Arduino PWM output, circuit for DRV101 2,3A

Hello guys!

Recently i started with a project for peltier heaters, and found this nice DDpak DRV101. And I need a little help. First ill explain.

I first thought of using this DRV101 with a direct connection to the Arduino. Than i found out that the device needed some fancy inverted voltage that varies from (0.1 to 3.7Volt) whereby 3.7 volt at 90% is closed and 10% closed at 0.1 Volt.

for the feedbackloop i am using a LT1025 compensator and a LTKA001 amplifier as Thermocouple Temperature measurement. The thermocouple part works fine. But i am having troubles with the DRV101 fancy inverted voltage range.

my actual build up looks like

LTKx001+LT1025 ---->Arduino---->LED to ground—>7400Nand---->BC547/BC557 signal booster transistors ---->DRV101. (The problem here is the signal booster which runs fine in the simulation but in the realtime is floating on 4 Volt level. At the moment the PID tries to kick in, to steer the DRV101, the signal floats from the ground, at a voltage level of 4V. at the 4 Volt the signal changes a little in shape but looks like a compressed sine wave with little pwm jitter inside. Compared to the reality the simulation does not fit. To save on time i am posting it here and i am asking for your help.

What would be the most adviseable course to take, and can it easily be done?

on two suggestions until now i have my doubts.

1 How can i invert my PWM output signal with the arduino PIDLibrary

2 Keep the PWM Output signal and invert the signal and boost it to the right levels of the DRV101.

your help is greatly appreciated!!!

I think we need to actually see a schematic of your circuit rather than a vague functional description.

Sounds far too complicated. Why not just PWM a MOSFET? You might have to adjust
the PWM frequency by manipulating the timer registers, but you external circuit is just
a resistor and an n-channel logic-level MOSFET...

The DRV101 has very high losses (1V Vsat or so for an amp), does something you
don't need/want (turn on full, then switch to PWM after a delay), is hard to control (needs
analog control of PWM duty cycle), and is loads more expensive than a suitable TO220
MOSFET (its also an un-friendly pinout for breadboard / stripboard).