NIKON Vibration Reduction (VR) "Keep Alive" Circuit
For TSE 2003, I used a NIKON 80mm - 400mm f/5.6 Vibration Reduction
(VR) lens on a NIKON F5 camera body to photograph
the eclipse from the flight deck of a QANTAS 747-400 ER aircraft over
Antarctica. Using a VR lens really does help reduce image
jitter and wander when observing a "fixed" target on a somewhat
unstable platform. The NIKON VR system, however, has two
significant drawbacks which would otherwise impede its
use in executing a fully automated imaging sequence (such as was done
TSE2003 and now planned for TSE2005).
1. To effectively use the Vibration Reduction feature, VR must be
activated one second before an exposure is taken. This is
done (as I perceive NIKON envisioned how most would normally use a VR
taking pictures via the shutter button on the camera body) by
depressing on the shutter button (what some documentation refers to as
a "soft touch") before pressing down fully to open the shutter.
2. Because the VR servos in the lens suck power from the batteries in
the camera body when running like you wouldn't believe (well, I suppose
would), the firmware logic in the camera will shut down VR appx 830
milliseconds after an exposure in progress completes. I believe
NIKON did this as
a way to "protect" the user from draining the batteries (also used for
winder, autofocus, etc.) too quickly (but not an issue for the duration
Because of these "limitations" (or, as NIKON would say, "features") the
use of VR is fundamentally incompatible with an automated exposure
sequence with the sutter being controlled externally though the
camera's electronic "shutter control line". (Automated eclipse
exposure sequencing by
is commanded through the camera's electronic "shutter control line",
which is pin # 4 on the NIKON's MC-27 10-pin electronic control port).
What one really wants is to enable/activate the VR
function BEFORE totality and have it remain active until it is shut
down by user control - after
the full sequence of externally commanded exposures completes. I
found this can be done with a relatively simple "VR keep alive"
is plugged into the camera body also through the MC-27 multi-pin
using pin #9. Pin #9 is multi-functional, and is used to
(a) external shutter control (pin 4 opens/closes exposures externally,
but must first be enabled via pin 9).
(b) autofocus (if you are using an AF lens and have the AF switched on
on the lens) - likely NOT for an eclipse!
(c) Vibration Reduction (if you are using a VR lens and have VR
switched on on the lens).
Basically, closing the pin 9 contact (returning it to ground) enables
all three functions, just as touching lightly on the shutter button
Hence, to enable VR one needs to close that contact (pin 9 to ground).
BUT, after an exposure is taken the camera will
autonomously disable VR even if the contact is still made. VR
must then be "re-enabled", which would be like "tapping" on the
button to turn it back on. Actually, if you tap "fast enough" -
is faster then the time-out after exposure completion that the camera
uses to shut down VR - it will remain up. So, I built a simple
to do the "tapping" for me - through pin 9 on the 10-pin camera control
Basically, it sends a series of "enable" pulses to the camera,
a bit faster than the 830 ms timeout interval. I found,
that if such a pulse is shorter than appx 17 milliseconds the camera
ignore it. Likely the firmware uses some sort of "debounce"
and (a)-(c) above won't be enabled if you "accidentally" did a very
soft-touch on the shutter button. I did find that with a pulse
the pulses would be robustly recognized by the camera. Hence the
alive" circuit, send a series of 23ms wide pulses (which close
pin-9 to ground connection) every 820 milliseconds (robustly faster
the timeout interval, also by experimentation).
Why not make the 23ms pulses longer? There is another
firmware interaction - which I haven't figured out how to beat (ideas
are welcome!). When you started a bulb mode exposure (shutter
open via pin 4), then send a VR enable (pin 9 closure) pulse, and then
you try to close the shutter (via pin 4) - the shutter WILL NOT
CLOSE until you open the contact on pin #9. I can only guess that
NIKON did this assuming that by
pressing "hard" on the shutter button to take a bulb exposure (shutter
remains open while depressed) you are not simultaneously "asking" to
VR as you would have done that before opening the shutter by a "soft
This makes sense IF you are operating the camera by its shutter
- but not via the electronic interface for eclipse photography.
for best compatibility with UMBRAPHILE, for example, I make the pulses
as short as possible. What this does mean is that there is about
3% chance (roughly one exposure out of 36) with an exposure sequence
with the keep alive timing, that the keep alive pulses will delay a
closing (and a planned exposure would be longer than desired, by the
of that 23 milliseconds). For now, I live with that...
So... here is the circuit (from my notebook):
Just a few notes.
1. The 7805 is a 5 volt DC voltage regulator, so you can power this
from a 9V transistor batter.
2. The NE556 is a timer chip, used here a mono-stable oscillator.
Actually the NE556 is a dual (two timer) package, you can use an
NE 555 (I just
happened to have a 556 when I built this). For more information
pretty complete on-line tutorial) on the NE555/556 see: http://www.uoguelph.ca/~antoon/gadgets/555/555.html
If you use a 555 rather than a 556, the pins are different - but
the functionality identical.
3. The W171 is a 5-volt fast acting relay (the DIP-25 just refers to
it being in a "dual in-line package"). The one I used was made by
Magnicraft. You can use a solid state (5V) relay or a reed replay
- basically anything that can be powered at 5V and provide an
open/close that is fast enough.
4. The outputs of the relay go to pin 9 and ground of the NIKON MC-27
connector. The pinouts on that connector are shown here: http://www.perfect4me.com/Products/mynikon/default.asp
The timer cadence and duty cycle (pulse width) are controlled by
adjusting resistors Ra and Rb and capacitor C (see above NE555
tutorial). The values given will result in the timing discussed.
This has worked with two different NIKON VR 400mm lenses - BUT
you may need to tweak that with a different lens - as I do not know the
robustness of the timing in the lens itself from lens to lens.
Also, I just used off the shelf components (10%-ish values) and
you too may want to check what you get
on an O-scope. If VR won't come on, you probably need to increase
the pulse width a bit. If VR comes on momentarily and then goes
you probably need to decrease the inter-pulse cadence a bit.
That's it! Hand's off VR "keep alive" for solar eclipse
That said. Here are two other advantages.
1. Carter Roberts had written to the Solar Eclipse mailing List: "One
thing to consider with the lenses with built in stabilizers is
that when the stabilizer is turned on the image usually drifts
at least on older designs of lenses, the manufacturer typically
to turn the stabilizer off when using a tripod. Since the
stabilizer goes off every time you stop to change shutter speed
there is the
potential to have smeared images. This strongly suggests taking a
bunch of shots at each setting (and hopefully use auto-exposure
to make that three different exposures) so even if the first shot
or two gets smeared the others will be sharp".
If you build this circuit and plug it in you don't have to worry about
2. To use external "shutter control" (as does UMBRAPHILE) on the F5
(not sure about other NIKON cameras) you MUST enable shutter control
is a shared enable signal on pin 9 along with VR enable) no longer than
30 seconds before your first use of the shutter control. And,
times out 30 seconds after an exposure. If you "turn on" the "VR
alive" circuit before totality, you also get shutter control "enable"
28 March 2005