PeakWright / Guides / File will not open
When a ChemStation file will not open
Seven specific failures, what each one actually means, and which of them can be recovered. Work down the list in order, because the cheap checks are at the top.
1. You are trying to open the folder
A .D is a folder, not a file. Nothing opens it because there is nothing to open.
Go inside and find the .ch or .uv file.
On a Mac a .D can appear as a single item. Right click and choose Show Package
Contents. Full map of what is in there:
what is inside a .D folder.
2. Unsupported file type
Every .ch and .uv starts with a number naming its layout. A reader
that does not recognise the number should stop rather than guess, and a good one tells you
which number it found.
| Number | What it is |
|---|---|
| 30, 130 | .ch, delta encoded body. UV, CAD, ELSD. |
| 179, 181 | .ch, plain 64 bit floats. FID style. |
| 31, 131 | .uv, diode array data. |
| 8, 81 | Older .ch. Handled by chromConverter, not by every reader. |
This message means the file is intact and this reader does not cover that version. Try another one before you give up, because the tools cover overlapping but different sets. The tools compared page lists what each one handles.
A reader that stops here is doing the right thing. The alternative is to guess at the layout and draw a chromatogram that looks plausible and is wrong, which is worse than no chromatogram in a way that is very hard to notice.
3. Too small to be a data file
The header alone runs to several hundred bytes, so anything smaller cannot be a run. In practice this means one of:
- A zero byte file, from an acquisition that failed or was stopped at the start.
- A truncated copy, from a transfer that was interrupted.
- A placeholder or lock file that happens to share the extension.
- A cloud storage stub that has not been downloaded yet, which looks like a file and is not.
Check the size on disk first. If it is a few kilobytes when the run should be hundreds, you are looking at a copy problem, not a format problem.
4. The .D folder arrived empty
This one is not recoverable. An empty .D means the folder was
copied by something that treated it as a single file, or a sync client moved the wrapper
without the contents. There is nothing inside to read.
Go back to the source and ask for it again, zipped. Zipping is the fix and it also prevents the next occurrence, because an archive is genuinely one file and cannot lose its contents in transit.
5. It is an archive, not a run
A .dx is an OpenLab archive holding what a .D folder would otherwise
spread across many files. A .zip containing a .D is the same idea by
hand. Neither is a data file, and a reader that expects a run will refuse both.
Extract it and open what is inside.
6. It opens, and the numbers look wrong
The worst case, because nothing errors. Two things produce it, and they look different.
The shape is right and the axis is nonsense
Scaling. The body of a .ch holds raw counts that have to be multiplied by a factor
stored in the header, and different file versions keep that factor at different offsets. A
reader looking in the wrong place produces a perfect peak shape on a meaningless axis.
Check the units on the y axis and check a peak height against a value you know. If the shape is right and the magnitude is out by a constant, this is it.
The shape itself is wrong
A misread body. With delta encoded files this shows as a trace that drifts steadily or steps the wrong way, because an error early in the decode is carried forward through every point after it.
This is a reader bug rather than a file problem, and the way to confirm it is to open the same file with a second tool.
How we check ours. Every value this reader produces is compared against the published output of rainbow, an independent open source reader, over eleven real files covering every version handled. All 1,487,470 values match exactly, and the check runs on every change. That is not a guarantee for a version nobody has tested, but it is the reason to trust the ones that are.
7. The time axis is wrong
Retention times that run to the wrong end point usually mean the start time, end time or sampling rate was read from the wrong place, or reconstructed by assuming a step rather than read from the file.
The check is simple: the last time value should equal the run length in the method. If your twelve minute run ends at 4.7 minutes, the time axis is being built rather than read.
Still stuck
Two things that genuinely help. First, try a second reader, because they cover different versions and a refusal from one is not a verdict. Second, if you have any access to the original ChemStation, ask for a CDF export, which every tool in the field reads.
If you have a file that nothing will open, we would like to see it. The file type number and the instrument that wrote it are the two facts that make a new version supportable, and a sample file makes it certain.
Try the file here
If it is a version we handle, it opens. If it is not, you get the file type number it found, which is the fact you need to take anywhere else.
PeakWright is not affiliated with or endorsed by Agilent Technologies. "ChemStation" and "OpenLab" are their trademarks.