The short version. Before placing the cursor on a photopeak, zoom in. At full scale the whole spectrum is squeezed across the display, so each pixel covers several keV and you cannot read the peak to better than that. Use the Expand control to narrow the view to the region around the peak, where each pixel is under a keV, and place the cursor there. Then make sure every operator does it the same way — that one habit removes most of the scatter from a peaking log.

What the spectrum view shows

The View Spectrum display plots counts against energy across the full range of the analyzer — thousands of channels of data. The photopeak is the hump; for Cs-137 it sits at 662 keV, for Co-57 at 122 keV. To “peak” the detector you place a cursor at the top of that hump and read its energy, either to confirm the calibration or to log the peak position for QA trending.

Why the full-spectrum view lies to you

Here is the part most operators never think about. The display is only so many pixels wide. When it is showing the entire spectrum at once, those thousands of channels are squeezed into that pixel width — which means each pixel on screen represents many channels, and several keV of energy. You simply cannot read a peak position to single-keV precision from a view where one pixel is wider than the precision you are after. Two careful people, both placing the cursor “on the peak” in the full view, can legitimately come away with readings several keV apart — not because either made a mistake, but because the view does not resolve any finer.

The fix: expand before you read

Every Genesys-platform counter has an Expand control in View Spectrum that narrows the displayed range to a slice of the spectrum — say, the few hundred channels around the photopeak. Zoom in like that and the same pixel width now covers a much smaller energy range, bringing the per-pixel resolution down to under a keV. Place the cursor on the peak in the expanded view and your reading is suddenly precise and repeatable.

The discipline that matters most. If two technologists share the QA, agree on one method and write it down: expand to the same range, place the cursor the same way, read the same value. Most of the well-to-well or week-to-week scatter people attribute to the instrument is really operator-to-operator variation in how the peak was read. Standardize the technique and your peaking log gets quieter overnight — without touching the detector.

What a healthy photopeak looks like

A healthy photopeak is a single, smooth, roughly symmetric hump sitting where it belongs in energy, with a clean low-energy region beside it. When you are peaking for QA, you are mostly confirming three things: the peak is where it should be (correct energy), it has a normal shape (one clean hump, not split or badly skewed), and it is well clear of the noise. A peak that is in the right place and a normal shape is a healthy detector — a few keV of day-to-day wander in its position is normal and expected.

Two shape clues are worth knowing. A peak that has become much broader than usual points to a resolution problem. A peak that appears split into two close humps is the classic signature of a cracked crystal. Neither is something a peaking habit fixes — but spotting them on the screen tells you when a reading problem has become a detector problem.

What to do

  • Always expand before reading a peak position. Never log a peak channel from the full-spectrum view.
  • Zoom to the same range every time so successive readings are comparable.
  • Standardize between operators — one written method, same cursor placement, same value.
  • Read the shape, not just the position — one clean, symmetric hump is healthy; broadening or a split peak is a detector question.
  • Expect a few keV of wander in a healthy peak; don’t chase it.

Further reading