was Re: Embedded System in Noisy Environment

[nospam]

Guy Macon <http://www.guymacon.com/> wrote:

and scroll down to
the section titled "Industrial washing machine control."

I don't buy that story as a general principle.

Collapsing fields in solenoids do not generate hundreds of volts with a
flyback diode, they generate enough volts keep whatever current they had
flowing.

Assuming an NPN drive when you turn off the solenoid the drive wire
transitions from ground to supply + a diode and that is a voltage spike
about the same size and speed as you got when you turned it on.

On turn off the only rapid changes of current are in the power supply loop
which stops supplying the solenoid and the diode which starts taking the
recirculating current. Putting the diode on the solenoid coil maximises the
power supply loop which is bad.

If the wiring is crappy with separate paths from the power supply and top
of the flyback diode to the top of the solenoid then you are turning off
one loop and turning on another. Putting the diode across the coil may fix
half that problem.

If the thing is wired properly then the flyback diode simply shortens the
size of the loop feeding the solenoid and that loop sees no rapid change of
current. The loop from the power supply to the flyback diode does see a
rapid change of current, hopefully the power supply is nearer the diode
than it is the solenoid out on the machine and you can short it out with
some decoupling capacitance anyway. .

The only other problem may be turning off the solenoid faster than the
flyback diode can turn on. Slowing down the driver or putting a bit of C
across the diode might fix that but it is not a problem I have experienced.

That's how I look at it anyhow, and I have driven lots of industrial
solenoids with microprocessor based systems without problem.



I have changed the thread because I hate advising information leeches from
google. If the OP actually bothers to read usenet and not just the typical
google "tell me how to fix my problem ...thanks" thread he started I'm
sure he will find it.

 

[Guy Macon]

nospam wrote:

Assuming an NPN drive

This was a big hexfet that went high-Z to turn off the solenoid
and a bunch of really big solenoid at the end of long wires
with a common return wire.

I have driven lots of industrial solenoids with microprocessor
based systems without problem.

As have I. You, no doubt, either designed the systems correctly
from the start or bought them from someone who did. The situation
I described was was a specific fix for a specific poorly-designed
system. I suspect that the fellow who started this thread also has
a poorly-designed system of some sort - he certainly isn't able to
say that he is driving industrial solenoids with microprocessor
based systems without any problems.

 

[Pig Bladder]

On Sat, 16 Apr 2005 11:59:12 -0700, Rich wrote:

Hi, this is the INFORMATION LEECH That started the thread.
....
Fortunately, there are many more people giving and receiving advise
from google groups (I recognized some of the ones at the thread for
years).
I'm rewarded when other Information Leeches as me reply my advise
telling they have weeks looking for that answer and solve the problem.

Big Bladder's Corollary to Sturgeon's Law:

Ninety-nine and forty-four one-hundredths percent of everything on the
internet is crap.
--
Love ya!
Pig Bladder on a Stick