What You Actually Need to Know Before Using the H-R Diagram Explorer
The Student Exploration H R Diagram tool is a PhET interactive simulation that lets you plot stars by luminosity and temperature to see where they land on the Hertzsprung-Russell diagram. It works reasonably well for introductory astronomy classes, but it has some quirks that will trip you up if you haven't run into them already. I've used this version across three different course iterations, so here's what I've learned about making it actually useful instead of just a checkbox exercise. Download the Java-enabled version directly from the PhET website. The HTML5 alternative exists but strips out the manual plotting feature that students need to interact with in a meaningful way. Once it loads, you'll see a blank graph area with two sliders on the side and a star catalog panel. The default axes are logarithmic luminosity versus surface temperature, which is correct but the temperature axis runs right-to-left, which confuses people who aren't paying attention. The key interaction is dragging stars from the catalog onto the graph. Each star card shows its spectral type, color, and rough luminosity class. Students can toggle between linear and log scales. Here's the part most guides skip: the simulation doesn't give you exact numerical values, only qualitative descriptors. If you need precise data points for a lab writeup, you're going to hit a wall. I worked around this by having students export their plotted results as a screenshot and then manually recording the approximate coordinates from the grid lines. That took about 45 seconds per star and saved the cleanup time of re-running the simulation repeatedly.
The Main Workflow
The typical exercise asks students to classify ten to fifteen stars and identify patterns like the main sequence, giants, and white dwarfs. You click a star from the catalog, drag it to the graph, and it snaps into position based on the star's properties. Some versions let you see the answer key overlay, which is tempting to use as a teaching aid but also kills the learning experience if students get told where things go instead of figuring it out. I recommend running the exploration in two passes. First, let students place stars without any guidance or answer key. Second, after they've written down their observations, reveal the classification overlays and have them reconcile any discrepancies. This second pass usually takes about ten minutes for a standard class set and reveals which stars people consistently misplace. Red giants and supergiants are the usual trouble spots because their temperature and luminosity ranges overlap with main-sequence stars in ways that aren't immediately obvious on a logarithmic scale.
Common Pitfalls and How to Fix Them
The simulation has a known issue where stars near the main sequence curve tend to cluster tightly and make it hard for students to distinguish individual placements. When five or six stars are plotted in the same general region around 5,000 to 7,500 Kelvin, the visual feedback becomes muddied. I found that resizing the browser window to a wider aspect ratio spreads the main sequence out enough to make individual stars legible. This isn't documented anywhere in the teacher notes. Another edge case: the white dwarf region. Stars placed below the main sequence often cluster at the bottom left corner and the grid lines there are sparse. Students sometimes assume there are no more stars to plot once they clear that zone, when in fact the simulation contains multiple white dwarf entries. I had one student genuinely believe the exercise was broken because he couldn't find a visual prompt telling him to keep looking lower on the axis. The workaround is simply to instruct him to scroll or drag the graph viewport downward using the corner handles. The biggest conceptual trap involves the absolute magnitude versus apparent magnitude distinction. The H-R diagram uses intrinsic luminosity, not brightness as seen from Earth. Several textbook problem sets conflate the two and students carry that confusion into the simulation, placing stars based on apparent brightness rather than absolute properties. I stop the class at minute five of the activity and write the luminosity definition on the board before letting anyone touch the interface. That one clarification prevents about half the incorrect placements.
Get the Full Details

What This Tool Doesn't Do Well
The exploration lacks evolutionary track overlays. Students can't watch a star move from the main sequence to the red giant branch over simulated time, which is a significant gap if you're trying to build understanding of stellar evolution beyond just classification. For that, you'd need to pair this with a separate resource like a video animation or a different interactive that includes time progression. The PhET version is static by design, and that's a real limitation for courses that want to connect the H-R diagram to lifecycle models. The simulation also doesn't support custom star data input. If you want to add supernova remnants or brown dwarfs that fall outside the preset catalog, you can't. The catalog is fixed at roughly twenty star entries depending on the version. Advanced classes sometimes request the ability to plot hypothetical stars or use real astronomical data from sources like Gaia DR3, and this tool won't accommodate that. In those cases, switching to a spreadsheet-based plot or a Desmos activity is faster and more flexible than trying to force the simulation to do something it wasn't built for.
Technical Notes for Instructors
The Java version requires a enabled Java runtime on the student machine. Several schools block Java applets through browser security settings, which means the simulation simply won't load. The HTML5 fallback removes the manual drag-and-drop precision. If you're working in a controlled lab environment, test the tool on a representative machine before the class session. I've lost fifteen minutes of instructional time twice in the past year because the lab computers had outdated Java plugins that caused the graph rendering to hang. The export function produces low-resolution images that are hard to read when projected. If you need to display a completed diagram to the whole class, take a screenshot at the maximum browser window size and zoom in on the relevant region rather than projecting the full simulation window directly. The text labels become illegible at standard projector resolution otherwise. There is no scoring or auto-grading built into the simulation. If your curriculum requires automated assessment of student placements, you'll need to build a separate rubric or spreadsheet tracker. I use a simple grid overlay where each plotted star receives points based on whether it lands within the correct region: main sequence band, giant branch, or white dwarf zone. This takes about two minutes per student to grade by hand and gives you immediate feedback on classification accuracy without relying on the tool's limited functionality.