Author Topic: Cheap thermistor vs 1-Wire DS18B20?  (Read 20631 times)

EdM

  • NewMember
  • *
  • Posts: 33
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #15 on: May 03, 2016, 09:05:53 PM »
@ Wh  Sorry for being out of touch for a while. 

Yes the aluminum was originally to keep things dry because the waterproofing on those ds18 was so bad I was killing off the sensors during overnight calibrations. But it turned out to be even more useful in the end because the thermal conductivity of the aluminum held all the sensors at exactly the same temp at the reference - dramatically more so than the water did.

Wrt putting the DS18 plastic in contact with the water: we install our loggers down to about 25 m, and the pressure has killed of several different IC temp sensors so far:
https://edwardmallon.wordpress.com/2016/04/06/field-report-2016-03-17-mcp9808-fails-under-pressure/

The DS18's have proved the most robust, but I am beginning to think is because the metal sheathing around those waterproof sensors is providing extra protection from the pressure at depth.  I am now beginning to look at thermistors, and I will probably epxoy them inside of the metal caps as well before we put them under water.

perky

  • Hero Member
  • *****
  • Posts: 873
  • Country: gb
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #16 on: May 04, 2016, 03:22:43 AM »
Thermistors, e.g. NTC resistors, are accurate and robust. That sounds like the best option.
Mark.

pcspinheiro

  • NewMember
  • *
  • Posts: 4
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #17 on: December 13, 2018, 06:28:06 AM »
Sorry to revive such an old topic but it's really a good match for my question. I'm a researcher and work with temperature-controlled solutions in small volume chambers. I have a rpoblem with a recent equipment acquisition (from a different, less known manufacturer) and the temperature controller has a probe based on the DS18S20 (maybe it's a typo as I can only find the DS18B20), the problem being that it's just way too large for this specific application. Temperature probes from other manufacturers are based on bead thermistors, so I was wondering if there is a bead thermistor with similar characteristicas that can be used instead. The input in both cases (bead or DS18S20) is BNC so I would like to just make a probe in house that can be connected to the temperature controller that is expecting to have connected a probe based on DS18S20. Can you please help, as this is not my domain? I know enough electronics to assemble the thing, but I'm lost with the overwhelming amount of different thermistors available. One important aspect is for it to be small.
Thanks in advance for any input.

Paulo from Portugal

TomWS

  • Hero Member
  • *****
  • Posts: 1930
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #18 on: December 13, 2018, 09:05:53 AM »
@pcspinheiro, it sounds as if your temperature controller is expecting a one wire probe, not a thermistor.  The two are not interchangeable.   It would be extremely difficult to translate a thermistor reading to a one wire signal. 

It may be that your temperature controller will accept either type of probe (on separate pins or some reconfiguration).  Can you post any information about your controller?

Tom

HeneryH

  • Full Member
  • ***
  • Posts: 229
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #19 on: December 13, 2018, 09:21:53 AM »
You can get just the DS18B20 by itself without the waterproof packaging.

https://www.digikey.com/product-detail/en/maxim-integrated/DS18B20/DS18B20-ND/420071


is the same as (I believe)

https://www.adafruit.com/product/381

pcspinheiro

  • NewMember
  • *
  • Posts: 4
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #20 on: December 13, 2018, 10:06:55 AM »
@pcspinheiro, it sounds as if your temperature controller is expecting a one wire probe, not a thermistor.  The two are not interchangeable.   It would be extremely difficult to translate a thermistor reading to a one wire signal. 

It may be that your temperature controller will accept either type of probe (on separate pins or some reconfiguration).  Can you post any information about your controller?

Tom

Thanks for the reply Tom. The temperature controller we are using is a Supertech Instruments TMP-5x series (don´t have the exact model number as I'm not in the lab). It's hooked up to their low voltage DC heater for electrophysiology applications. The main problem we have is that the chamber for the experiments is really small (this is normal and best left unchanged) and the probe that they sold us is huge in comparison. Warner Instruments, for example, have a temperature controller that uses a thermistor probe that's no bigger than 1x3 mm and this is ideal, but was way off our budget. Actually, the people that bought the system didn't even realize it had no probe included, so I suggested a thermometer-based calibration of the temperature (which works, but needs a long wait before experiments to reach stability). Therefore, not knowing about these differences in the temperature probes I suggested we try to find a thermistor that has the same characteristics as the DS18B20...

Anyway, it seems like it won't be possible then, but I appreciate your help.

TomWS

  • Hero Member
  • *****
  • Posts: 1930
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #21 on: December 13, 2018, 12:22:22 PM »
Thanks for the reply Tom. The temperature controller we are using is a Supertech Instruments TMP-5x series (don´t have the exact model number as I'm not in the lab). It's hooked up to their low voltage DC heater for electrophysiology applications. The main problem we have is that the chamber for the experiments is really small (this is normal and best left unchanged) and the probe that they sold us is huge in comparison. Warner Instruments, for example, have a temperature controller that uses a thermistor probe that's no bigger than 1x3 mm and this is ideal, but was way off our budget. Actually, the people that bought the system didn't even realize it had no probe included, so I suggested a thermometer-based calibration of the temperature (which works, but needs a long wait before experiments to reach stability). Therefore, not knowing about these differences in the temperature probes I suggested we try to find a thermistor that has the same characteristics as the DS18B20...

Anyway, it seems like it won't be possible then, but I appreciate your help.
Do you need the external sensor?  According to the user manual:
Quote
  It is possible to eliminate the usage of the external temperature
sensor, because in every heated object manufactured by Supertech there is a built-in
internal temperature sensor. Please, do NOT use the external sensor, except under
particular circumstances.


pcspinheiro

  • NewMember
  • *
  • Posts: 4
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #22 on: December 14, 2018, 04:36:34 AM »
Unfortunately yes, we need the external temp sensor because the perfusion chamber for the electrophysiology recording is not from the same manufacturer and, therefore, it does not have any built-in sensor. We have calibrated the temperature needed at the heating element that gives us 30ºC at the bath using a thermometer, but this will obviously not be accurate anymore once a microscope objective is immersed in the solution and starts conducting heat away, and in this case the thermometer won't fit the chamber for a new calibration. So yeah, we're kinda stuck until we find a probe small enough for the chamber. Whoever bought this should have considered these things, and it would have been better to try and find the additional € than to have something that's not really practical or functional. I guess, based on the info I already got from you, that using a thermistor-based probe is not possible at all?

I'll break the bad news to the person that bought this heating system...

Thanks a lot for your help!
Cheers,
Paulo.
« Last Edit: December 14, 2018, 05:53:02 AM by pcspinheiro »

TomWS

  • Hero Member
  • *****
  • Posts: 1930
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #23 on: December 14, 2018, 07:53:37 AM »
I guess, based on the info I already got from you, that using a thermistor-based probe is not possible at all?
Well, nothing is 'impossible' so it is 'conceivable' to make such a translator circuit, however, it probably isn't practical.  You would need an intermediate circuit that had the analog sensing front end to 'read' the thermistor value, some processing (digital or analog) to linearize the thermistor curve, and then some digital circuit to emulate a DS18B20.   It may be worthwhile taking a moment to search for a one wire analog voltage sensor, but then there is no guarantee that the SW in your controller would recognize it.

Sorry for the bad news,
Tom


HeneryH

  • Full Member
  • ***
  • Posts: 229
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #24 on: December 14, 2018, 09:38:03 AM »
Did you see my post above about a small DS1820B sensor?  Maybe I'm missing something but I thought that was the answer.

ChemE

  • Sr. Member
  • ****
  • Posts: 419
  • Country: us
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #25 on: December 14, 2018, 01:57:38 PM »
Translating from one-Wire to thermistor sounds like more work than just spinning up a new controller that works with bead thermistors.  All you need is a Moteino w/o a radio, a SSR, and the thermistor.

Regarding DS18B20 vs DS18S20 they are virtually the same: https://www.maximintegrated.com/en/app-notes/index.mvp/id/4377

EDIT: Sticking a BNC plug on the end of a DS18B20 would not be difficult at all by the way.  You'll just need to figure out if the controller is expecting the pullup resistor to be present or not.  One can easily configure the uC in such a way that the resistor isn't needed (which I always do).
« Last Edit: December 14, 2018, 02:08:19 PM by ChemE »

pcspinheiro

  • NewMember
  • *
  • Posts: 4
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #26 on: December 17, 2018, 05:23:15 AM »
Did you see my post above about a small DS1820B sensor?  Maybe I'm missing something but I thought that was the answer.

Thanks, I did see that, but I can't find any data about its physical dimensions. Plus, if the encapsulated version is as big as shown in the picture then it's WAY too big. I need something smaller than the resistor being shown for size comparison...

Right now I think the easiest solution is to just live without the chamber probe and painfully calibrate (once, as the room itself is strictly temp. controlled) the temperature with a thermometer in similar conditions as the ones used for recordings... I was not the one creating this problem (and I strongly recommended the Warner system), so I'm not going to go to great lengths to solve it..

I really appreciate all your help and information!
Cheers,
Paulo.
« Last Edit: December 17, 2018, 07:36:38 AM by pcspinheiro »

jra

  • Jr. Member
  • **
  • Posts: 81
  • Country: us
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #27 on: December 17, 2018, 02:29:16 PM »
There is a OneWire emulator available https://github.com/orgua/OneWireHub that might handle the translation piece.  You would have to pick an appropriate temperature sensor and figure out how to read the temperature and use the emulator to make it look like a DS18B20.  See https://www.instructables.com/id/Arduino-1-wire-Generic-ClientSlave-Device-Sensor/ for an example.  Feasible yes, practical not so sure.

HeneryH

  • Full Member
  • ***
  • Posts: 229
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #28 on: December 17, 2018, 03:57:09 PM »
Thanks, I did see that, but I can't find any data about its physical dimensions.
Here are its dimensions https://www.analog.com/media/en/package-pcb-resources/package/pkg_pdf/to-92t/T-3-1.pdf

TomWS

  • Hero Member
  • *****
  • Posts: 1930
Re: Cheap thermistor vs 1-Wire DS18B20?
« Reply #29 on: December 17, 2018, 04:00:50 PM »
There is a OneWire emulator available https://github.com/orgua/OneWireHub that might handle the translation piece.  You would have to pick an appropriate temperature sensor and figure out how to read the temperature and use the emulator to make it look like a DS18B20.  See https://www.instructables.com/id/Arduino-1-wire-Generic-ClientSlave-Device-Sensor/ for an example.  Feasible yes, practical not so sure.
Great tip!  Thanks!  This could be useful!

A quick search for Thermistor library yielded quite a few results.  I think you may have to review the quality of each of these, but the electrical interface is easy enough.  Thermistor and 1 resistor.   If anyone has any experience with a decent thermistor library, speak up!

Coupling these two should be straight forward.

Tom