I had a similar question, which was answered previously. However, it differs in usage of Pandas package with it.
Here is my previous question: matplotlib - No xlabe
I just got the same issue because I was mixing plots made with matplotlib and made with Pandas.
You should not plot with Pandas, here is how you could replace:
pd.Series(range(10)).plot(ax=ax42)
with
ax42.plot(pd.Series(range(10))
As Scimonster mentioned above for me it worked when I plotted all the pandas before creating twinx axes.
I had few plots in twinx which were also coming from Pandas Dataframe objects (x,t plots). I created separate lists before starting the plot and then used them to plot after plotting the first pandas plots.
to summarize my work flow was 1. Creating lists for twinx plots 2. opening plot and plotting all pandas plots with normal axes. 3. creating twinx axes 4. plotting lists on twinx axes
fortunately, this flow is working for me