[volt-nuts] Temperature controller for ovenizing and temperature cycling

Sun Feb 2 06:00:40 EST 2014

Hi and thanks for feedback,

The LTC2057 I used for feedback has an open loop gain of 150dB and
therefore hard to stabilize with a thermal feedback only. Adding electric
feedback too, brought the total overnight stability to a total of
0.3mCptp, less than 0.00005Crms.

Without searching much, the commercial offerings I found, are specified at
0.1C to 0.01Crms. If you know some with better specs, commercial or,
circuit or other, please point...

IC vs discreete and LM35 vs NTC, I could use either depending on drift data.
LM35 is specified max 0.08%/sqrt1000h at max temp (100C or 150C depending
on version). I'd expect a few orders of magnitude lower at 30-50C.

The best NTC I find are from US Sensor Inc, with specs down to 0.001% per
year for reference probes. However their datasheet are very thin and drift
data is very genericly specified.

The big advantage of the LM35 is that it can very easily be used for a
numeric display.

I was asked about my voltage ref; it is a LM327, and for now with standad
0.5% 50ppm/c thin film resistors as divider. Any combination that is "good
enough" could of course be used.

For larger power systems I'd want to use voltage controlled PWM output to a
power MOSFET transistor or a solid state relay.

Jan

On Fri, Jan 31, 2014 at 6:00 PM, <volt-nuts-request at febo.com> wrote:

> Send volt-nuts mailing list submissions to
>         volt-nuts at febo.com
>
> To subscribe or unsubscribe via the World Wide Web, visit
>         https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
> or, via email, send a message with subject or body 'help' to
>         volt-nuts-request at febo.com
>
> You can reach the person managing the list at
>         volt-nuts-owner at febo.com
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of volt-nuts digest..."
>
>
> Today's Topics:
>
>    1. Temperature controller for ovenizing and temperature      cycling
>       (Jan Fredriksson)
>    2. Re: Temperature controller for ovenizing and temperature
>       cycling (ed breya)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Thu, 30 Jan 2014 20:10:40 +0100
> From: Jan Fredriksson <jan at 41hz.com>
> To: volt-nuts at febo.com
> Subject: [volt-nuts] Temperature controller for ovenizing and
>         temperature     cycling
> Message-ID:
>         <CAFoWNwA9jkETwbe=
> mEMPfEYQVXKO+kCMHa-0RJO9EFX9CXnaPA at mail.gmail.com>
> Content-Type: text/plain; charset=ISO-8859-1
>
> I just built a small heater / regulator board for ovenzing components. I
> should also be usable for temperature cycling. The heater outputs up to
> around 0.6W, runs on 12V.
>
> It is designed with a voltage sensor and opamp that compares the sensor
> output to a fixed (setpoint) voltage divider. The opamp output drives a
> small power transistor. Pretty much like the LTZ1000 datasheet circuit but
> with a few tweaks.
>
> I have used the LM35 temperature sensor as sensing element, and compared it
> to a reference voltage, derived from a zener.
>
> The LM35 has an output of 10mV / C. I gave it a separate buffered output,
> amplified it by a factor 10, so I can easily monitor the temperature with a
> voltmeter (ie 5.0V = 50C)
>
> The heater resistors give about 0.4W output on average and the regulator
> loop switches on/off about once per second at best.
>
> It has been running since last night. Short term stability has been about
> 0.002 - 0.004C peak to peak (2-4 milli Celcius (mC) ) over a few minutes,
> at a nominal temperature of 33C. The temperature has varied by around
> 10mC as I have tried different component values.
>
> I have not really measured the drift so far, as I have been tinkering too
> much with the circuit. It seems to be no more than one or two mC so far.
>
> I am prettu sure that If I place the sensor and heater behind copper plate
> or heatsink, the other side of the plate would consistently see less than 1
> mC rms oscillations. Long term drift should only depend on the LM35, a
> resistor pair and the zener reference.
>
> If someone is interested, I will get back when I have built a few more
> boards, tuned the component values, made some longer tests.
>
>
> ------------------------------
>
> Message: 2
> Date: Thu, 30 Jan 2014 11:29:06 -0800
> From: ed breya <eb at telight.com>
> To: volt-nuts at febo.com
> Subject: Re: [volt-nuts] Temperature controller for ovenizing and
>         temperature cycling
> Message-ID: <201401301929.s0UJTJAl010424 at mail42c40.carrierzone.com>
> Content-Type: text/plain; charset="us-ascii"; format=flowed
>
> I think you'll get better stability by using a resistive thermistor
> as the sensor. As far as I know, all commercial oven circuits for
> OCXOs and the like use RTDs, and not IC sensors, for best performance.
>
> Ed
>
>
>
> ------------------------------
>
> _______________________________________________
> volt-nuts mailing list
> volt-nuts at febo.com
> https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
>
> End of volt-nuts Digest, Vol 53, Issue 25
> *****************************************
>