OpenSource Gene detector - Needs PWM peltier control & Amp'd Photodiode

[Fluffee]

Heya

I've taken on a project that is a tad out of my league in terms of circuits... Basically its an aluminium block being heated to 63°C for 20 - 60 minutes. This block has sample tubes in them in which a reaction (LAMP - type of Isothermic PCR) takes place producing flurecence if a gene is present, and I hope to measure this using an amp'd photodiode. My aim is to have the device as simple, cheap and fixable as possible, as it is intended to go out to sea/be used by the public.

Heres what im using

Arduino Uno

2 x Peltiers
Temperature Differential (DT) (Th=27 degrees C) : >=68 degrees C
Optimum Input Voltage (Vmax) (Th=27 degrees C): 15V
Optimum Input Current (Imax) (Th=27 degrees C):8A
Maximum Cooling Power (Qmax): 68.09W

OPT101P-J - PHOTODIODE
**broken link removed**

LM335Z temperature sensor

BUZ71 N Gate MOSFET

As for the peltiers, I know this has been asked before, but ive read many conflicting things such as low voltages are bad for peltiers, high frequency pwm is bad and the various ways of smoothing the pwm signal....

My circuit works so far, but the oscilloscope im using is cheap, plus I dont realllly know how to use them... So not 100% sure if the signal is doing damage to my peltiers =(


My questions so far....

1) Currently im using a LPF with R - 0.3 and C - 220uf. This gives me a cut off of around 2khz according to an online calculator, is this correct.

2) Ive heard of people using LC circuits and DC-DC bucks, what are the pro's/cons of these?

3) I can either output a pwm freq of ~3khz or ~7khz, im assuming I should go the 7?

4) I'd like to stear clear from using PID to keep it simple. Im assuming I can just program it to keep within 63 fairly easily (It can stray a degree or two)

5) Oh, ive read some people placing a resistor between the pwm output and the Mosfet, what does this achieve?


I think thats more than enough for now, I'll leave the the photodiode stuff till I do some testing with it. Any help is muuuuch appriciated. Sorry for being an electrical noob, my passion lies with biology, but I have been edging into electrics for the last year or so =)

Cheers

-Fluffee

 

[pinout]

Hi Fluffee,

I've worked on these for design and repair (of commercial products) for several PCR reactor blocks and reaction stations over the years, so might be able to help.

Firstly Peltiers 'prefer' smooth DC, ripple and PWM reduce their efficiency, but not necessarily their operating life as far as I'm aware. Mechanical and thermal stresses are what cause most failures in my experience.
Low voltage (smooth DC) has not been an issue in my experience, in fact at the desired operating point this reduces the power in the module which is good.

Depending on the thermal inertia in the block, you might find a simple switching circuit works better than PWM, i.e. Peltier on full and at a point near the desired temperature Peltier off. Then Peltier on for a short while, then off..... But you can see the problem! This is a bang-bang servo and needs to be in a control loop, but just a proportional control might be all you require. It also only requires a switch to power the Peltiers and a low on resistance power MOSFET would work well.

The mechanical mounting of two Peltiers will probably be the stumbling point. I have done this successfully working with the manufacturers, but the problem is you need the metal surfaces to be ultra flat and the Peltiers should be lapped for a constant thickness (used in matched pairs).
They have to be torqued up slowly and precisely to avoid the ceramic or elements fracturing.
Sepearteheat sinks for each Peltier is a good solution here.

The screws also need to be considered as they are thermal short circuits between the hot and cold faces. Plasic screws can be used but they are not mechanically rigid, so don't torque up as well.
The best soltion we found (simple engineering) is to have a clearance hole for the screw in one plate and an insulating locating washer (top hat shape) between the screw and the block. The washer part is kept thin and a metal washer placed over it to spread the load.

Back to the Peltiers, they are only required for cooling, so what temperature do you need to cool down to?
For a cheap public product could you consider a heated resistive block and a fan on the heatsink for fast cooling to ambient? From memory you need to go below ambient to stop the PCR process, but you need to refresh my memory here!

I could go on for hours on this but I hope this helps as a starter with your thoughts.
I'll try to follow your posts and discussions, but send a private message if I drop off.......

 

[Fluffee]

Thanks heaps for the advice =)

Im actually using the peltiers to heat the block! Is this strange? It seems to work ok.... We chose them since they are compact and seem fairly energy efficient.... The Thermal cycler we pulled them out of used simple resistive elements to do the heating and a fridge compresser/oil for the cooling. As this is is a special kind of PCR which occurs at just one temperature.

https://en.wikipedia.org/wiki/Loop-mediated_isothermal_amplification

I like the idea of just turning them on and off as needed, but wouldnt this cause the thermal stress isue?

As for mounting, I've just stuck them onto the bottom of the block (already extremely flat) using thermal glue, It seems to work well, though in the finished design I was thinking of screws, but thats a great idea!

Any thoughts on the smoothing of the DC power?

 

[pinout]

Re: OpenSource Gene detector - Needs PWM peltier control & Amp'd Photodiode

Hi thanks for more info. I wasn't familiar with LAMP method, just thermocycled PCR.

I shall read up, but initial thoughts are if you just want heating resistors, heat pads/mats etc., will be much more robust and lower cost and easier to service.
I would really only use Peltiers when cooling is required, i.e. for high precision around the set point or if I need to get back to ambient quickly. Using a restive heater with a thin low thermal mass plate and even a finned plate with fan (PC processor type) will be much cheaper and robust.

Converting power to heat is 100% which ever method you use, but if you can also extract heat from the environment then Peltiers exceed 100% when heating, but I'm not sure that's a powerful 'selling point' for such a device, but that's your choice.

If you go the Peltier route, there are some guidelines on ripple, but any general purpose supply should be good.
Switching the Peltiers should not really stress them too much as that's what they are designed to do, when correctly mounted.


There is some useful data and guides at:
**broken link removed**
http://www.tetech.com/FAQ-Technical-Information.html#faqs

**broken link removed**
I found this statement in the above link

" ripple should be kept below 15% to maximize power output to the TE modules. Unfiltered DC (i.e., rectification only) is very inefficient and quite problematic if the peak voltage exceeds VMax "

At that site I also found this page
**broken link removed**
which shows a picture of the mounting method I mentioned that we used.

I'll come back later in the week, but keep your thoughts and questions coming.

I used products from a company called TED (Thermo Electric Devices) but I think they have merged now, so can't track them down.

It seems you have done a lot with the Peltiers, but I would also consider the restive heater at this stage in the design even though it's less exciting!

- - - Updated - - -

Forgot to share this

http://www.freefullpdf.com/

This is a great search facility that allows you access to "open source published research papers", I find it very usefull on my budget!

 

[Fluffee]

Re: OpenSource Gene detector - Needs PWM peltier control & Amp'd Photodiode

I just read somewhere about placing the LPF on the arduino side of the mosfet. This makes so much more sense! Wastes far less power... Still curious about the DC-DC Buck and LC solutions though.

I spose the peltiers were the choice since we had some laying around tbh Their flatness also helps for mounting, the resistive ones are rods, so not sure how to best mount them....

I do like the idea of robustness, would the heating pads continue to heat for a while after being turned off though? Would this mess with trying to keep a set temperature?