modification for avoiding temperature oscillations: see at bottom
modification for driving blue laser diodes: see at bottom as well
I got hold of one of these classic controllers. At some point I'll measure
long-term stability and other factors that would be important for holography.
Here are some remarks:
The unit didn't at first seem to work - this was due to an open interlock,
for which there is a little plug at the backside. Needless to say that first
thing to do was to short this for good.
By changing the current sense resistor, one can modify the current range.
I read somewhere that one can extend it to 1.6A max or so. The resistor in
question is shown in the figure below. Normally (for 1A max) it is 1Ohm, 1%,
5W. I found my unit to be equipped by a 10Ohm resistor instead, limiting the
max current to 100mA. After reinstating the 1Ohm resistor, I can go up to
1A, however the display doesn't show it right - it displays 100.0mA when the
current really is1A. Obviously this had been modified by the previous owner,
but I couldn't figure out what else was done - no further resistors anywhere
of the board seem to have been changed, it seems an issue of programming the
microprocessor; indeed, all what needs to be changed is the decimal point,
and from tracing the circuit, that seems to be controlled directly by the
processor.
What is very preculiar, is that there is no soft-start feature!.
Insted the current is switched on (after a delay) by a relay, and so goes
up in a microsecond or so. I was skeptical and checked the current carefully
with a scope: indeed, there is a ca 30mA overshoot on the diode upon turn-on.
This can destroy a low current diode instantaneously! Below a scope trace
for a 30mA setting, there is overshoot to approx 90mA; the current limit was
set to 40mA. For a 100mA setting, I found overshoot to ca 130mA, and for an1A
setting, the overshoot is still in the order of 30mA. So don't use
this driver if you have a diode that cannot take more than 30mA!
It is a good idea to have a small series resistor (1-10Ohm) close to the diode
and a cap of 100nF or so parallel to the laser diode. And a Schottky diode
anti-parallel to the
LD plus a 100Ohm resistor
are a good investment too.
Shown here is a current overshoot to over 90mA for a laser diode setting
of 30mA.
This was measured across a low-induction 1Ohm series resistor near a dummy
diode.
The temperature stability is supposedly 0.1C, which corresponds to a wavelength
shift of approx 0.03nm. This may not be good enough for holography applications. But it is not specified under what conditions, so some measurements are in order.
The TEC controller is uni-directional and not bi-directional - that is, it
can cool only; the idea is, of course, that the cooling is balanced by the
heating from the diode. This can be non-optimal for some applications,
where fast response time is required.
I have tried this controller with TEC's in various different laser heads, and in every single case the temperature oscillated - apparently the feedback loop is designed to work with specific SDL laser heads. Thus, the TEC controller as-is appears not be useful for stabilizing other laser heads, and in particular is not suitable for holography applications.
I now changed the feedback circuit so that it is now stable with all my laser diode mounts. The time constant for the integrator was too short, but the following simple operation remedies this. See this part of the circuit diagram:
Simply replace R74 by 1Mohm und C33 by 4.7uF (or more; preferably MKS2 film capacitor or similar). They are easy to find as the PCB shows the component labels, they are at the center of the PCB somewhat to the front (see the blue circle in the picture above). The little modification has the following effect:
To the left: temperature swing due to oscillation in original setup. To the right: Damped response after modificaton.
Clearly now the nice gadget can be resurrected from the shelf and actually be used for something ;-)
The short-term stability seems sufficient for holography applications (x-axis is scaled by ten):
Further modification: The SDL-800 has a lock-out mechanism if the diode becomes unconnected. It triggers at approx 3.5V, which means that blue diodes such as the omnipresent blu-rays or the stunning 1W Nichia diodes, which have a forward voltage in excess of 4V, cannot be driven. This called for an improvement. It turns out that just one resistor needs to be changed, namely R23 needs to be changed from 68k to 100k (this refers to the schematics in the manual, it is there pretty much to the right). On the PCB it is located roughly at the center somewhat to the right. After the change, the cutoff voltage is increased to about 5.5V and the Nichia can easily be driven up to 1A or so; watch out whether your TEC can handle the heat load of more than 4W, as the SDL800 can yield only 1.5ATEC current. It may be necessary to use a TEC with higher resistance or several TEC's in series.
modification for avoiding temperature oscillations: see at bottom
