CFL's and oscillating inverter

[Scott Willing]

I helped a friend upgrade an extremely modest cabin system, and while
I was at it, threw in a refurbished Samlex 600W isolated output MSW
inverter to replace a non-isolated Statpower 250W.

The inverter primarily runs a string of four CFL's. One bulb is OK,
but when a second bulb is turned on, the additional (non-linear,
highly capacitive of course) load is seen by the Samlex as an
overload. It limits the current, tries again, limits the current,
tries again et cetera ad infinitum. Net result is a very lousy strobe
effect.

Not surprisingly, throwing a single incandescent bulb (or just about
any other load) into the mix swamps the effect of the ugly CF bulb
loads and we can turn 'em all on no problems.

We could simply switch back to the other inverter but the idea was
that I wanted my friend to have enough inverter juice to run a few
small kitchen appliances. Annoying to put in a larger inverter only to
have a problem running 26W of light.

I'm looking for a cheap and simple fudge that does not compromise
safety or use significantly more power.

As in the past, Samlex was very helpful and forthcoming with my
original inquiry, though all they've done so far is confirm my
suspicions. I'm awaiting a response to this follow-up question that I
thought I'd post here as well:

Mightn't we resolve this with a little series inductance, and if so,
what would a good experimental value be? I don't have the junk
collection that I used to, so I'd probably have to order something. I
might have something with a PI RFI filter on the AC input...

BTW, the bulbs in use at present are 13W Sylvanias. We're going to go
over their next time with a bunch of our no-name generic brand CFL's
to see if we get lucky with a different ballast design. This is very
much "on the edge" e.g. just the manufacturing variables between the
different Sylvanias is enough to cause inverter cycling or not
depending on the order you turn them on.

(Wiring's good BTW.)

-=s

 

[Danno]

Here's my first guess.

Get a kill-a-watt or other type watt meter. Read the power in watts and
the power factor. This can give an estimate of the level of capacitive

Rory

My Kill-A-Watt meter doesn't work on my MSW inverter, but it does work on
utility power and a SW inverter.

 

[Scott Willing]

I have one CFL in my (grid powered) house. My Kill-a-watt says 23 watts
and power factor of 0.59

I am not sure that this power factor carries across CFL's in general.

The irony here is that my real problem is the surge from charging the
capacitor in the CFL that -- at least according to one source -- they
added to the circuit to try to improve the power factor.

MSW driving non-linear switch-mode converter... I'm not sure that
power factor per se -- i.e. a steady-state issue -- is really the
problem. It's more of a transient behavior issue.

I'm also highly skeptical that a meter like the Kill-a-watt could make
sense of non-sinusoidal input either.

La la laaa.... still haven't gotten over there with a pile of
alternate brand CFL's to try that, but I'll bet it will be different.
Like maybe she'll be able to turn on two bulbs and the third one will
start the light show. We'll see.

Thanks for your thoughts lads.

-=s