Do I understand you correctly, you really expect the TM search to show you matches for a single word from a segment with dozens of words? If so, then you unfortunately misunderstand, how a TM works.
A TM will deliver you matches for WHOLE segments (similar by the percentage set in minimum match value to what has been translated in the past). A TM match will however NEVER deliver a match for a single word out of such a segment.
If you need matches for single words, use a Multiterm termbase.

But if you are referring to the search function, so it cannot find "hydraulisch" if you are searching for "hydraulische " (with a trailing space. Try to search for "hydraulisch", then it will find all words with "hydraulisch", including "hydraulische", "hydraulischer" and so on.

Same will apply to the concordance - if you search for a word with a trailing space, this will reduce the amount of what can be found, as the space is also searched for. Try to select words without trailing spaces instead...

Well, I am not searching for the whole segment, dear Jerzy. And you really dont need to suspect me in being a total idiot (thanks again:/). There has always been an option, at least in 2014, of marking a word or a phrase in a source text and then pressing F3-button. By that this word or phrase was placed into a concordance search string and the program looked for all TUs containing this word. Please study the screenshot carefully and you will see that there is only one word in the search string above.
What happens now is that F3 works, the word is being placed into the concordance search but the fuzzy matches are not appearing.

What I need to find out is how to make this search less sensitive. There should be an option. And from my experience, I never needed to do this juggling with endings to find the same rooted words. It is also quite time consuming. I was used to Studio showing me an exact match first and after fuzzy matches, so that I could find what i needed in a split second.

I would really appreciate your help on this issue, if you know, of course

Actually, I think you should be fine with the defaults, which are 70 and 30. If you had the minimum match value set to 100% that would cause only strings that are identical to the search string to be found.

The only thing is, that as Jerzy has suggested, the trailing space in your second search is probably causing the issue and not allowing other partial matches to show, because they don't actually match the string (with the trailing space).

As a suggestion, I see that you're not using a termbase, and for things like these terms, a termbase might be more helpful than Concordance, and it wouldn't require any intervention on your part, the term would be recognized and indicated for you as soon as the segment becomes active. 

Unknown said:

Didnt know it has become such a complicated issue in the new Studio

It hasn't really... it's the same as before but Fragment matching is new so I was just suggesting you see if this will help you.  It's capable of picking out single words from a TM source and matching it with the appropriate target.  So may be the sort of enhanced feature you are looking for.

yes sure, here

Hi Maria Pyatkova 

While you're trying to figure this out I thought this may be helpful... if nothing else give you some confidence that this is working with your sort of content in Studio 2017:

Hi Maria,

Something happened to me just now that made me think of your problem. I have a project received from a client that had the option to use different TMs for the specific language pair enabled. I didn't realize this initially and added my TMs under All Language Pairs. It wasn't until I realized the TMs didn't seem to be updating that I checked this option and it was enabled. Unchecking the box fixed the issue. Could this also be checked in your case?