I was suggesting you try more mass on the cold side, but looking at your drawing again and seeing your report of drift caused by holding your hand near the heatsink, I think drastically increasing the heatsink mass would be better. It's a nice miniature design, but I can see how the multiple thin fins would efficiently couple changes in the ambient air temperature and more importantly air currents to the Peltier hot side, thus changing its operating point to which the control system would then have to react. I have in mind one of those big heavy cast aluminum multifin heatsinks often seen on high power audio amps. Or just a big heavy chunk of metal. Anyway, it's something to try when your reserves of patience have been replenished
Diode Stabilization
-
Jonathan
Diode Stabilization
You are so close to making the TEC work, since mode stability and high contrast in the fringes is not a problem when the PID is tuned. To get a handle on the drift, I agree that more mass in the system is needed, on both sides, as well as maximum protection of the heat sink from all external influences. That's clear from what happens when you put your hand 2" from the fins, or add cold drops of water, etc.
Keeping air currents away is necessary. An enclosure of some sort may be enough, as you're planning to build. The closer it is to being a closed system, the better. But for this to work the hot side needs sufficient mass so that overheating isn't a problem. Also, more mass on the cold side will make for less vulnerability of the diode to external heating influences which may arise from the current associated with the thermistor and/or the TEC.
Keeping air currents away is necessary. An enclosure of some sort may be enough, as you're planning to build. The closer it is to being a closed system, the better. But for this to work the hot side needs sufficient mass so that overheating isn't a problem. Also, more mass on the cold side will make for less vulnerability of the diode to external heating influences which may arise from the current associated with the thermistor and/or the TEC.
-
Joe Farina
Diode Stabilization
I suggest using a small cold plate and a large heat sink. A smaller cold plate will have a faster response to the cold side of the TEC. I would also suggest insulating the cold plate to isolate it from the surrounding air. As to the size of the heat sink, it would be far better to make it too large rather than too small.
-
Colin Kaminski
Diode Stabilization
Since everyone was so helpful I decided to dust off the TEC and give it another try. I made a heatsink that is simply a 1/4" aluminum plate 4"x4". I has almost the same surface area as the old heatsink but is much heavier and the emissive area is farther away from the heat load. This made it much more stable to touching the heatsink and just generally being near the TEC. I only had 30 minutes to rough tune the PID but it is already better. I should get more time to test it later this week.
Thank you, Jonathan, Tom and Joe for your help.
Thank you, Jonathan, Tom and Joe for your help.
-
Greg Garriss
Diode Stabilization
Howdy folks,
I just found this forum and have been reading through the interesting postings. Please excuse me if I'm going over old ground. Most techniques used to stabilize laser diodes use the TEC to keep the diode at a fixed temperature ( selected so that by adjusting the diode current you can avoid mode hop regions. Really expensive low power diodes ( telecomm, etc ) actually come with individual characterization "curves" so you have a good idea where to start.) This doesn't necessarily mean cooling the diode. I've been involved in several measurement systems that had the best long term performance ( weeks ) when the diode was kept above room temperature. A "best practices" technique for diode laser stability involves "embedding" the diode in a relatively large thermal mass and then insulating the package with just the back side of the TEC available to attach to a huge, nearly "infinite" heatsink ( like the optics bench ). The TEC doesn't have to work very hard ( doesn't generate much heat )and by having a massive heatsink on the other side of the TEC, you have a place to pump heat to/from that doesn't vary much( the mass integrates rapid changes in the thermal environment ). The sensored mass wrapped around the laser does roughly the same thing as the big mass ( on a much smaller scale )which helps stabilize the TEC controller's PID loop. Heatsinks with fins are a bad idea from a laser stability view since they're very dependent on air movement, making the system much more sensitive to environment changes. Hope this is useful info.
I just found this forum and have been reading through the interesting postings. Please excuse me if I'm going over old ground. Most techniques used to stabilize laser diodes use the TEC to keep the diode at a fixed temperature ( selected so that by adjusting the diode current you can avoid mode hop regions. Really expensive low power diodes ( telecomm, etc ) actually come with individual characterization "curves" so you have a good idea where to start.) This doesn't necessarily mean cooling the diode. I've been involved in several measurement systems that had the best long term performance ( weeks ) when the diode was kept above room temperature. A "best practices" technique for diode laser stability involves "embedding" the diode in a relatively large thermal mass and then insulating the package with just the back side of the TEC available to attach to a huge, nearly "infinite" heatsink ( like the optics bench ). The TEC doesn't have to work very hard ( doesn't generate much heat )and by having a massive heatsink on the other side of the TEC, you have a place to pump heat to/from that doesn't vary much( the mass integrates rapid changes in the thermal environment ). The sensored mass wrapped around the laser does roughly the same thing as the big mass ( on a much smaller scale )which helps stabilize the TEC controller's PID loop. Heatsinks with fins are a bad idea from a laser stability view since they're very dependent on air movement, making the system much more sensitive to environment changes. Hope this is useful info.