Send, sent; end, *ent?

The absence of any immediate answer to this interesting question confirms my belief that it is not a subject which lends itself well to a Q&A site such as this. The history of English verb forms is a complex subject and each of the verbs you mention would merit a reply in itself. To give an idea of what might be involved, the OED records the past tense of send as appearing in the following forms between its first appearance in Old English and the 15th century: sende, seonde, sænde, sænte, sennde, sente, seende, send, sont, sent, sendet, sendyd, seended and sended. In addition, Bruce Mitchell points out in his ‘An Invitation to Old English and Anglo-Saxon England’, that in Old English it could also have past tense sendede. He notes that ‘the d of the ending –d(e) is not absorbed into the root’. Such a feature may be one of the clues to understanding how similar verbs, if not necessarily this one, developed the forms they have today.

Shakespeare used both "blent" and "blended" as the past tense of "blend". For example, from Twelfth Night,

Where every something, being blent together,
Turns to a wild of nothing, save of joy,
Express'd and not express'd.

and from Troilus and Cressida,

This blended knight, half Trojan and half Greek.

From Barrie England's citation of the OED, the past tense of "send" wasn't fixed until the 15th century. And people still say "on bended knee". I'm not sure that there's any reason other than pure chance why some of these verbs ended up regular and others didn't. For many of these verbs, it certainly seems to have happened much later than the transition from Old English to Middle English (circa 1100).

In fact, Old English had multiple conjugation patterns even for weak verbs (verbs ending in a dental suffix), and it seems that the differences were relevant to the development of irregularities in the past-tense forms of some verbs. Send and end did not originally form their past tenses in the same way.

Send, bend, rend had contracted past-tense forms already in Old English: the OED shows forms like sende, bende, rende as being used in Old English, and marks sended, bended as Middle English, not Old English, past-tense forms. (The past participle shows different behavior.)

The important thing is the division between Class I weak verbs and Class II weak verbs. Class I weak verbs contained a derivational suffix [(i)j~i] in Proto-Germanic. In the past tense, this was subject to deletion after a heavy verb root by the time of Old English:

In the past tense forms, a mid vowel (phonetically schwa) appears before the preterite /d/ in verbs with a light root but not in verbs with a heavy root

(p. 510, "The Dental Preterites in the History of English", by Aditi Lahiri)

The verb bend is from a Class I weak verb, bendan

The Oxford English Dictionary gives the following summary of the historical forms of the verb bend:

Forms: past tenseOE–ME bende, ME–15 bend, ME bente, ME– bent, 15– bended. past participleOE bended, ME y-, i-, ye-bent, ME–15 bente, 15 y-, i-bente, bende, ME– bended, bent

The OED's Proto-Germanic etymology shows -ja- in the infinitive, a characteristic of this class of verbs:

Germanic *bandjan, < bandjâ- ‘string, band’

The Old English past-tense form bende shows vowel deletion/contraction and coalescence of the dental suffix with the final consonant of the root. But in the Old English past participle (ge)bended, deletion did not originally occur because the /d/ was word-final rather than being followed by a vowel as in the past-tense form. In Middle English, there seems to have been influence between the past participles and past tense forms that resulted in all verbs of this type eventually having a leveled paradigm. (Aside from a general tendency towards leveling these two forms, the loss of schwas in ME would have eliminated the original phonological motivation for having contraction in the past tense but not in the past participle.)

The process leading to the devoicing that we see in the modern past-tense and past-participle form bent (and in a number of other verbs) is more obscure. The linked Lahiri paper attempts to explain it, but I don't fully understand the explanation given.

The verb end is from a Class II weak verb, endian

Verbs of this class originally had a derivational suffix containing a vowel that was *ō in Proto-Germanic. The OED etymology of end shows this characteristic o in the given Proto-Germanic etymon:

< Old Germanic *andjôjan , < *andjo- end n.

For some reason, verbs of this class have an -i- in the Old English infinitive form, but that's not relevant to the development of the past tense.

The past tense of Class II weak verbs was originally formed with this vowel before the d. The vowel was regularly subject to shortening because of the lack of stress on it (resulting in Old English past tense forms ending in -ode, with short o) and it was frequently subject to weakening/fronting/raising (resulting in past tense forms with ed(e) and id(e)). But this ō-derived vowel was not subject to deletion as early as Class I verbs were subject to deletion of their [j~i]-derived vowel.

Of course, verbs with a regular weak conjugation in Modern English don't always come from Class II weak verbs (many regular weak verbs ending in d or t come from either borrowings or part-of-speech conversions after the Middle English period).

And weak verbs with irregular past tense/participle forms don't always come from Class I weak verbs. However, this division is important for understanding how verbs ending in d and t with irregular weak past tenses developed, and also for understanding how verbs ending in other consonants that show irregular vowel shortening in the past tense (e.g. keep/kept).