Homemade High Power Diode Laser

koliaco · Post by **koliaco** » Thu Feb 12, 2009 9:48 am

a_k wrote:MAX1978 is also a better choice because it needs fewer external parts. With MAX1968 you would need several additional precision opamps.

Unfortunately MAX1978 comes in a tiny QFN48 7x7mm package. MAX1968 is in a larger TSSOP package.

I got the solution for this already, just purchased the Schmartboard few days ago to mount the MAX1978, there is a video on youtube showing how to solder it. http://www.youtube.com/watch?v=qaMPp_5t ... 1&index=60

wler · Post by **wler** » Thu Feb 12, 2009 1:39 pm

koliaco wrote:Have anyone in here used the TEC controller MAX1978?

I have read the post "Wavelength stability requirements" about using the MAX1968, but since its accuracy can only be up to 0.01C compared to 0.001C of MAX1978 so I guess MAX1978 would be a better choice?

Cheers

The stability depends mainly on the frontend opamp and you can do well with the MAX1968. The 78 is too hard to handle due to the package. But you need to be very careful with the design of the circuit, otherwise you won't get the stability. See my web page for PCBs and detailed analysis.

a_k · Post by **a_k** » Thu Feb 12, 2009 9:16 pm

The package in the video is a TQFP with visible pins. MAX1978 comes in a TQFN package without pins, just pads at the bottom side which are not visible from above or from the sides of the chip. The problem with the TQFN stuff is, besides the small footprint, that there is no way of visually inspecting the result in detail.

I'm not sure how easy it will be to solder a TQFN component pad by pad with a soldering iron. Usually it is done by reflowing and the chip aligns itself.

koliaco · Post by **koliaco** » Sun Feb 22, 2009 9:39 am

a_k wrote:The package in the video is a TQFP with visible pins. MAX1978 comes in a TQFN package without pins, just pads at the bottom side which are not visible from above or from the sides of the chip. The problem with the TQFN stuff is, besides the small footprint, that there is no way of visually inspecting the result in detail.

I'm not sure how easy it will be to solder a TQFN component pad by pad with a soldering iron. Usually it is done by reflowing and the chip aligns itself.

I have soldered this thing now, it took me the whole weekend, hopefully everything is gonna work well.

With inspections regarding connections- all working well, no short circuiting, I just adapt my Multimeter probe to have clips to hold two needles to test the pins. You can actually see the pins of MAX1978 from sides. the only thing is that they are only about 0.5mm high.

Colin Kaminski · Post by **Colin Kaminski** » Sun Feb 22, 2009 4:45 pm

Wow, those are big caps. What is the voltage rating on them?

wler · Post by **wler** » Sun Feb 22, 2009 4:56 pm

Actually it is important to use high quality film caps and not electrolytic or tantals, for stability. Those are generelly quite a lot larger than the latter, but also bipolar which is needed here. At any rate one needs to use high-quality components to achieve good stability. That applies in particular to the diode driver as well.

Incidentally there is a thread in a german forum about TEC controllers and one guy is doing PCB's for the MAX1968, see here. Though stability is not a major concern for those guys who use lasers as kind of lamps only

Hans · Post by **Hans** » Sun Feb 22, 2009 7:54 pm

How did you solder it. Could you touch the pins with a very thin soldering iron tip?

How did you connect the fooling fins on the back of the board?

Solarenemy · Post by **Solarenemy** » Mon Feb 23, 2009 5:48 am

koliaco wrote:
a_k wrote:The package in the video is a TQFP with visible pins. MAX1978 comes in a TQFN package without pins, just pads at the bottom side which are not visible from above or from the sides of the chip. The problem with the TQFN stuff is, besides the small footprint, that there is no way of visually inspecting the result in detail.

I'm not sure how easy it will be to solder a TQFN component pad by pad with a soldering iron. Usually it is done by reflowing and the chip aligns itself.
I have soldered this thing now, it took me the whole weekend, hopefully everything is gonna work well.

With inspections regarding connections- all working well, no short circuiting, I just adapt my Multimeter probe to have clips to hold two needles to test the pins. You can actually see the pins of MAX1978 from sides. the only thing is that they are only about 0.5mm high.

That MAX1978 comes in what we call a castellated package in the industry. The connections are made on the bottom and small indentations on the side of the device. It is very hard to do without a special purpose iron and tips made by Metcal. Also you have to heat the board and the device evenly so solder will flow under it to the pads. otherwise the device acts as a heatsink. The solder will draw up to it but will not go under and or adhere to it's pads. It will look connected to a bare eye and even good magnifying glass. But with a powerful magnifier you will see a gap between the device and the solder. Testing the connections the way you are is only testing for continuity from the solder pad that would be the long silver part you are touching to the point the other end of your meter is attached to. You are not verifying if the solder connection to the device is good or not. The only way to test this is through a computer aided scan test that can scan the inputs and outputs and verify they are all connected or a boundary scan will also verify connection by reading impedance of the device and board but again you need a computer aided test system. The only way your going to be able to verify for sure the connections are good since you chose this device is to power it up and simulate or place it in a real condition and verify it works. Of course you also stand a chance or damaging it as well. But for this type of device package it is the only way to test it without expensive equipment.

koliaco · Post by **koliaco** » Mon Feb 23, 2009 11:45 pm

Hans wrote:The only way to test this is through a computer aided scan test that can scan the inputs and outputs and verify they are all connected or a boundary scan will also verify connection by reading impedance of the device

When I soldered this thing, the solder iron did not run to all the pins, and they are still of golden-yellowish-copper color. I used the needle to only touch that part of the pin which has no solder. Also when my circuit is complete, I could verify that by having connection running through different pins (eg. two separate resistors connected to two internally connected pins are then connected).

I actually used my cheap 40w soldering iron with 6mm dia tip - just fined it to very small point at tip and tinned with solder.

koliaco · Post by **koliaco** » Mon Feb 23, 2009 11:56 pm

Hans wrote:How did you solder it. Could you touch the pins with a very thin soldering iron tip?

How did you connect the fooling fins on the back of the board?

No i could not touch the pin, the solder on the pcb just melted and I pushed it towards the pins as in video in youtube, and it worked. I know its not the best sort of connection, but connected.

The cooling fins I got from local auction, they are 1.5x2x0.5cm chipset heatsinks and have sticky pad (but I removed it). under pcb where the chip sits there is a 5mm square filled with a bit of solder and it has a ~1.5mm dia hole to fill in with solder, thus connecting it to the MAX1978 on top. I just put my cooling fins on top of that square and filled it with solder again.