On Jan 11, 5:52*am, wrote:
On 11 Jan 2009 07:03:37 GMT, "DoN. Nichols"





wrote:
On 2009-01-11, willray wrote:

Greetings all,

I'm working on balancing my rotary converter and need
a sanity check.

(Basic statistics for the record. *1125RPM 15HP idler.
*Feeds off a 125A breaker in the shop panel, through

* *[ ... ]

Just for gits and shiggles, rather than just tweaking
the balancing caps with an amp-clamp and DVM, I thought
I'd build a full-blown current and voltage monitoring
system for it.

For current monitoring, I've stuffed 3x 100:5 current
transformers into the 3P subpanel, and brought them out
to current-sensing resistors.

* *Good -- that gives you isolation.

For voltage monitoring I need voltage in the range -1V -- 1V
to match my scope inputs, so I need to build a voltage-divider.
This would be no big deal, except I just tossed myself a puzzler
when getting ready to solder the bits together.

One could build 3 bridges (say 2MOhm and 2KOhm roughly,
just to make the numbers reasonably neat), and bridge
L1-L2, L2-L3, and L3-L1, thereby being able to measure
the between-leg voltages. ?Alternatively one could bridge
L1-C, L2-C, L3-C (and possibly tie C to ground/N). *I've
doodled an ASCII version of this below. *Hopefully it survives
being posted. *L1, L2, and L3 are as traditional. *L1-a is 2MOhms,
a-d is 2KOhms. *The other legs are symmetric.

* L1 * * * * * * * * * L2
* *. * * * * * * * * * .
* * _ * * * * * * * * _
* * *| * * * * * * * |
* * * _ * * * * * * _
* * * *| * * * * * |
* * * * _ * * * * _
* * * * *. a * b .
* * * * *| * * * |
* * * * * _ * * _
* * * * * *| * |
* * * * * * _ _
* * * * * * *. d
* * * * * * *\
* * * * * * */
* * * * * * *\
* * * * * * *. c
* * * * * * *\
* * * * * * */
* * * * * * *\
* * * * * * */
* * * * * * *\
* * * * * * */
* * * * * * *.
* * * * * * *L3

This setup appears allows metering of both the between leg
voltages by metering across both middle .1% resistors on
a leg pair ( ie, a to b, b to c, c to a) and the
Leg-Neutral voltages if the d is tied to ground/N.

I hadn't originally thought of using the "Wye connected" meter
setup, but I can't currently think of a downside. ?I'm
wondering if anyone who does this stuff more regularly might
know some reason not to do this, that I haven't thought of.

* *Yes -- one very good reason -- have you tried measuring the
voltage from ground to the center point? *I'll bet that it is a nasty
range above ground, since you have no isolation between the incoming
power line and the rotary converter.

* *Get three identical filament transformers with input voltage
appropriate to the leg-to-leg voltage. *You can then tie one side of
each secondary to ground and safely measure voltages (taking into
account the ratio of the transformer, of course.

Hmm - while I'm at it, another question -- Is there any reason
that I can't tie together one leg of each current-sensing
resistor for the current monitor part, and use this as a
"common", for metering purposes?

* *Yes -- since the current transformers give you isolation.

* * * * * * * * * * * * * * * * * I'm suspicious that my
scope inputs don't have isolated grounds.

* *Most don't. *The probe ground is connected to the safety ground
pin on the power cord.

* *You want the isolation that the current transformers give you,
and similar isolation on the voltage side with the filament
transformers. *(Those used to be cheap enough and common enough so you
should be able to pick up three identical ones.

* *Good Luck,
* * * * * *DoN.

* *I see you're aiming for the luxury of both current and
*voltage measurement.

*When there is significant unbalance, the apparent 3 phase
balance of the current set and the voltage set can be very
different.

* * With both sets of figures available you may be tempted *to
aim for a rough average betweeen voltage and current balance.

* * * * Because motors are essentially current driven devices
this is not the way. Balanced current should be your aim even if
this results in larger voltage unbalance.

* * * * * * * * * * *Jim

What do you think of displaying the current transformer output on the
scope to watch the phase angle? A vectorscope would be really nice but
I don't know how to show 3 phases on an X-Y display.

jw