Exploring Thorlabs MTDEVAL TEC Driver

In a [»] previous post, I have shown how thermoelectric cooling could be applied to a cheap Thorlabs CPS532 laser diode to tune its mode and reduce its overall bandwidth to increase the resolution of the [∞] OpenRAMAN spectrometer.

During that evaluation, I was using an expensive bench driver (Thorlabs TED200C – about 1,100 EUR ex-VAT) to control the TEC element (a TECD2S also bought from Thorlabs). Since I had previous experience with the MTD415TE board (see [»] here), I thought it would be a good opportunity to test if it was good enough to drive my laser cooler assembly.

To avoid soldering anything, I relied on Thorlabs MTDEVAL1 board which comes with a very nice software to view the temperature and TEC settings (current/voltage). The software also allows for PID tuning as we have shown in [»] this post. Since I wanted some kind of definitive solution, I made a housing in 3D printing and included a USB cable connector, power connector, TEC connector and a fan. The resulting box is shown in Figure 1.

The complete system is functional and can drive the CPS532 laser at the 40°C set point that I discussed previously. Using the procedure for PID tuning, I found constants of P=2.340 A/°C, D=1.253 As/°C and I=1.092 A/°Cs with a cycle time of 30 ms. The system reaches the set-point within a few minutes and can recover quickly from thermal load variations. And while it’s difficult to check the exact temperature stability because the program does not export raw data for post-processing in either Excel or Matlab, we can still see in Figure 2 that the temperature is well stabilized within ±0.01 C° over a period of 30 sec. According to the MTD415T datasheet, this is at the temperature reading limit of the IC (datasheet says “better than 10 mK, typ. 2 mK”) and consists probably of a lot of electrical noise or read-out noise rather than actual temperature variations.

Despite the system is operational, there are two things that I am not 100% happy in the system of Figure 1.

First, the overall thing is extremely bulky. When you know that the MTD415T IC is about 2 cm² and that the box is above 230 cm², it feels like a lot of wasted space. This is mostly due to the size of the USB connector and the MTDEVAL1 board itself. While I used to be a big fan of these USB connectors, I realized that they’re clearly not made for small spaces as it was extremely difficult to squeeze the USB cable in the box without hurting the soldered USB connector on the MTDEVAL1 board.

Second, the MTDEVAL1 is relatively expensive for what you actually get (140 EUR ex-VAT). The board only contains a FT232 IC for USB communication and a voltage regulator⁽¹⁾. The board does not come with the TEC driver itself (you need to purchase it separately), nor USB cable, power supply or TEC cable.

A thermal analysis revealed that the fan was a good idea when putting the board in a closed box. Most of the heat was actually generated by the voltage regulator although the MTD415T chip itself went to ~35°C in natural convection conditions using a small heatsink on top of it. Using the fan, its temperature did not go beyond 30°C inside the box.

Due to the two remarks (size and cost), I’m now considering making a custom PCB for the MTD415 IC and hold it directly on the laser cooler block. This however poses the problem of the programming of the MTD415T because we would need either to reprogram a complete software and manage serial communication with the IC, or use the daughterboard version (MTD415TE) and move the IC to a MTDEVAL board every time we would like to change the settings.

Since I’m not a big fan of either of these solutions, I started experimenting with the Thorlabs software provided with the MTDEVAL1 board since it’s relying on a very common FT232 IC.

It appears that Thorlabs made its own variation of the FT232 Virtual COM Port driver and ships it with the Thorlabs brand on it. On the other hand, it’s still possible to access the FT232 eeprom data using the FTDI software FT_Prog as shown in Figure 3.

From that data, I was relatively confident that I could directly hook-up a FT232RL board (such as [∞] this one) to the MTD415T by plugging the RX/TC lines (warning: FT232 TX must be connected to MTD415 RX line and vice-versa). The board can be accessed at 115,200 bauds through a program like Termite and commands sent as human-readable text (e.g. Te? to retrieve the current temperature). All the commands are very well described in the MTD415T datasheet.

A very good news is that there does not seem to be any protection in the Thorlabs Software at all because I was able to connect to my own FT232RL IC without any issue. The only thing I noticed was that the software does not display any name when enumerating the COM ports (while it does for the MTDEVAL1 board). I tried playing with the VID/PID and string descriptors in FT_Prog but wasn’t able to program the FT232RL EEPROM because I actually got a Chinese clone with no eeprom! (as confirmed by [∞] this post since my S/N was A50285BI). Conclusion: don’t buy your ICs on Amazon… I will have to repeat the experiment with an official FT232 IC.

Nonetheless, the conclusions of these tests are very positive as I was able to show that it’s very easy to hook up a FT232RL IC to the MTD415T and still get access to the Thorlabs software for PID tuning and modifications of the temperature set-point.

So, what’s next?

From the experimentation performed today, I will start working on a custom board to attach directly to the TEC cooling block. However, since designing and manufacturing a PCB takes time, I decided to split this post in half and already present the first part today. Also, not all decisions are taken concerning this new board. For instance, there is the problem of the current limitation of the USB 2.0 protocol from which the FT232RL relies on. We can only get 500 mA from the USB host which is not enough so using the FT232RL directly would require having a separate power supply for the board. Another possibility is to use the new FT232HP which supports USB3 power delivery but which is more complex to use.

Be sure to stay tuned for updates about this board :)

Want to discuss this further? Check out our new [∞] community board!

I would also like to give a big thanks to Young, Naif, Samuel, Sebastian, James, Lilith, Alex, Stephen, Jesse, Jon, Sivaraman, Cory, Karel, Themulticaster, Tayyab, Marcel, Kirk and Samy who have supported this post through [∞] Patreon. I also take the occasion to invite you to donate through Patreon, even as little as $1. I cannot stress it more, you can really help me to post more content and make more experiments!