[Previous Months][Date Index][Thread Index][Join - Register][Login]
[Message Prev][Message Next][Thread Prev][Thread Next]
[IP] Re: Feedback instabilities
>John, I thought this was getting to technical for the list. BUT damping
>or delays are always bad for feedback systems. They allow the systems to
>get closer to instability. Whenever you have such in system, the design
>goal is to try to compensate for this, often with some derivative control.
>I haven't done much system design since college, but that one of the golden
>rules was to avoid delays and integrals whereever possible. The damping of
>the insulin into the blood with NPH or Lente is what causes most of the
>instability problems without the pump. When we do have that long promised
>noninvasive BG meter, we will see just how bad the damping of the insulin
>input is, even with H in the pump. Then maybe your suggestion of an iv
>injection to eliminate this damping would make sense!
Ah! A nice email to brighten up my afternoon...
I'll leave you to ponder your own integrals and derivatives :-)
Personally, I think that insulin delivery through subcutaneous fat must
soon go. It's so problematic. With extreme body-piercing now being in high
fashion, insulin infusion points going straight to the liver will soon be
the rage. Much better than iv. Perhaps with 5% NN304 mixed in with Regular,
you will have some DKA-protection in the event of pump disconnection. The
implantable pump seems to be fraught with practical problems, and may be
some way off.
I'm not so interested in non-invasive bg meters. Diabetes treatment _is_
invasive. Far more interesting are the continuous meters. Now that Minimed
have a continuous (if invasive) bg monitor, when do you reckon they'll put
the whole shebang together, and give us a hands-free automated control? Is
it just a question of time and money and a bit of software?
All the best -
mailto:email @ redacted
Insulin Pumpers website http://www.insulin-pumpers.org/
for mail subscription assistance, contact: HELP@insulin-pumpers.org