switch boxes that made installation of other, larger dimmer
-- including many/most X-10 models -- impractical. This is/was a good initial trade-off in my opinion -- especially if they promptly identify the problem and supply the fix for the (apparently) small minority of cases where this (apparently minor?) problem arises.
Someone that actually has the flickering problem could broadly assess the proposed fix by adding a pair of inductors (in parallel with each other for adequate current capacity and to minimize likelihood of saturation) from an X-10 WS-467 (I measured ~44uH each) in series with the red load lead of an INSTEON dimmer to raise the total nominal inductance to ~40 uH.
If that solves the problem, a permanent solution that many local authorities would likely consider compliant with the National Electrical Code would be to install a RU- (UL component) rated toroid core with conventional 14awg insulated wire looped through it. The number of turns needed for a given inductance depends in large part on the composition of the core.
One can calculate the resulting inductance for various toroid compositions from the data at rfcafe.com, eg:
There are numerous sources of toroids on the primary and surplus market.
I have a large supply of red+brown T-36 toroids from
... Marc Marc_F_Hult