EDIT 2020-04-03:

--recreate-pods --wait is not recommended anymore. As Jorden pointed out one way is to add a checksum annotations that will implies to restart the pods if any file change. see https://helm.sh/docs/howto/charts_tips_and_tricks/#automatically-roll-deployments for reference doing so.

ORIGINAL ANSWER

To complement the answer of @stan-bondi, you can do :

helm upgrade --recreate-pods --wait -f new_values.yaml nginx-controller nginx-controller

This is often needed when you juste changed a configMap or secrets that wont be detected as a change in the release itself.

This is how I update the current chart with new values, without upgrading the chart version:

helm upgrade --reuse-values -f values.yaml {release-name} {release-path} --version {fixed-version}

For example:

helm upgrade --reuse-values -f prometheus/values.yaml prometheus-operator stable/prometheus-operator --version 5.7.0 --namespace monitoring

I use a fixed version of the installed chart, and add --reuse-values flag to ensure that I keep the previous values I used.

You can do:

helm upgrade -f new_values.yaml nginx-controller nginx-controller

This will update the revision of your chart which can be viewed using:

helm ls

or more specifically:

helm get nginx-controller

For helm chart upgrades, check this link from the docs: https://docs.helm.sh/helm/#helm-upgrade

In the Deployment(or StatefulSet) yaml file, and if you are using ConfigMap or Secret you can add a checksum like below:

kind: Deployment
...
spec:
  template:
    metadata:
      annotations:
        checksum/config-env: {{ include (print $.Template.BasePath "/configmap-env.yaml") . | sha256sum }}

...

This will detect a changed in the configMap that wont be detected as a change in the release itself.