As I wrote, "This time the hydrogen didnt stay in the reactor."
In both cases, the hydrogen was a result of overheated zirconium encased fuel rods reacting with water.
A problem with the hydrogen bubble at Three Mile Island, which had nothing to do with an explosion, was the simple fact that it could displace water in the reactor, further exposing the core.
http://www.southernstudies.org/2009/04/post-4.html
[font face=Times, Times New Roman, Serif][font size=5]Investigation: Revelations about Three Mile Island disaster raise doubts over nuclear plant safety[/font]
[font size=3]...
The Thompsons aren't the only ones who have produced evidence that the radiation releases from TMI were much higher than the official estimates. Arnie Gundersen -- a nuclear engineer and former nuclear industry executive
turned whistle-blower -- has done his own analysis, which he shared for the first time at
a symposium in Harrisburg last week.
"I think the numbers on the NRC's website are off by a factor of 100 to 1,000," he said.
Exactly how much radiation was released is impossible to say, since onsite monitors immediately went off the scale after the explosion. But Gundersen points to an inside report by an NRC manager who himself estimated the release of about 36 million curies -- almost three times as much as the NRC's official estimate. Gundersen also notes that industry itself has acknowledged there was a total of 10 billion curies of radiation inside the reactor containment. Using the common estimate that a tenth of it escaped, that means as much as a billion curies could have been released to the environment.
Gundersen also offered compelling evidence based on pressure monitoring data from the plant that shortly before 2 p.m. on March 28, 1979 there was a hydrogen explosion inside the TMI containment building that could have released significant amounts of radiation to the environment. The NRC and industry to this day deny there was an explosion, instead referring to what happened as a "hydrogen burn." But Gundersen noted that affidavits from four reactor operators confirm that the plant manager was aware of a dramatic pressure spike after which the internal pressure dropped to outside pressure; he also noted that the control room shook and doors were blown off hinges. In addition, Gundersen reported that while Metropolitan Edison would have known about the pressure spike immediately from monitoring equipment, it didn't notify the NRC about what had happened until two days later.

...[/font][/font]