Help Peltier elements

[:: Marcolino ::]

Hi, I've a question:
I build a termosensor circuit with termosensor (lm335) and peltier elements
(68W 15V) for cooling my box.
The circuit job is:
- tx > 20°c relais turn on so peltier is on so cool my box
- tx <20°c relais turn off so peltier is off so stop the cooling
regulated with a trigger to avoid micro oscillation of relais.

the problem is after 3 days the circuit go well but my peltier elements
broken... I think than peltier element suffer on/off continuos switch so
have you same ideas to resolve problem?

Thanks

Marco

p.s.
sorry for my english but I'm foreign

 

[legg]

Hi, I've a question:
I build a termosensor circuit with termosensor (lm335) and peltier elements
(68W 15V) for cooling my box.
The circuit job is:
- tx > 20¡c relais turn on so peltier is on so cool my box
- tx <20¡c relais turn off so peltier is off so stop the cooling
regulated with a trigger to avoid micro oscillation of relais.

the problem is after 3 days the circuit go well but my peltier elements
broken... I think than peltier element suffer on/off continuos switch so
have you same ideas to resolve problem?

Running the junction with a bang-bang controller and a lot of
hysterisis means running it with maximum thermal stress. If the
attachment method (source and sink) is too rigid, cracking may result.

If you simply ran it linearly, with degenerated temperature regulation
(ie allow 2degrees rise between 0 and 15V applied range), you'd be
operating at a minumum stress cycles. Reducing the rigidity of the
attachments would reduce the actual stress, regardless of cycle count
and rate of change.

RL

 

[colin]

Hi, I've a question:
I build a termosensor circuit with termosensor (lm335) and peltier elements
(68W 15V) for cooling my box.
The circuit job is:
- tx > 20°c relais turn on so peltier is on so cool my box
- tx <20°c relais turn off so peltier is off so stop the cooling
regulated with a trigger to avoid micro oscillation of relais.

the problem is after 3 days the circuit go well but my peltier elements
broken... I think than peltier element suffer on/off continuos switch so
have you same ideas to resolve problem?

Thanks

Marco

p.s.
sorry for my english but I'm foreign

peltiers that ive seen are fairly robust all that goes wrong usualy is when
they get too hot the solder melts and they fall apart. i asume you have a
big enough heatsink on the cold side of the peltier?

Colin =^.^=

 

[Rene Tschaggelar]

Hi, I've a question:
I build a termosensor circuit with termosensor (lm335) and peltier elements
(68W 15V) for cooling my box.
The circuit job is:
- tx > 20°c relais turn on so peltier is on so cool my box
- tx <20°c relais turn off so peltier is off so stop the cooling
regulated with a trigger to avoid micro oscillation of relais.

the problem is after 3 days the circuit go well but my peltier elements
broken... I think than peltier element suffer on/off continuos switch so
have you same ideas to resolve problem

Peltiers are not intended for such kind of operation.
And you need a fan on both sides to move the warmth or the cold.
What was the switching periond ? 30 seconds, 1 minute, 10 minutes ?

Rene

 

[Roger Hamlett]

peltiers that ive seen are fairly robust all that goes wrong usualy is when
they get too hot the solder melts and they fall apart. i asume you have a
big enough heatsink on the cold side of the peltier?

The 'maximum working temperature', of most Peltier elements like this,
unless higher temperature units are deliberately used, is about 80C, so if
the heatsink is poor, it could easily explain an accelerated failure. They
often have insulation that fails before the slder...
The other question would be how the supply is arranged. If (for instance),
it was going overvoltage when the Peltier module is disconnected, then a
very high current could be flowing through th module, when the relay
switches back on.
Personally, I'd prefer to see a PWM regulation system, limiting the
current through the Peltier, and controlling the temperature as well.

Best Wishes

 

[Bill Sloman]

Hi, I've a question:
I build a termosensor circuit with termosensor (lm335) and peltier elements
(68W 15V) for cooling my box.
The circuit job is:
- tx > 20°c relais turn on so peltier is on so cool my box
- tx <20°c relais turn off so peltier is off so stop the cooling
regulated with a trigger to avoid micro oscillation of relais.

the problem is after 3 days the circuit go well but my peltier elements
broken... I think than peltier element suffer on/off continuos switch so
have you same ideas to resolve problem?

Thanks

Marco

p.s.
sorry for my english but I'm foreign

Bang-bang control is a very bad idea for Peltier elements. The problem is
that the heat transferred by the Peltier effect is linearly proportional to
the current through the junction, but the heat generated in the Peltier
junction by its ohmic resistace increases as the square of the current, so
running a junction at half the current all the time works a lot better than
running at twice the current for half the time.

There is a formula you can use to calculate the nett heat transferred by a
Peltier junction as a function of current - I've put it in print twice (once
in Measurment Science and Technology in 1996, and last year in Review
Scientific Instruments) and much the same formula was published in EDN
around 1998, but I've not yet seen it on the web.

When you are using a Peltier junction as a cooler, the cooling effect
reaches a maximum at a specific current - the manufacturers data sheets
specify this for junction on essentially perfect heat sinks - and if you put
any more current through the junction you actually get less cooling effect.
Rather higher currents are known to melt the (low-melting) solder used to
put the things together - I've not managed this, but one of my bosses
claimed to have unintentionally achieved such a disassembly.

 

[:: Marcolino ::]

period is 2 mins.

thanks,
Marco


"Rene Tschaggelar" <[email protected]> ha scritto nel messaggio

Hi, I've a question:
I build a termosensor circuit with termosensor (lm335) and peltier elements
(68W 15V) for cooling my box.
The circuit job is:
- tx > 20°c relais turn on so peltier is on so cool my box
- tx <20°c relais turn off so peltier is off so stop the cooling
regulated with a trigger to avoid micro oscillation of relais.

the problem is after 3 days the circuit go well but my peltier elements
broken... I think than peltier element suffer on/off continuos switch so
have you same ideas to resolve problem

Peltiers are not intended for such kind of operation.
And you need a fan on both sides to move the warmth or the cold.
What was the switching periond ? 30 seconds, 1 minute, 10 minutes ?

Rene

 

[:: Marcolino ::]

I think now to reduce cell current to half manner to reduce switch timing
and power cell.

Thanks

Marco

 

[Rene Tschaggelar]

2 minutes ?
Within these 2 minutes, the applied power creates a temperature
difference which makes some warmth to flow against the usual gradient.
Then when switched off the lost heat spreads and again the gradient
between ambient and to be "cooled" settles. This kind of operation
gives the lowest possible efficiency. Much lower than the 2 minutes
average current as constant DC. Get or make yourself a controller.
A manual on how it is done is downloadable :
http://www.ibrtses.com/products/tec.html


Rene

 

[Rene Tschaggelar]

Marcolino,
I remember Melcor, one of the manufacturers of Peltiers has
a rather good brochure
http://www.melcor.com/

Rene

 

[Tom Seim]

Bang-bang control is a very bad idea for Peltier elements. The problem is
that the heat transferred by the Peltier effect is linearly proportional to
the current through the junction, but the heat generated in the Peltier
junction by its ohmic resistace increases as the square of the current, so
running a junction at half the current all the time works a lot better than
running at twice the current for half the time.

One project I was involved with had problems with Peltier coolers
because an unregulated DC supply was used to drive them. The ripple
was reducing their cooling capacity for the reason you mentioned. We
switched to regulated supplies; problem solved.

 

[Rich Grise]

I think now to reduce cell current to half manner to reduce switch timing
and power cell.

Don't reduce it by exactly half - reduce it by an adjustable amount,
and don't switch it at all. Simply use regulated DC, and make
the control loop's response time really slow, so you avoid overshoot.