Manuscripts of the TR

When creating his Greek text, now called the Textus Receptus (TR) by many, Erasmus used just seven Greek manuscripts. Erasmus’ version in its first edition was called, Novum Instrumentum omne, diligenter ab Erasmo Rot. Recognitum et Emendatum, non solum ad Graecam veritatem verum etiam ad multorum utiusq; linguae codicum eorumq; veterum simul et emendatorum fidem, postremo ad probatissimorum autorum citationem, emendationem et interpretationem, praecipue, Origenis, Chrysostomi, Cyrilli, Vulgarij, Hieronymi, Cypriani, Ambrosij, hilaryj, Augustini, una cum annotatines, quae lectorem doceant, quid qua ratione mutatum sit. The original version was a bilingual Greek/Latin parallel edition with Greek in the left column and Latin in the right. It was originally intended to be a revision of Jerome’s Vulgate, a translation much hated by the KJV-Only crowd. The TR is not the early version of Erasmus’ Novum Instrumentum but is based on the later editions from the third edition onwards.
The codices used by Erasmus were (the names used are according to the Gregory-Aland system):

First Edition

Miniscule 1: a twelfth century manuscript containing the entire NT except Revelation. This forms part of f¹ or “Family 1” in conjunction with miniscules 118, 131, 209, 1582, and others. The order of books is Gospels, Acts, General Epistles and Pauline Epistles with Hebrews as the last. The Gospels are “Caesarean” text type while the remainder is “Byzantine”. Scrivener believed that the following verses came from this manuscript:

Mathew 22:28; 23:25; 27:52; 28:3.4.19.20;
Mark 7:18.19.26; 10:1; 12:22; 15:46;
Luke 1:16.61; 2:43; 9:1.15; 11:49;
John 1:28; 10:8; 13:20.

(A plain introduction to the criticism of the New Testament for the use of Biblical students, pg 382)

Miniscule 2: probably eleventh or twelfth century. This 248 page codex contains the full text of the Gospels. The text is Byzantine and contains numerous scribal errors. Von Soden placed this miniscule in Family K^x along with about fifty other late miniscules. Von Soden believed the family appeared around the tenth century and diverged from Family E. The TR is based largely on this family. Miniscule 2, along with the rest of the family are distinguished largely by the presence of the Pericope Adultera.
Miniscule 817: as with 2816 this codex is dated to the fifteenth century.
Miniscule 2814: dated to the twelfth century containing the book of Revelation and a commentary by Andreas of Caesarea (when he lived is unknown but he has been placed between the fifth century to the ninth century. This is the only Greek source of the book of Revelation in the TR and is missing the last six verses which Erasmus supplied with a back translation from the Vulgate.
Miniscule 2815: a Byzantine text-type codex dated to the twelfth century of 216 pages and containing Acts, the General Epistles, and the Pauline Epistles in that order. It is textually similar to 206, 429, 522, and 1891.
Miniscule 2816: a Byzantine text-type 287 page codex dated to the fifteenth century. The codex contains Acts, the Pauline Epistles, and General Epistles. Erasmus used some of the marginal readings rather than the main body text.
Miniscule 2817: a twelfth century Byzantine text-type codex of 387 pages containing all of the Pauline Epistles and ends on Hebrews 12:18. Estienne used this along with Codex Bezae and numerous others in his 1550 Editio Regia.

Second Edition Additions

The second edition of Erasmus’ text is the basis for Luther’s German translation.
Miniscule 3: a twelfth century Byzantine text-type codex of 451 pages containing all of the NT except Revelation. This codex does not contain the Pericope Adultera. As with Miniscule 2 Von Soden placed this codex in Family K.

Third Edition Additions

Miniscule 61: The first and second editions did not contain 1 John 5:7 called the Comma Johanneum, which was added on the basis of its inclusion in this fifteenth century miniscule also known as Codex Montfortianus. It contains the entire NT with the books in the order Gospels, Pauline Epistles, Acts, James, Jude, I and II Peter, I, II, II John and the Revelation. This codex is the first discovered Greek manuscript containing the Comma Johanneum. The manuscript was apparently copied from a tenth century manuscript held at Lincoln College in Oxford. The original codex did not contain the Comma which was added in this copy from the Latin.
The manuscripts are the main source of the text for the King James or Authorised Version of the Bible as well as the Luther Bible. The version of the Textus Receptus used was the Bezae version which, in addition to the manuscripts listed above, also used Codex Bezae or Codex Cantabrigiensis (D^ea or 05) and Codex Claromontanus (D^P or 06) both of which were in the hands of Theodore Beza at the time.
Cantabrigiensis: This 5th century uncial codex also known as Codex Bezae is 406 pages and is believed to have originally had 534 with Greek on the left page with Latin on the left. It contains most of the Gospels, III John, and Acts. There is evidence of at least nine separate correctors between the sixth and twelfth centuries. The order of books is Matthew, John, Luke, Mark, III John, and Acts. The Greek is Western text-type while the Latin is considered to be an old form of the Old (pre-Vulgate) Latin. One form of the longer ending of Mark is included, as is the Pericope Adultera while John 5:4 is omitted and Acts is about 10% longer than the TR.
Claromontanus: this 533 page Western text-type uncial codex is dated to around the 6th century AD and contains all of the Pauline epistles. In addition it contains a catalogue of the OT and NT omitting Philippians, I and II Thessalonians, and Hebrews and includes the Epistle of Barnabas, The Shepherd of Hermas, Third Epistle to the Corinthians, Acts of Paul, and the Revelation of Peter.
I am not sure if these are all the individual manuscripts that went into the TR as it was when the KJV was translated from it. It appears that contrary to the KJVO position the TR is actually based on a limited number of manuscripts – the rest (including many Byzantine manuscripts) mostly, but not 100%, in agreement with the 50-60 in Family 1 and Family K that directly support nearly all readings in the KJV.