The big TEC driver thread!

Quote:

---

Originally Posted by dnar

Ok Sleepless in Germany, I'll have a butchers hook over the next few days. Time to get drunk as it's my 47th orbit of Sol today. I would not like to see this finalised until V1.0 is nailed.

BTW, with your Maxim choice, take a look at pricing. The MAX233 is something like 3 x the cost of MAX232A.

---

Well happy birthday to you dude.... 47... those were the days my friend (I'm 52). Wish you all the best and I'll drink one on your birthday on this side of the Globe.

Jeah I did compare prices on the 232 and 233. Here the 232 is about 3€ and the 233 about 5€ max. (the site is down at the moment, can't double check!)

Sure we're gona nail V1.0 first, but I just couldn't see you striding on by yourself to V2.0. Maybe I'll just send both to ya for your sleepless nights.

Cheers!

Quote:

---

Originally Posted by dnar

Time to get drunk as it's my 47th orbit of Sol today.

---

HAPPY B'DAY dnar....
Have a drink on me...

Jerry

Happy birthday dnar

Cheers,*DZ
Ed

@dnar: Happy 47th, dnar! I'm only 4 years behind you ;)

@solarfire: Some suggestions for the schematics:

v1/v2 - Shouldnt we be monitoring the state of the LMD18200's overtemp line with the micro, not just with a LED?

v1/v2 - Could we add support for an RTD temp sensor, as well as NTC? They are more accurate, and the small ones have VERY fast responses. Not that much more expensive than an NTC therm.

v2 - Better to use a 10 pin header instead of a DB9 connector
v2 - Suggest a MAX231 instead of 233.. Cheaper, very common, same DIP pinout as max232, and only needs 0.1uf caps.
v2 - Remember to tie unused rs232 driver TTL inputs to ground, and rs232 line side inputs to -9v


I have a PCR reactor (VERY fine temp control, better than 0.01C) kicking around at work, so I'll have a look at how it drives its peltier for some more ideas.

Cheers,

Pete

Quote:

---

Originally Posted by lasersbee

HAPPY B'DAY dnar....
Have a drink on me...

Jerry

---

Thanks! I did.

Quote:

---

Originally Posted by Solarfire

Well happy birthday to you dude.... 47... those were the days my friend (I'm 52). Wish you all the best and I'll drink one on your birthday on this side of the Globe.

---

Thanks!

Quote:

---

Originally Posted by megaton

Happy birthday dnar

Cheers,*DZ
Ed

---

Thanks!

Quote:

---

Originally Posted by ogoun

@dnar: Happy 47th, dnar! I'm only 4 years behind you ;)

@solarfire: Some suggestions for the schematics:

v1/v2 - Shouldnt we be monitoring the state of the LMD18200's overtemp line with the micro, not just with a LED?

v1/v2 - Could we add support for an RTD temp sensor, as well as NTC? They are more accurate, and the small ones have VERY fast responses. Not that much more expensive than an NTC therm.

v2 - Better to use a 10 pin header instead of a DB9 connector
v2 - Suggest a MAX231 instead of 233.. Cheaper, very common, same DIP pinout as max232, and only needs 0.1uf caps.
v2 - Remember to tie unused rs232 driver TTL inputs to ground, and rs232 line side inputs to -9v


I have a PCR reactor (VERY fine temp control, better than 0.01C) kicking around at work, so I'll have a look at how it drives its peltier for some more ideas.

Cheers,

Pete

---

Thanks. All good suggestions. I thought the MAX232A was identical to MAX232 except for the 0.1uF caps. I am not expecting the bridge to get hot let alone warm, being a low Rds FET bridge.

Quote:

---

Originally Posted by ogoun

@dnar: Happy 47th, dnar! I'm only 4 years behind you ;)

@solarfire: Some suggestions for the schematics:

v1/v2 - Shouldnt we be monitoring the state of the LMD18200's overtemp line with the micro, not just with a LED?

v1/v2 - Could we add support for an RTD temp sensor, as well as NTC? They are more accurate, and the small ones have VERY fast responses. Not that much more expensive than an NTC therm.

v2 - Better to use a 10 pin header instead of a DB9 connector
v2 - Suggest a MAX231 instead of 233.. Cheaper, very common, same DIP pinout as max232, and only needs 0.1uf caps.
v2 - Remember to tie unused rs232 driver TTL inputs to ground, and rs232 line side inputs to -9v


I have a PCR reactor (VERY fine temp control, better than 0.01C) kicking around at work, so I'll have a look at how it drives its peltier for some more ideas.

Cheers,

Pete

---

Ok.. V1.00 is wrapped up as far as I’m concerned. To the LMD18200's overtemp question, we're all I/Oed out.

And for the RTD temp sensor in V1.00 would require more PCB-Space which is already bigger than I was hoping for. I also did a part cost assessment (Germany), without the PCB, and it comes to roughly 24€. Prices were sourced at hobby level suppliers. Attached are the V1.00 schematic, placement and rout doc’s. The Eagle files won’t be posted until dnar has implemented and tested the Software and confirms the build. If the build and R values confirm some of the trimmers may be replaced with standard 1% resistors.


Also attached, another prelim update of the V2.00 schematic and PCB-placement plan. I ended up using a MAX232A due to sourcing issues here of the MAX231 (special order?), the only supplier which had it stock had it priced 3 fold of the 232A. Can we live with the 232A?


As for RTD temp sensor in V2.00, I’ll have to take a closer look at that one.

Is there a particular reason why everyone is set on a 10 pin header instead of a DB9 connector for the RS232? The pads I’m using are for PCB-Edge mounting (see the Placement plan).


Cheers!


Hey Wayne! How’s the head doing there buddy?*retardo

Nice clean job on the PCB...
I would stay with the RS232 connector. They are easy
to find and cheap to buy and easy to wire..
And if you want to add USB all you need is a standard
RS232 to USB Converter Cable for around $2.00 that
will eliminate the use of extra PCB real estate for USB
connectivity.


Jerry

Quote:

---

Originally Posted by Solarfire

Ok.. V1.00 is wrapped up as far as I’m concerned. To the LMD18200's overtemp question, we're all I/Oed out.

And for the RTD temp sensor in V1.00 would require more PCB-Space which is already bigger than I was hoping for. I also did a part cost assessment (Germany), without the PCB, and it comes to roughly 24€. Prices were sourced at hobby level suppliers. Attached are the V1.00 schematic, placement and rout doc’s. The Eagle files won’t be posted until dnar has implemented and tested the Software and confirms the build. If the build and R values confirm some of the trimmers may be replaced with standard 1% resistors.


Also attached, another prelim update of the V2.00 schematic and PCB-placement plan. I ended up using a MAX232A due to sourcing issues here of the MAX231 (special order?), the only supplier which had it stock had it priced 3 fold of the 232A. Can we live with the 232A?


As for RTD temp sensor in V2.00, I’ll have to take a closer look at that one.

Is there a particular reason why everyone is set on a 10 pin header instead of a DB9 connector for the RS232? The pads I’m using are for PCB-Edge mounting (see the Placement plan).


Cheers!


Hey Wayne! How’s the head doing there buddy?*retardo

---

Hi mate, the heads is ok! I had a big sleep to shake this cold. Runny nose, sneezing all the time and it's 37c here!

I'll checked the files, all good. I did contemplate a separate Vin for the bridge. It could make sense to use 5V for small TEC's. BTW, with MAX232A I have always used non-polarised 100nF caps. I can't recall if this is recommended or not.

I prefer an IDC10 header for several reasons, the PCB size is reduced and you can locate the panel mount D9 away from the PCB easily. Personal preference, that is all. I feel temp monitoring of the bridge is not necessary as total dissipation should be minimal with low Rds FET's hard on or off. Those bridges are designed to switch higher loads at higher voltages than we would ever encounter. It will be interesting to test sans heatsink, as I feel that also won't be a requirement. Look at the Beer Keg controller, a tiny SMD switch, that did not even get warm in my tests to 4A at 12V...

On the subject of NTC etc, I have also used LM35 which provide an accurate 10mV/C. I have not used them in high speed applications however, so can't comment if they are more suitable than NTC in this application. I do like them for the fact you can epoxy the flat face to the heatsink for improved thermal bonding when compared to the curved facets of NTC thermistors.

I plan to spend some time on code for V1 this weekend.

I know you guys all ready have this stuff down to a sience and more than likly have any and all this infomation like the TEC contorler elec schematics and such, bt just in case you don't or some one like me needs it here it is, yes some of it was collected from PL's TEC thread.
I hope this will some what be helpfull to some member here. Enjoy!






Attachment 23682Attachment 23681Attachment 23674Attachment 23675Attachment 23677Attachment 23676Attachment 23678Attachment 23679Attachment 23680

Quote:

---

Originally Posted by dnar

I'll checked the files, all good. I did contemplate a separate Vin for the bridge. It could make sense to use 5V for small TEC's.

---

Unfortunatly this bridge's Vs is 12-55V which is why current limitation is essential!

Quote:

---

Originally Posted by dnar

BTW, with MAX232A I have always used non-polarised 100nF caps. I can't recall if this is recommended or not.

---

In the MAX232A datasheet 0u1 poled caps are used..

Attachment 23683

Quote:

---

Originally Posted by dnar

I prefer an IDC10 header for several reasons, the PCB size is reduced and you can locate the panel mount D9 away from the PCB easily. Personal preference, that is all.

---

Ok.. IDC10 header implemented

Quote:

---

Originally Posted by dnar

I feel temp monitoring of the bridge is not necessary as total dissipation should be minimal with low Rds FET's hard on or off. Those bridges are designed to switch higher loads at higher voltages than we would ever encounter.

---

Yeah.. I don't think so either

Quote:

---

Originally Posted by dnar

On the subject of NTC etc, I have also used LM35 which provide an accurate 10mV/C. I have not used them in high speed applications however, so can't comment if they are more suitable than NTC in this application. I do like them for the fact you can epoxy the flat face to the heatsink for improved thermal bonding when compared to the curved facets of NTC thermistors.

---

For the NTC's I have a 4mm bore between Diode and TEC 8mm into the mount, embedded in thermal compound.

Attachment 23684