The all important parameters for our applications are xx Rds(on) @ xx Vgs.
The Rds on is the effective source to drain series resistance at a given
source to gate voltage Vgs. Note that beam circuit output voltage that
would be applied as Vgs are generally in the range of 3V-5V.

According to the current Digikey catalog guide (p299-p300), for the Zetex
TO92 packaged 2N7000P-ND Rds = 5 ohm @ Vgs=10V and for the ZVN2106A-ND Rds =
2 ohm @ Vgs = 10V. Those are not particularly good specs for Beam circuits
and a 74ACxxx (advanced complimentary mosfets) device can give better
performance at low voltage.

A better choice at the same price of $1.02 would be the Zetex ZVN4206A-ND
Rds = 1.5 ohm @ Vgs = 5V or at $2.10, the ZVN4306A-ND looks good with Rds =
0.45 ohm at 5V. There is a TO92 P-channel mosfet ZVP4105A-DS with Rds = 10
ohm @ Vgs = 5V. I am not aware of any!! TO92 complimentary mosfets that have
excelent low voltage like those readily available in tiny SMD packages For
example, look at the two complementary devices in one 8-MSOP package
ZXMD63C02XCT-ND with Rds=0.2ohm(N) and 0.4ohm(P) @ Vgs = 2.7V and it costs
only $2.28! Now all we would need is two of these dual transistors and two
resistors (shown per attached), mounted on a small carrier board with 0.1
inch terminations and you are looking at a nearly ideal 4 transistor
non-smoking h-bridge for our beam applications.

Regarding the second question, power is an important consideration but
threshold symmetry, frequency stability and spurious oscilations are other
parameters. Some 74ACxxx devices have thresholds that are almost perfectly
at Vcc/2 over a range of Vcc. That means that time constants will be less
sesnitive to Vcc variations. 74ACT/HCTxxx has relatively fixed threholds
offset from Vcc/2 at Vcc = 5V. Note however that the threshold will be
symmetrical near between 2.5-3V. The upshot of this is that for 74HCTxxx
bicores/ucores the time constant increases when the Vcc drops from 6V to 4V
which can help compensate for the slower motor speed as the battery voltage
drops, thereby maintaining a slower but equal length stride. Finally, the
74ACxxx devices require fast input transistions, short supply leads and good
Vcc filtering in order to avoid spurious output oscillations. I suspect that
many beam designs based on 74ACxxx oscillators violate FCC rules by
radiating RF in the FM band something that can be easily verified by
scanning the band with a portable FM radio. In addition, these high
frequencies are conducted via common supply conductors through out the Beam
circuit and account for the tendency of multiple 74ACxxx bicores to
synchronize which is only good if you need that.

wilf













wilf


-----Original Message-----
From: Jean auBois
To: beam@yahoogroups.com
Sent: 5/20/01 4:39 AM
Subject: [beam] couple-o-questions

First: I notice that Solarbotics sells the ZVN2106 FET. From the name,
one
can infer that it is an n-channel device. Another device, the ZVP2106
also
exists which is, naturally enough, a p-channel device. Not only does
Solarbotics not carry it, but when I went looking elsewhere
(www.future-active.com, to be precise) I found that they require that
you
call them to get a quote on price/availability.

So my question is: how can the ZVN2106 be used as such a
general-purpose
device? For example, I haven't seen a bridge yet that doesn't require
both
PNP and NPN transistors...


Second question: so HCT parts use less power and can sink/source less
current than AC parts, however, the latter are obviously preferred in
the
Solarbotics catalog. Presume you've got an infinite power source so
that
doesn't enter into your decision: when making Nv neurons and both kinds
of
bicores and all the rest of this stuff, of -all- of the possible chip
technologies, which has the best overall characteristics? Which
characteristics make the difference?


Thanks bunches,

JaB


Your use of Yahoo! Groups is subject to
http://docs.yahoo.com/info/terms/