I know that in Biblical Hebrew (and presumably other varieties), we see the construct "noun adjective", with the noun unmodified, such as parah adamah (red heifer, parah = heifer). We also see the genitive (noun-chain) construct, "noun noun", which usually means "noun (of) noun", and in that case the first noun is modified: eishet chayil (woman of valor, eishet instead of ishah), birkat kohanim (blessing of priests, birkat instead of b'racha).
I recently noticed both olah tamid and olat tamid (burnt offering) in Numbers 28, but I don't see why the grammar would be different between the uses. The former is what I'm used to seeing, for example in Num 28:3:
וְאָמַרְתָּ לָהֶם--זֶה הָאִשֶּׁה, אֲשֶׁר תַּקְרִיבוּ לַיהוָה: כְּבָשִׂים בְּנֵי-שָׁנָה תְמִימִם שְׁנַיִם לַיּוֹם, עֹלָה תָמִיד.
But a little later we see (in 28:6):
עֹלַת, תָּמִיד--הָעֲשֻׂיָה, בְּהַר סִינַי, לְרֵיחַ נִיחֹחַ, אִשֶּׁה לַיהוָה.
I see the comma there (which was not in the book I first saw this in). Punctuation, of course, is editorial (and I understand that it is related to the trope, which is also not original). In both cases the text seems to be describing a continual olah (burnt offering). And later, in 28:10, we see another olat formation without any complicating punctuation:
עֹלַת שַׁבַּת, בְּשַׁבַּתּוֹ, עַל-עֹלַת הַתָּמִיד, וְנִסְכָּהּ
What distinguishes the cases that use the olah formation from the ones that use the olat formation? They all look the same to me, semantically; does the syntactic difference mean something?