This is just one example: In the word "father", there is the interdental voiced fricative. However, in Old English, the word is fæder with a voiced alveolar stop; it is also fader in Middle English. The German word is Vater which has the voiceless stop from the voiced stop, which makes sense if the original sound was a voiced stop, but the Gothic word fadar and the Old Norse word faðir both contain fricatives. And I've never read anything about the High German Consonant Shift affecting anything twice. There are also other words like that, such as the word "word". So is the voiced stop the original sound or was there some amount of th-stopping in Proto-West Germanic?
Th-stopping of original Proto-Germanic voiced /d~ð/ to /d/ in all contexts is normal for Old English. It seems to be a common feature of West Germanic languages. The modern-day /ð/ in "father" is due to later changes.
There's a relevant ELU question here: /ð/ → /d/ shift in English
When researching my answer to it, I found Dental fricatives and stops in Germanic: deriving diachronic processes from synchronic variation, by Bridget Smith 2007 which I think gives a good summary of the development.
Basically, Proto-Germanic is reconstructed as having a series of voiceless plosive phonemes *p *t *k, a series of voiceless fricative phonemes *f *þ *h, and a series of voiced phonemes that neutralized the distinction between plosive and fricative (written sometimes as *b *d *g and sometimes as *ƀ *đ *ǥ).
The voiced series derived for the most part from the PIE breathy-voiced/aspirated voiced series, and from Verner's Law voicing of the PIE voiceless series (which in non-Verner environments became Proto-Germanic voiceless fricatives).
There is some uncertainly about whether Verner's Law applied after Grimm's Law (voicing the fricatives [ɸ θ x] to [β ð ɣ] which then experienced fortition in some cases to [b d g]) or before (voicing the plosives [p t k] to [b d g], distinct in some way from the reflexes of PIE *b *d *g, which then experienced lenition in some cases to [β ð ɣ]).
The reflexes of the voiced series in German are basically always plosives; voiced /b/ and /g/, and /t/ resulting from devoicing of earlier /d/.
The reflexes of the voiced series in Old English are more dependent on the environment and the specific phoneme. As far as I know, *đ became /d/, realized as [d], unconditionally (e.g. hēafod < *hauƀuđą). *ƀ became /b/ word-initially, after nasals and when geminate, and merged with the reflex of *f elsewhere (intervocalically to the voiced allophone [v], and word-finally to the voiceless allophone [f]). *ǥ stayed a single phoneme (ignoring the split of velars into palatal and non-palatal forms), but with allophonic variation between [g] and [ɣ], and possible word-final devoicing to /x/. (You can see a more complete summary of the sound changes from PIE up to English here: https://digilib.phil.muni.cz/bitstream/handle/11222.digilib/131584/Books_2010_2019_072-2014-1_12.pdf?sequence=1; be aware that it uses "ʒ" to represent [ɣ] for some reason)
In Old Norse, it seems that Proto-Germanic *đ became [ð] intervocalically, but Wikipedia indicates that this should be considered an allophone of /d/.
The phone [ð] occured in Old English as an allophone of /þ/. In the development to modern English, there are a number of cases of substitution of [d] with [ð], especially before -er (as Yellow Sky mentions). Smith says this is usually attributed to analogy or influence from Scandinavian languages. The reverse change of [ð] to [d] also occurred in some words, and there are a number of words ending in -der that did not undergo this change and retain Old English /d/.
I'll reproduce from the ELU post the list of affected words that I know of:
OE d [d] > ModE th [ð]:
- OE mōdor > mother
- OE fæder > father
- OE gad(e)rian > gather
- OE tōgædere > together
- OE weder > weather
- OE hider, hwider, þæder/þider > hither, whither, thither
OE þ [ð] > ME d [d]:
- OE spīþra > spider
- OE rōþor > rudder
- OE morþer/morþor/morþur/morþre > murder (this word may have been influenced by French forms)
- OE byrþen/berþen > burden (unusually, this one does not end in *-er)
OE d > ModE d:
- OE hlǣd(d)er > ladder
- OE blǣdre > bladder
- OE nædre/næddre > adder
- OE mædre/mæddre/mædere > madder (the plant)
- OE fōdor > fodder
- OE ūder > udder
Spelling with -th- (15c.) reflects widespread phonetic shift in Middle English that turned -der to -ther in many words, perhaps reinforced in this case by Old Norse forms; spelling caught up to pronunciation in 1500s (compare mother (n.), weather (n.)).