TPSM84824: Can TPSM84824 sink current

Hi Joe, the TPSM84824 will be able to sink current but will be limited by the LS sinking current limit of -3.4A typical. But I would like to understand better the conditions where they need this sinking ability. What is going on at the output where they expect the device to have to sink current continuously?

Regards,
Kris

Hi to all,

in my application I have to drive a bunch (4) of Peltier modules. The Peltier modules have cooling and heating ability, depending on the direction of the current that flows through them.

Each Peltier module requires a maximum of 3A (-3A) at a voltage of 4.5V(.4.5V) when heating(cooling) at maximum power.

It should be possible to seamlessly control the voltage (or current) of each Peltier module from -4.5V..+4.5V (resp. -3A..+3A).

Constant current operation would be preferred, but constant voltage operation is also ok, because each Peltier module is part of a separate temperature control loop.

My idea is to do it in the way that is shown in the picture; would that be possible?

Hi Christian,

If I understand your drawing correctly, the supplies on the left and the one on the right all need sinking capability, right? To get the negative voltage across the TEC, you would set the left TPSM84824 supply lower than the buck at the right, which means the TPSM84824 devices would sink current.

Based on the sinking current limit for the TPSM84824, I would not recommend using it where it would have to continuously sink current.

Perhaps a different topology would work better. Here is an app note that gives an example of using a buck-boost converter to drive a TEC.

The other alternative I may suggest is to use an H-bridge (or 4 switches) to drive each TEC, where you would use the H-bridge to control the polarity of the voltage applied across the TEC. The supply to the top of the H-bridge could be powered by a buck module like TPSM84824 with adjustable output.

Hi Kris,

please see my answers in RED:

"If I understand your drawing correctly, the supplies on the left and the one on the right all need sinking capability, right?"

Yes, correct, both sides need sinking capability

"Based on the sinking current limit for the TPSM84824, I would not recommend using it where it would have to continuously sink current."

Is there a technical reason not to do this or could I give it a try? Should I use a TPS54824 with external inductor for this experiment rather than a TPSM84824 module?

"Perhaps a different topology would work better. Here is an app note that gives an example of using a buck-boost converter to drive a TEC."

I am aware of this app note. The problem is that the proposed buck-boost converter can only supply TECs with much lower voltages and currents (-0.4A...+1.6A/-0.4V..+1.6V) than I need: My requirements are +/- 4.5V, +/- 3A, which is way more than such a buck-boost converter can supply!?

"The other alternative I may suggest is to use an H-bridge (or 4 switches) to drive each TEC, where you would use the H-bridge to control the polarity of the voltage applied across the TEC. The supply to the top of the H-bridge could be powered by a buck module like TPSM84824 with adjustable output."

This is probably a good idea, but I have to be able to control the voltage seamlessly between -4.5V and 4.5V, because the TEC is part of a control loop. The thermo-electric cooler must be able to change its cooling output or heating output continuously and it must also be possible to set very low cooling outputs and heating outputs

In order to do this, the buck module and the H-Bridge of your proposal must meet two crucial requirements:

- the buck converter must be able to output voltages down to 0 Volts. The TPSM84824 has a lower limit of 0.6V, but not zero.

- the H-Bridge must be able to work with a top supply of down to 0 Volts. If i look through the H-Bridges I can find the DRV8850, but this one needs even 2 V as a minimum supply.

Can you give me advice?

Hi Christian,

The reason I would not recommend TPSM84824 is that its negative (sinking) current limit is too close to your required sinking current of 3A. The -3.4A I mentioned is also just a typical value, not min/max. If the sinking current limit is tripped, the FETs will be tristated until HS FET turns on and off again, and HS FET will be turned on again only when output voltage has fallen enough such that the control loop issues a HS turn on, which may not happen if current limit has tripped before the output has been discharged sufficiently.

We have TPS543B20 which has a much lower negative current limit of -23A and TPS546A24A which has -10A typical negative current limit, but both have output voltage range of 0.6V to 5.5V, meaning that you would need to dynamically adjust the voltage on both sides of the TEC to achieve +/-4.5V across the TEC.

I referred to the buck-boost example with the TPS63020 just to share the concept, not meaning that it exactly met your voltage and current requirements.While the TPS63020 would not work, perhaps there is a buck-boost controller (external power FETs) that could be used in a similar fashion to meet your requirements. Let me see if the boost controller team has suggestions.

Regards,
Kris

Hi Kris,

The buck-boost-controller TPS55288 could be a good choice, what do you think? Can you please ask the buck-boost controller team if they would recommend that part for that purpose or if they would prefer another one?

Regards

Christian

Hi Christian,

Yes I had asked the buck-boost team for inputs and they pointed me to LM5176. Good find on the TPS55288. I do not see an issue with using TPS55288 either. It does not appear to have any reverse/sinking current limits to worry about.

Regards,
Kris

Hi Krist,

They are facing issue in related to Transients and ripple, resulting board resetting, not booting up, BLER issues and stability.Kinldy advise

RT2053V2_Ripple_data_Resonous.xlsx

Regards

yashwitha

Hi Yashwitha,

I believe this is with regards to your other thread post, but it appears you have followed up on an unrelated thread. Let us continue this discussion on your original post.

Thanks,
Kris