The problem here is that the display is driven using a timer interrupt. The temperature sensor, being a one-wire device, has to have very careful timing to communicate with it. As a result it disables interrupt so it has exclusive use of the CPU and can properly predict its timing while it's communicating. And so the display can't refresh at the same time.

So what can you do? Well, nothing really. You can't have perfect timing for the temperature sensor while at the same time interrupting it for displaying the value.

So there's really two main options:

1. Change the temperature sensor. Use an SPI or I2C temperature sensor that has a real hardware interface to communicate with it so you don't have to use big-banging.

2. Change the display. A non-multiplexed display or a display with its own built in driver (most often with some serial protocol like I2C or UART) would mean you don't have to use interrupts to update the display.

Basically you have two technologies that won't sit together well on the same chip.

The problem here is that the display is driven using a timer interrupt. The temperature sensor, being a one-wire device, has to have very careful timing to communicate with it. As a result it disables interrupt so it has exclusive use of the CPU and can properly predict its timing while it's communicating. And so the display can't refresh at the same time.

So what can you do? Well, nothing really. You can't have perfect timing for the temperature sensor while at the same time interrupting it for displaying the value.

So there's really two main options:

1. Change the temperature sensor. Use an SPI or I2C temperature sensor that has a real hardware interface to communicate with it so you don't have to use big-banging.

2. Change the display. A non-multiplexed display or a display with its own built in driver (most often with some serial protocol like I2C or UART) would mean you don't have to use interrupts to update the display.

Basically you have two technologies that won't sit together well on the same chip.

Update:

I missed the sevseg.refreshDisplay() function call since the website had scrolled it off the bottom of the code block. Bah!

Ok, so it's not using an interrupt. That could be a good thing, or it may be meaningless. It takes time to do the temperature reading, and it's that getting in the way of the refreshing.

There's a chance you may be able to push the update "into the background" by attaching it to a timer interrupt, but if the OneWire communication does disable interrupts then the problem will still persist (I don't have OneWire to hand to look at the code right now).