another k-type question

I'm working on a project where I need to read a K-Type thermocouple at temperatures in the 1100C / 2000 F range.

The software part is straight forward, but the electronics side is new to me.

From what I read the AD595-AQ (http://www.sparkfun.com/datasheets/IC/AD595.pdf) looks to be the tool for the job, but I have a couple of questions before I move forward, or look for a different solution. Since the analog pin only supports 5V I don't see a way to actually accomplishing this as the AD595 exceeds 5V at ~500C. Is there another way of reading this, or a circuit / component that can be used to "recompute" or step the resolution down to a percentage of 5V?

10 bit resolution should be fine since it would still give me 1.25C/2.25F per unit, I just need the entire ~12.5V of the amp across the 10 bits.

The second question depends upon the ability to actually use the 595, and relates to power. In order to get the higher temp readings, it needs to have +15V, and (if I understand the specs) -10 on pin 7 to balance the additional 10V added to pin 11.

How would I go about accomplishing this?

I'm not looking for working solutions, just a nudge in the right direction.

Thanks in advance for any help with this. Rob

10 bit resolution should be fine since it would still give me 1.25C/2.25F per unit, I just need the entire ~12.5V of the amp across the 10 bits.

A voltage divider (2 resistors with appropriate values) would do this. Search the forums as this comes up often.

In order to get the higher temp readings, it needs to have +15V, and (if I understand the specs) -10 on pin 7 to balance the additional 10V added to pin 11.

One way is to use an external power supply that has a negative supply.

There are dedicated charge pump ICs to increase the voltage, and inverters to make a negative voltage. Sorry, I'm not familiar with any specific examples.

A typical rs232 transceiver (e.g. MAX232) happen to contain both a charge pump and an inverter, if you tap it at the right place. I'm not sure if it supplies the voltage and current you need, though.

Sorry to keep pimping Maxim devices, but the MAX6674 does 0 to 1024C on a single 5V supply, if 1024C is enough (yeah, I know 1024 is less than 1100).

-j

You can divide the 12.5V down to 5V using a voltage divider with precision resistors. It depends on your accuracy needs. A voltage divider with a 14k and 21k resistor will get you the right ratio, and if both are 1% precision then add about +/-1.2% error to your output voltage before it gets to the A/D. 0.5% and better resistors are available, but for more money of course.

I don't see in the datasheet why you would need a negative supply, unless you plan on measuring negative temperatures.

You are right that you will need a supply voltage ~15V. You can use what's called a "boost converter" or "step-up converter", either one that uses inductors (e.g., LT1316) or a "charge-pump" type converter (e.g., LTC1261). That last part looks like a very simple solution since it can act as a "voltage tripler", putting 15V out when given 5V in.

Thanks so much. These should get me going in the right direction.

I don’t see in the datasheet why you would need a negative supply, unless you plan on measuring negative temperatures.

On a second read through I see where it does specify for negative values only.

Sorry to keep pimping Maxim devices, but the MAX6674 does 0 to 1024C on a single 5V supply, if 1024C is enough (yeah, I know 1024 is less than 1100).

No problem. I actually read several other posts you made, and did the research on the MAX6674. Unfortunately I have to be able to get to 1040C and would like to be able to get up to 1200C.

After doing some research, I believe that I could use any of these three to handle the 5V to 15V step up:

LME0515SC: http://www.cd4power.com/data/power/ncl/kdc_lme.pdf

HPR102C: http://www.cd4power.com/data/power/ncl/tdc_hpr1xxc.pdf

EC1SA03N: http://www.cincon.com/data/products/dcdc1_1/EC1SA.pdf

Am I missing anything here?

The first one is isolated, which you don't need - you'll have to tie the grounds together anyway.

Looks OK to me; I suspect price may become a factor. Some of these switcher modules that are canned are a bit expensive. Discrete ICs with a handful of components may be a bit ugly and messy, but can me lots cheaper.

Not sure what the overall power situation is like (e.g. how you are feeding the arduino), but boosting, say, a 12V supply may be easier than boosting 5V, if you plan to run the final product off a wall-wart or vehicle supply anyway.

-j

Thanks for responding.

The overall cost on these really isn't prohibitive for me at this point. They are only a couple of dollars more than using one of the SOIC converters + misc caps and resistors. When you add in the cost of replacing everything when I put it together wrong and blow it up, the cost goes down considerably ;)

As far as power goes. I'm not sure what the final solution will be yet. The Arduino will be communicating with an app on a PC, providing feedback, and allowing for remote control of a relay. For initial testing it will be within USB reach, so power will be coming from the USB. If I need to have more distance from the PC I will add an Ethernet shield for communication. At that point I would use a wall wart.