Peltiers die rapidly from thermal fatigue when controlled by thermostats. Fast PWM is okay, but it leads to excess I**2 R losses unless you filter it. You really want current control rather than voltage, so ideally you want to use a huge inductor in series.
One method that works well for that is (surprise) two cap multipliers in inverse parallel, one PNP and one NPN. You want to put the Peltier in the collector leads, where the circuit looks like a huge inductor.
Cheers
Phil Hobbs
Den torsdag den 10. september 2015 kl. 00.12.57 UTC+2 skrev Jeff Liebermann:
also looks quite sketchy, random piece of wood used as insulation, peltier with broken off corner
-Lasse
It looked more like a copper base under the Peltier junctions than wood. However, I can't be certain. I did notice the broken corner, which I don't think will seriously effect operation. However, you're right that it does present a risky purchase. Probably easier to just buy some Peltier junctions and run them from a bench power supply.
I recently bought a small B&D refrigerator for $12 on CraigsList. It had a bad Peltier unit which I replaced with one I had previously purchased on eBay for about $6. I also had to repair the power supply because an output diode had gotten hot and desoldered itself. It seems that the temperature control connects to the 12V supply and adjusts the voltage - I can hear the fans change speed when I turn the dial. A full discussion is here:
You can get a 72 watt Peltier module, including fans and shipping, for about $17:
Peltier modules are perhaps 1/3 to 1/4 as efficient as conventional cooling, but they have advantages of size, complexity, audible noise, and cost (if you get surplus units).
It should be easy to make a thermistor or RTD based linear control, or even use PWM with some filtering. Maybe a cheap DC motor control module or buck module.
Paul
The idea would be to tape small parts or breadboards to the top, over some gap-pad stuff, and maybe throw some bubble wrap or something overall for insulation, and run the temp from maybe -20 to 80C. I can't imagine the DUT dissipating more than a watt. A package with interface might cost $1000 or so maybe.
There are tons of fancy chem lab hot plates, but few that will go hot/cold. They tend to be expensive.
Temperature testing is really useful in ensuring that a design is robust. Right now we are testing some 0805 resistors for TC and voltage coefficient.
Den torsdag den 10. september 2015 kl. 00.52.13 UTC+2 skrev John Larkin:
for $1000 I'm tempted to spend the weekend to build on, think I have some 40x40mm peltiers, they are 70W I think
-Lasse
On Wed, 09 Sep 2015 14:27:23 -0700, Jeff Liebermann Gave us:
Wow. That one went right over your head.
Our big one is by Thermotron, and it's very blue.
2-stage cooling goes to -70C. That's some cold beer.
What temperature? Would a toaster oven and a piece of fire rated ceramic tile work?
I'm not sure exactly what you mean by "inverse parallel"in this context but it sounds like a mongo heatsink is in order.
If just 2 cooled temp points were acceptable, bunging it in a domestic fridge freezer might work.
NT
On Wed, 09 Sep 2015 20:25:03 -0400, krw Gave us:
Mongo likes beans...
Not only copper but it's water cooled.
Still in business....
Cheers
I've been considering building something based on this:
They look like a pain to solder for a one-off project. I considered buying two of the eval boards but those are big.
Chris
If you build the peltier into a plug that fits the neck of a wide-mouth vacuum flask then it might be able to achieve lower final temperatures for a larger volume, probably enough for many products to fit into. That is something I have been thinging about for a while.
It is worth looking for a peltier that doesn't use low-temperature solder, as there are some that go to much higher temperatures than others.
Chris
Why would the Peltier device care if you control the current or the voltage? Won't both accomplish the same thing? The temperature is the value being regulated, not the voltage or the current.
I see further down in the thread JL and DLU get into their usual pissing match. I guess I can just skip that part.... or do like they did in that cable show with the robots in the theater seats, where they make up their own dialog. John: You're a butt head... Decadent: No, I'm Bevis and YOU'RE Butthead.
in:
nur
, gup
m mas
n
Technically speaking, it's the current that regulates the heat transfer, no t the voltage. The resistance of the Peltier junction isn't all that well-d efined.
And in fact the heat transfer per unit current - which you have to keep tra ck of to keep the temperature control loop well-regulated - depends on the temperature difference across the Peltier junction.
Sloman A.W., Buggs P., Molloy J., and Stewart D. "A microcontroller-based d river to stabilise the temperature of an optical stage to 1mK in the range
4C to 38C, using a Peltier heat pump and a thermistor sensor" Measurement S cience and Technology, 7 1653-64 (1996)spells out what's going on (and if you are one of the few regulars on the g roup who hasn't yet had a copy, e-mail me at snipped-for-privacy@ieee.org and I'll organise a reprint). In our application the heat transferred per until curr ent could change by a factor of seven from one end of the range to the othe r.
On Thu, 10 Sep 2015 12:56:27 +1000, Chris Jones Gave us:
Much simpler to just use LN2.
*ding*
Drink!
Have something to add? Share your thoughts — no account required.
Ask the community — no account required