This is only a preview of the March 1992 issue of Silicon Chip. You can view 39 of the 96 pages in the full issue, including the advertisments. For full access, purchase the issue for $10.00 or subscribe for access to the latest issues. Articles in this series:
Articles in this series:
Articles in this series:
Articles in this series:
|
REMOTE CONTROL
BY BOB YOUNG
Voltage losses in speed controllers
Following last month's article introducing the
topic of speed controllers for electric motors, we
present more figures on motor performance and
voltage losses in typical FET circuits.
Having looked at the performance
of some commercial controllers last
month, let us now look at the figures
for an unloaded motor running in both
directions and driving directly from
short leads (no speed controller), to
establish some sort of additional
benchmark-refer to Table 1. The columns in the table are as follow: Vs =
Battery Voltage at the terminals; Vm =
Voltage at motor terminals; Is = Instantaneous start up current in amps;
Ir = Sustained running current in
amps.
Table 1 last month - see page 71.
In Table 1 below, we are losing 0.02
volts with no speed controller in circuit and a lower current. The 2-digit
resolution of the meter, and the fact
that the batteries were losing their
charge as we measured, made accurate measurements very difficult to
achieve but the general thrust of the
problem is there. Does this mean that
Australian automotive wire is at least
twice the resistance of what appeared
to be a lighter American wire? On the
face of things, it certainly seems so.
TABLE 1
Comments
Vs
Vm
Is
lr
12.47
12.45
17.80
2.64
Anti-clockwise direction & timing. No load.
12.44
12.42
19.56
1.92
Clockwise, incorrect rotation & timing. No load.
The interwiring was 75mm of 10A
automotive cable. The small anomalies in the figures are due to the
resolution of the digital meter and the
use of two separate battery packs.
I am at a loss to explain these figures as they are nothing like what I
would have expected. Unfortunately,
I did not have the facilities to measure the unloaded RPM but at least we
have some guide as to current and
voltage drop in the leads, the latter
being 0.02 volts across 0.0075 76 ohms.
Here we have 10-amp automotive
cable giving almost twice .the resistance of the wire and speed controller
in the Novak installation listed in
78
SILICON CHIP
As I said earlier, beware the power of
the milliohm.
To compound the problem, I would
have expected the motor to run at a
higher current when running clockwise than in the correctly timed anticlockwise rotation. This was not the
case and I would be very interested in
retiming this motor at some later date.
Paralleled FETs
To further illustrate the point, Table 2 lists some system losses with
the prototype speed controller fitted.
This table was compiled very early in
the development of the speed controller and shows the effect of paral-
leling FETs to reduce the "ON" resistance of the switching network. If you
will recall, in an earlier column I stated
that the number of FETs to be used
and what type would be decided as a
result of this work.
Incidentally, although it is not a
good idea to parallel bipolar transistors, FETs are just the opposite and it
is quite in order to parallel these devices.
These figures were taken using 10 x
1.8Ah cells, 10 x 4 Masters propeller
and the standard-wind Kyosho 360ST
motor, as tested above. This motor
was timed to run anticlockwise and
was fitted with a 2.5:1 gearbox, the
interwiring again being the poor quality 10A automotive cable. Due to the
fitting of the gearbox, the motor had
to be run in reverse.
Column 1 shows the number ofFETs
fitted to the speed controller. Vd represents the system losses calculated
by subtracting the voltage across the
motor from the battery voltage measured under load. These losses include
wiring harness resistance, voltage
drop across the FETs, contact resistance and meter shunt. The leads in
this case were quite long.
Two things become very obvious
from a quick glance at Table 2. Firstly,
the results are not very good at all.
Secondly, the FETs are the major
source of losses. There is, however, a
law of diminishing returns in regard
to adding additional FETs to reduce
the "ON" resistance. At some point
the cost and weight of the additional
FETs will outweigh the gains.
This is in keeping with Ohm's Law.
The same applies to wire thickness.
Adding copper beyond the optimum
point will add considerable weight
for only very insignificant gains. It
appears that a better approach here is
to use high quality cable, probably
This model shows that electric flight can be applied to quite large models. This
model of an Australian Air Force Caribou has a wing span of 2.4 metres. It was
built a few years ago by David Masterton who also built a model of a B-36
bomber which had six electric engines.
test equipm ent cable or some equally
high quality cable.
Insufficient gate drive
As th e FETs used were IRFZ44s
and these have one of the lowest "ON"
resistances in th e range of readily
available FETs (28mQ), there was obviously a problem somewhere in the
FET drive circuitry. Measurement of
the Gate voltage showed only 7.73V.
It is obvious th at this is not enough
drive voltage to push the FETs into
saturation.
As a result of these tests, I decid<:Jd
to change th e voltage doubler driving
the FET gates to a voltage tripler. This
resulted in an increase in Gate drive
voltage from 7.73V to 12.21V. The results of this are sh own in Table 3. All
other test parameters remain u n changed .
Table 3 shows a substantial increase
in RPM for a reduction in the number
of FETs and FET heating when compared to Table 2. Note also that there
TABLE 2
No.
Vd
l(A)
RPM
1
0.62V
12.6A
7100
2
0.42V
12.4A
7200
3
0.38V
14.0A
7600
4
0.33V
14.2A
7800
is still a substantial loss in the interwiring. In the first item on Table 3, we
have lost 0.63V (46 mQ) across the
wiring which was deliberately made
very light. In line 2 of Table 3, there is
only a 0.15V (llmQ) loss but it still
seems high to my mind.
Mind you, there is a 100A meter
shunt in this circuit and the wiring
harness is longer than that used in a
model , but it does illustrate the point
in regard to wiring. Do not just grab
the nearest piece of wire, even if it
does look thick enough. There is little
pointin spending $12.50 on each FET
if you are going to lose more voltage
across a 5-cent piece of wire.
Alternative FETs
Incidentally, there are a number of
very useful FETs which cost a lot less
than the IRFZ44 and deliver almost
the same p erformance. One in particular which I have been testing gives
an ON resistance of just 23mQ. Six of
these in parallel will give us the required 4mQ ON resistance required to match
the commercial units .
More on these in a future issue.
Comments
There is also no fus eFET very hot
holder in this circuit but
fusi ng is a must, particuFETs hot
larly for aircraft. This is
another potential source
FETs warm
of unwanted milliohms
FETs OK
and fo r this reason the
finished design will feature a PCB
track fuse option, to reduce bulk and
parasitic resistance.
Table 4 gives a final set of figures, to
illustrate problems of a completely
different type. These figures were
taken with a prototype speed control
fitted with two FETs only; this time
using a Leisure 05 Stock motor driving an 8 x 4 propeller on a direct
drive.
In th e figures of Table 4, we see the
usual problems of insufficient FETs
and voltage drop in the interwiring,
resulting in a total of 0.74 volts. However, look at the battery voltage under
load. Wh y is it so low and why is the
voltage drop much higher than usual?
The answer here lies in the much
higher running current of 25 amps.
The direct propeller driver provides a
much greater load and thus the armature RPM is lower than the geared
version , resulting in the higher current. The low battery voltage is an
indication of poor quality batteries.
These were sold to us as high current
1.8Ah batteries. When I cut the pack
open to find what brand and type
they were, I found that they were just
simply labelled "Japan: (no brand, no
type number). Do not forget that these
figures w ill all change in flight as the
propeller unloads.
I have gone into a great deal of
detail in the foregoing material to drive
home the point that in high current
electric motor applications you cannot just grab the n earest piece of wire ,
battery and motor. Great care must be
exercised if you intend to get the best
from your system.
Th e one lesson which stands tall
from th e rest is do not underestimate
the power of the milliohm. As stated
earlier, there is great scope in electrics
for careful system design and the
whole thing can become very absorbing indeed.
However, having outlined the foregoing shortcomings , I must say that ,
all things con sidered, the system
shows a lot of promise. By the time I
have finished with the development
work and installed all six FETs, we
will have a good cheap controller with
performance comparable to some of
the commercial units.
Dangers of electric props
One last point for tyro electric fans
(I love this pun). Do not get casual
about the dangers of using electric
M ARCH 1992
79
~
~
- -
The electronics magazine for the enthusiast
Four good reasons why you should
subscribe to SILICON CHIP
1.
You get a 12.5% discount on the newsstand cover price. Recently, we ha d to increase the
cover price of the magazine because of r ising costs but we have held the subscription to the
old price so you get the benefit.
2
•
You'll never miss an issue. Sometimes it can be very hard t o find SILICON CHIP in the
newsagents because it sells out early or gets buried under other magazines. By taking out a
subscription you don't have to search for it - it arrives in your letter box in mint condition.
We wrap it in tough pla stic to make sur e of that.
3•
Get a discount on t he binder too. You'll want t o store and protect your issues of SILICON
CHIP so when you take out a subscription you get a fur ther discount on the binder. By taking
out a 2-year subscription and buying two binders at the same time, you get even bigger savings.
And we despatch the binder(s) with your first issue so you
have it right fr om t he start.
4 • From time to time, we will have special promotions
and offers in t he magazine, like the Bose Lifestyle compet ition featured in the September to December 1990 issues.
When you are a subscriber you are automatically eligible for
these and you get the chance to benefit.
Interested in these savings and benefits? Just fill in the
coupon on the following page and send it with your
remittance. You'll be glad you did. Why not do it today?
I hope to see you as a supporter soon.
Leo Simpson, Publisher
Just fill in the order form on page 93 ~
80
SILICON CHIP
JUNK MAIL
Most magazines sell their subscriber lists to mail order companies, to earn extra income.
Be assured that we will not do
this. Your privacy will be respected and you will not be deluged
with unwanted mail.
MAGAZINE
BINDERS
Speciafly designed binders for
SILICON CHIP are always available. Made with a distinctive high
quality 2-tone vinyl, you can obtain them at a discount when
you take out or renew your subscription . The price details are
on the coupon overleaf.
TABLE 3
No.
RPM
l(A)
Vd
Protect your valuable issues
Comments
2
0.77V
13.7A
7200
0.14 V across FET S/D pins. Thin Fig.8 speaker
cable used for interwiring.
2
0.3V
14.2A
8000
0.15V across FET S/D pins. 1OA automotive
cable used for interwiring. FETs warm
motors, particularly with propellers.
In more than 30 years of power flying
with internal combustion motors, I
had very few accidents because I
treated those propellers with great respect. However, I did not treat the
electric driven propeller with the same
respect - at first.
In my first week of electric experimenting, I had two nasty accidents.
One involved a gearbox which I damaged by switching on the power without adequately securing the motor. As
a result, the propeller hit a nearby
object and cracked the gearbox housing. This was repaired with superglue
and seemed OK - that is, until the
aluminium strap holding down the
motor began to stretch. At this point,
the propeller, which had not been
balanced, began to shake the motor
and the gearbox disintegrated and the
propeller and drive shaft went hurtling around the workshop. It gave me
a hell of a fright.
The second accident was more serious and occurred when I accidentally connected the positive lead of
the motor directly to the positive battery terminal instead of the positive
speed controller lead. At this time ,
the speed controller was switched off
but the battery terminal was live and
the motor burst into life. I was standing in front of the motor with my
wrist just inside the propeller arc.
The tip of the propeller hit my metal
watch band and skidded off into the
flesh of my wrist, slashing it across
the arteries.
If it had not have hit the watchband
first, I may have ended up in hospital
that day. As it was, I had a bruised
wrist and a cut which drew blood. I
saw a friend of mine do the same
thing with a 60-size IC motor and he
was not so lucky. He did end up in
hospital.
So the rules are simple. Polish off
the flash on plastic propellers, balance them properly, and never work
in line with the propeller or with your
hands inside the prop arc. It's also a
good idea to always switch off the
power to both the battery and the
speed controller. And finally, because
they are only electric motors, do not
get casual. They deliver every bit as
much power as an internal combustion motor and are just as dangerous.
Design features
So here we are at the end of another
column. As a result of the foregoing
work, the design of the completed
controller is now beginning to firm
up. A brief summary of the major
points is now in order.
Firstly, the design calls for a simple
and cheap controller, which will give
forward only speed control. Secondly,
all components are to be readily available. This precludes my favourite
servo amplifier chip, the NE544 which
is now obsolete. Thus, we will be
using standard op amps as the active
elements.
Other features will include a ZkHz
switching rate , a voltage tripler for
operation from a low cell count, provision for up to six FETs, a PCB track
fuse option, and finally a dynamic
braking option. All of this will be in
surface mount technology.·
Thanks to ABC Models at Bexley,
Hobbyworld at Hurstville and Moore
Park Model in Armidale for the help
given in preparing these articles. sc
Silicon Chip Binders
These beautifully-made binders
will protect your copies of S ILICON
CHIP. They feature heavy-board
covers & are made from a
distinctive 2-tone green vinyl. They
hold up to 14 issues & will look great
on your bookshelf.
* High quality
* Hold up to 14 issues
* 80mm internal width
*gold-coloured lettering
logo printed in
on spine &
SILICON CHIP
cover
Price: $A 11.95 plus $3 p&p each
(NZ $6 p&p). Send your order to :
Silicon Chip Publications
PO Box 139
Collaroy Beach 2097
Or fax (02) 979 6503; or ring (02)
979 5644 & quote your credit card
number.
Use this handy form
l
----------Enclosed is my cheque/money order for
$_ _
_
or please debit my
O Bankcard
O Visa
O Mastercard
Card No:
Card Expiry Date __/__
Signature _ _ _ _ __ _ _ _ _
TABLE 4
lr
Name _ _ _ _ _ _ _ _ _ __
Vs
Vm
Is
12.7
12.7
12.0
2.60 No load, no speed controller, 75mm of 1OA auto cable
8.44
7.70
50.0
25.0
Comments
10,800RPM on an 8x4 propellor, with two FETs
Address_ _ _ _ _ _ _ _ __
_ _ _ _ __ _ P/code _
_
_
·-----------·
MARC H 1992
81
|