Fine grained temperature control with a Peltier/TEC

[Gutierrez]

I'll preface this with I'm a computer scientist not an engineer. So feel free to correct away. I'll never learn if you don't.

I'm trying to work out a simple control for creating small temperature changes on an aluminum surface +/- 0.1C/sec to +/- 0.4C/sec. The peltier looked like a good fit for the device so I've ordered two 7w 20mm peltiers online to mess around with.

I've been looking around for simple ICs to control the TEC and I'm hunting around bi-directional motor controllers using H-Bridges. So my question is, given the ratings of the peltier, what specs should I look for in the chip or dip?

Here are the specs on the TEC:
Type: TEC1-04902
10W
Couples: 49
Imax (A): 2
Vmax (V): 5 (usually working <=5v)
QCmax (W):7
Dimensions: about 20 x 20 x 4mm
Tmax △C: 67

I read that the optimum current for a peltier is around 75% of the Imax, so I would think that means I should use a supply of 1.5A at around 4.5V (6.75 watts), is that even right?

The problem I run into is finding an IC that does that. The closest I have come that looks simple to work with is TI's L293 which outputs from 4.5V to 36V and 1A per channel.

If I were to supply 4.5V at 1A to the chip to drive the Peltier would that still work for my application?

My second problem is I need to regulate the temperature of the surface with a decently rapid response time. Is measuring off a LM35 and using the result to regulate PWM on the peltier via the L293NE a good fit? Will all the rapid temperature changes (albeit small) junk my peltier too quickly?

[BMorse]

Check out the LMD18200T H bridge:

•Delivers up to 3A continuous output
•Operates at supply voltages up to 55V
•Low RDS(ON) typically 0.3 per switch
•TTL and CMOS compatible inputs
•No "shoot-through" current
•Thermal warning flag output at 145°C
•Thermal shutdown (outputs off) at 170°C
•Internal clamp diodes •Shorted load protection
•Internal charge pump with external bootstrap capability

The LMD18200 is a 3A H-Bridge designed for motion control applications. The device is built using a multi-technology process which combines bipolar and CMOS control circuitry with DMOS power devices on the same monolithic structure. Ideal for driving DC and stepper motors; the LMD18200 accommodates peak output currents up to 6A. An innovative circuit which facilitates low-loss sensing of the output current has been implemented.

B. Morse

[Gutierrez]

Originally Posted by BMorse

Check out the LMD18200T H bridge:

B. Morse

I ask for simple, you give me simple. A PWM signal will be easy to swing and this will definitely take the load I need to pass without blinking. This will be a great chip to prototype with for a single TEC. Makes me sad National Semiconductor isn't as sample happy as Texas Instruments, though.

[Gutierrez]

Sent Texas Instruments and e-mail and it turns out that they actually produce a whole set of TEC drivers right on spec with my coolers!

Here's the list for parts in series DRV59X.

You can get chips that use an external PWM or use their own internal PWM depending upon your needs.

I'm not really in love with the fact it's PowerPad (32-HLQFP), I'd rather have straight DIP to prototype with, but I'm sure I can find/make a board to solder one to.

They also are a decent price ranging from $6.20 with no PWM to $24 with built in PWM on Digikey for each unit. And best yet... They are on the sample program!

[SgtWookie]

You won't likely find a DIP that can pass enough current.

Most of the newer stuff is SMT. There are very few DIP IC's that are capable of sourcing/sinking more than an amp or so current anymore. A couple of those are the IRLD014 and IRLD024; single logic-level N-ch power MOSFETs with Id=1.7A and Id=2.5A respectively.

But, you'd be better off with an integrated solution than playing around with discrete components.

[Gutierrez]

I'm plugging away on the circuit design and schematics right now. I think I've got the major components and layout for it done. But I wanted to add some flair: LEDs that light showing whether the TEC is hot or cold on the top side. Luckily a coolers performance is based on current flow and PWM. And LEDs are only deal with single direction current flow and use PWM to dim. So it seems I should be able to indicate what's going on with the TEC using a blue and red LED. My problem? 1.5 Amps of current and 4.5 volts that needs to run LEDs. So how can I drive the LEDs off the PWM? I know a blue led runs around 4.5V, but what about RED leds that use lower voltage? I'm working with SMD components, but the board is up to me, so any footprint for the components is workable. From design tips to component recommendations, any help would be greatly appreciated.

[SgtWookie]

Have you tested your TEC to see what the actual voltage across it is with 1.5A and 2A applied?

Going back to look at the TEC specs, it gives a max of 2A, and 5V, but that doesn't necessarily mean the TEC has a flat response of 2.5 Ohms.

Anyway, just figure your current limiting resistors (Rlimit) as:
Rlimit >= (Vsupply - Vf_LED) / Desired_LED_Current

Use the Typ Vf @ current specification for your calculation