Posts: 16
Joined: Mon Jun 18, 2018 12:57 pm

Power supply and element suggestions for my first project ever

Mon Jun 18, 2018 2:07 pm

Hi, all out there.

I'm new to this community and in electronics world at all. I'm happy to announce that I'm about to begin my first project ever and I would like to read your suggestions and additions to my so far research on that topic.

Let me first describe in few lines the idea of that project. All in all is that I want to heat and cool a small volume of solution around 100 to 500 microliters. Since I'm coming from a web developing background, I was though to use raspberry pi zero W, in order to create an AP (access point) which will serve a specific web page where a user will set something like that:

Code: Select all

1 - Ramp up to 65 C and stay @ 65 for 20 min
2 - Ramp up to 95 C stay @ 95 C for 10 min
3 - Ramp down to 20 C and stop
So as I already mentioned, I chose as the main platform for that project the raspberry pi zero W. Further, I broke down the whole project into smaller pieces.


For that stage, I conclude on two different approaches.
1) Peltier element.
While searching online for a Peltier that can reach out the temps that I want, I found these two [ 1, 2 ]. Peltier is a good idea because it gives you the chance to use it for cooling too. For this to happen polarity should be changed and thus an H-bridge is needed. I found this one on eBay and it seems really widely used.

2) Heating element.
This is another way to heat up the solution. I'm not sure if this approach makes sense or not and it good to be "judged" from you. So far I found this heating element on eBay too but I really don't know if this is going to work for me. Any ideas here are welcome.

Independently of using Peltier for cooling or not, I take the existence of a fan for granted.

As you can imagine, in order to keep the temperatures as stable as possible I'm going to need a temperature sensor like this or that one.

This is the last sub-problem that I figured out. The whole project I would like to be powered from only one socket/cable and not more but as you can see, for example, the Peltier elements, want about 15.5-15.7 V to work out while at the same time raspberry pi zero W wants 5V. So which is the best approach for achieving something like this? Are there any module/boards that have one power input and can split voltages into different values? How did they call?

As I already mentioned, this is my first ever project. I have a programming background but not that much in electronics. So If you think that all the above are totally wrong or that I have some mistakes/missing elements etc. please enlighten me with your knowledge. I would very thankful. Of course, if you believe that I need to change some elements please also let me know.

P.S In your answers, act like there is no budget problem.

Posts: 1231
Joined: Thu Oct 09, 2014 2:50 am

Re: Power supply and element suggestions for my first project ever

Mon Jun 18, 2018 7:42 pm

You did not mention the time frame for coming to the new temperature, but the peltiers you give as example are huge, power hungry and super costly. I bumped up against a few of these problems in a past life so that is my frame of reference. My task was to maintain the temp of small volumes of flowing liquid and on another task, to raise and maintain about 500 uL in 60 seconds from ambient to 38C +/- 1C.

Clearly you will need a heat sink for cooling the chip surface and depending on time frame and the thermal mass of your container, a fan may be needed as well. The thermal mass of your sample is quite small and will probably be less than the thermal mass of your sample container. Factors such as the thermal conductivity of the materials involved and the thermal resistance of the various interfaces: heatsink to air, chip to heatsink, chip to sample container and sample container to solution under test...not necessarily all inclusive, but representative of what you are facing. The time constant of the measuring system is a factor that needs consideration as well

I am currently using the ds18b20 that you show in "this" example and have had good luck with it but I think it's tolerance is on the order of +/- 0.5C so depending on your need for accuracy, you would require some calibration. That is not a big deal for a one-off but might be in a production setting. If you require greater accuracy, the dual thermistor element devices from outfits like YSI (if still in business) might be a better choice but would require additional circuitry for A-D conversion and a stable source of power. Also, the ds18b20 in the metal enclosure as shown seems to require quite a bit of time to equilibrate to a new temp and it has considerable thermal mass relative to your sample size. For that matter, so do the peltier chips.

There are numerous, small boost and/or buck regulators to be found on eBay and other sites. As the peltier is the real current hog, you will probably want to start out with a voltage suited for it. The RPi and associated bits will require much less and you can step down from the supply for the peltier to the 5V that is needed for a 0W and its associated bits.

It is all possible. There are a few problems to be worked out but the bits you contemplate will work...but do try to downsize the peltiers if you can. There are tons of H-bridges on the market and you will just need to match the power capability and voltage drop of your bridge to the current requirements of the peltier you choose and the voltage of the power supply you choose. Also, I am sure you realize that you will not be able to get away with a "bang-bang" type of control unless you can tolerate overshoot. Some sort of a PID algorithm will be needed, with some tuning. It is also possible that some of the tasks to be done here are better handled with a simple controller like the Arduino with its built in A-D conversion and outputs that are set up for PWM-analog.

The issues I mention above are not all-inclusive, but are representative of what you will face in this project. Good Luck!

Return to “Automation, sensing and robotics”