dissenting.
Although I agree with the majority that the post-conviction court ruled that the shackling issue was not raised by petitioner’s defense counsel on direct appeal,11 disagree with *49its conclusion that it necessarily follows that petitioner is entitled to a new trial as a matter of law. In my opinion, once the post-conviction judge determined that the trial record did not establish a sufficient factual basis for shackling, he should have continued the hearing to determine whether shackling was, in fact, necessary. If he had found that shackling was not necessary, he should then have granted petitioner a new trial. If, however, he had found that shackling was necessary, as the original trial court subsequently found on remand, he should have denied petitioner post-conviction relief.
Common sense dictates that a procedural error, such as the trial court’s failure to designate properly in the trial record the factual basis supporting his order to shackle petitioner, does not require a new trial when it is later shown at an independent factfinding hearing that the procedural error did not prejudice the defendant. The substantive issue is whether petitioner was unnecessarily shackled. A new trial is not required to make that determination.
A similar approach was followed by the Supreme Court in State v. Brewton, 238 Or 590, 395 P2d 874 (1964). The defendant, who had been convicted by a jury of murder, contended that the trial court had erred in admitting his oral confessions “for the reason that the admissions were not voluntary and were secured from him in violation of his right to *50due process of law under the United States Constitution.” 238 Or at 593. The Supreme Court ruled that, in view of the then recent United States Supreme Court opinion in Jackson v. Denno, 378 US 368, 84 S Ct 1774, 12 L Ed 2d 908 (1964), the trial court had erred in admitting the confession without first “fully and independently” resolving the issue of voluntariness. However, rather than simply remanding for a new trial, it allowed the state the opportunity to elect between a hearing by the trial court to determine whether the defendant's confessions were voluntary or a new trial.
The court stated that, if the state elected to prove that the confessions were voluntary,
“the trial court shall hold a hearing and determine whether [the defendant’s] confessions were voluntary. If the court finds the confessions were voluntary, it shall make an appropriate finding and enter a new judgment of conviction based on said finding and the verdict heretofore returned by the jury.” 238 Or at 604.
If the trial court were to find that the defendant’s confessions were not voluntary, it was instructed that the state was to be granted the option of retrying the defendant or releasing him from custody. See also State v. Addicks, 28 Or App 663, 560 P2d 1095 (1982).
I see no reason why it would have been inappropriate for us on direct appeal to remand the case to the trial court for a similar factfinding hearing to determine whether shackling was necessary, if petitioner’s defense counsel had raised the issue on appeal.2 Therefore, because the post-conviction judge was fully authorized to conduct such a hearing himself, I believe that he should have done so. Accordingly, I agree with the majority that the post-conviction judge erred in remanding the case to the trial court to conduct the requisite factfinding hearing.3 However, because petitioner did receive a full and fair hearing by the original trial court on remand from the *51post-conviction court,4 I would hold that the post-conviction court’s error was harmless.
Although it is not raised by either party, in my opinion the trial court’s error was harmless. In other words, petitioner’s defense counsel did not err in failing to raise the shackling issue on appeal.
In State v. Schroeder, 62 Or App 331, 338, 661 P2d 111, rev den 295 Or 161 (1983), we determined that the trial record failed to show a substantial necessity for requiring the defendant to wear leg shackles at his trial. However, because of the overwhelming *49evidence of guilt, we concluded that the error was harmless beyond a reasonable doubt and affirmed the defendant’s conviction. I believe that the trial court’s error was harmless in this case for the same reason, as evidenced by our recitation of the state’s case in petitioner’s appeal:
“Defendant was an inmate at the Oregon State Penitentiary when two prison guards were assaulted by an inmate wielding a metal pipe. Waterman, one of the guards assaulted, testified he was on the fourth tier of the prison cell block preparing to lock the inmates in their cells. He was accosted by two inmates; one had an ice pick-type knife and the other had a metal pipe. Waterman later positively identified defendant as the inmate with the pipe. The other inmate ordered Waterman to give him the cell block keys. When Waterman attempted to grab the inmate, defendant struck him with the pipe two or three times. He fell down the stairs to the landing on the third tier. As he rolled over onto his back, defendant was standing over him with the pipe and struck him two or three times. Waterman grabbed defendant’s pants leg. Beers, the victim of the other assault, testified that he heard a warning whistle and ran to the third tier of the cell block. He saw defendant striking Waterman with a metal pipe. He grabbed defendant, who turned and struck him two or three times with the pipe. Defendant was finally subdued. Another guard who was on the fourth tier witnessed the assault on Waterman and identified defendant as the assailant with the pipe.” State v. Guinn, 56 Or App 412, 414, 642 P2d 312 (1982).
To the extent that this conclusion is inconsistent with our holding in State v. Kessler, 57 Or App 469, 645 P2d 1070 (1982), I would overrule that opinion.
The post-conviction judge, who was from a different judicial district, was assigned to hear this case. Although he implicitly concluded that further factfinding was required, he should have done it rather than delegate that task to a different court.
The issue on remand to the trial court was not the trial judge’s credibility, as the majority seems to imply. Rather, the issue was whether a sufficient factual basis existed at the time of trial to support the judge’s order that defendant remain shackled. Therefore, because the trial judge had no independent knowledge of the need for shackling before the issue was raised at trial, his testimony was not required on remand to determine whether shackling was necessary. Furthermore, the post-conviction judge determined that the trial judge was not prejudiced against defendant.