STARTER always holds two versions of a drive's configuration: the offline project on your PC and the online parameter set in the drive itself. They drift apart whenever someone changes something on the keypad without updating the project. The comparison function — available once you are connected to the target system — lists every parameter where the two differ, which is the fastest way to answer "what did somebody change on this machine?". The comparison tells you what differs; deciding which side is correct is an engineering judgement, and getting it backwards either loses a genuine fix or restores a mistake.
- A project file is a snapshot, not a live mirror. It only matches the drive if someone kept it updated.
- Comparison is a diagnostic, not a repair. Read the list before deciding to upload or download.
- Some differences are expected — device name, IP, motor ID results on a replaced motor. Learn to recognise the benign ones.
- A comparison run before every change is the cheapest documentation habit in drive engineering.
Two Copies of the Truth
New users often assume the STARTER project is the drive. It is not. It is a file that was accurate at the moment someone last uploaded or downloaded, and nothing keeps it synchronised afterwards.
| Offline project | Online drive | |
|---|---|---|
| Lives on | Your PC, as a project folder | The control unit's memory |
| Updated by | Upload (Load to PG) | Download, keypad edits, BOP-2 changes, PLC parameter writes |
| Available when | Always, no drive needed | Only while connected |
| Represents | What was configured at some past moment | What the machine is actually doing now |
The asymmetry is the point. The drive has many ways to change; the project has one. So the project is almost always the stale copy, and it is the one you should suspect first when they disagree.
What Makes Them Drift Apart
| Cause | How it shows up |
|---|---|
| Keypad or BOP-2 edits during a breakdown | Ramp times, current limits or minimum frequency differ, with no record of why |
| A different engineer used a different project copy | Wholesale differences across many parameters |
| A motor identification run was repeated | Stator resistance and inductance values differ slightly |
| Drive replaced without updating the project | Device name, IP and serial-linked values differ |
| A PLC wrote parameters acyclically | Ramp or limit values change with no human involved |
| Macro changed on site | p0015 differs, and with it a large block of BICO interconnections |
The first row is the common one and the most valuable to catch. Someone fixed a problem at 2 a.m. by lengthening a ramp, the machine has run well ever since, and nobody wrote it down. That change is real engineering knowledge, and it exists only in the drive until a comparison finds it.
Learn to audit a drive properly
Hands-on STARTER training covering comparison, backup discipline and change control on live SINAMICS hardware.
Running a Comparison in STARTER
The comparison needs both copies available, so you must be connected to the target system with the relevant project open.
| Step | Action |
|---|---|
| 1 | Open the project you believe matches this drive |
| 2 | Connect to target system |
| 3 | Select the drive object and start the offline/online comparison |
| 4 | Let STARTER read the drive — this takes a moment on a full parameter set |
| 5 | Review the difference list before touching anything |
Two habits make this far more useful. Export or screenshot the difference list before acting, because once you upload or download it is gone. And run the comparison before you make any change of your own, so the list reflects what was already there rather than a mixture of somebody else's changes and yours.
Reading the Results
Not every difference is a problem. Sort them mentally into three groups.
| Group | Examples | Action |
|---|---|---|
| Expected | Device name, IP address, operating hours, serial-linked values | Ignore. These differ by design on a replaced or renamed unit. |
| Meaningful | Ramp times, current limit, minimum frequency, macro, control mode, motor data | Investigate. Someone changed behaviour deliberately. |
| Alarming | Motor nameplate data, telegram selection, safety configuration | Stop and establish which is correct before proceeding. |
A difference in motor nameplate data is the one to treat seriously. It usually means either the motor was changed and the project was not updated, or the project was downloaded from a machine with a different motor. Both are conditions that damage equipment quietly, so resolve them against the physical nameplate rather than against either copy.
Deciding Which Side Wins
The comparison gives you facts. The decision is yours, and there are only two options.
Upload — the drive is right. Choose this when the drive is running well and the differences look like deliberate on-site tuning. You are capturing knowledge that exists nowhere else. This is the safer default, because it changes nothing about the machine's behaviour.
Download — the project is right. Choose this when the drive has been disturbed, when someone made changes you can show were wrong, or when you are restoring a known-good configuration after a fault. This changes the machine, so it needs justification and a fresh upload of the current state first.
When you genuinely cannot tell, upload. A machine that is running correctly with undocumented settings is a documentation problem; a machine you have just downloaded the wrong configuration into is a production problem.
Whichever you choose, record the reason somewhere durable. The comparison list plus one sentence explaining the decision is worth more to the next engineer than any amount of tidy project structure.
Step-by-Step Lab: Find a Change Nobody Documented
Hands-on- A G120 on a training rig with a verified backup project taken from that same drive.
- Ideally a colleague to make the change while you are not watching — the exercise is much more realistic that way.
- Estimated time: 25 minutes.
Prove the two copies match
Open the backup project, connect, and run the comparison before anything is changed.
Have someone change three things from the keypad
Ask a colleague to change three parameters — a ramp time, a minimum frequency and a current limit — save with p0971 = 1, and tell you nothing.
Find them with a comparison
Reconnect and run the offline/online comparison again.
Export the list before you act
Save or screenshot the difference list. Then decide, for each parameter, whether the drive or the project should win — and write down why.
Sync in the safe direction
Upload the drive into the project, then re-run the comparison.
See what a macro change looks like
Change p0015 on the drive, save, and compare again.
You have the skill if you can hand your colleague the difference list and have them confirm it names exactly the parameters they changed, and if you can explain for each one whether you would upload or download and why. That judgement, not the button click, is what the comparison function is actually for.
Frequently asked questions
What is the difference between the online and offline project in STARTER?
The offline project is the file on your PC, accurate as of the last upload or download. The online data is what is actually in the drive right now. Nothing keeps them synchronised, and because the drive can be changed from the keypad, the project is usually the stale copy.
How do I find out what someone changed on a drive?
Connect to the target system with the matching project open and run the offline/online comparison. It lists every parameter where the two differ, with both values, which is far faster than checking parameters by hand.
Which differences are normal and can be ignored?
Device name, IP address, operating hours and serial-linked values differ by design on a renamed or replaced unit. Ramp times, current limits, macro selection and control mode are meaningful. Motor nameplate data, telegram selection and safety configuration should stop you until you establish which side is correct.
Should I upload or download after finding differences?
Upload when the drive is running well and the changes look like deliberate on-site tuning — you are capturing knowledge and changing nothing. Download only when you can justify replacing the drive's behaviour, and take an upload of the current state first. When unsure, upload.
Why did a comparison suddenly show dozens of differences?
Usually because the macro parameter p0015 was changed. A macro rewrites a whole block of BICO interconnections at once, so a single change produces a long difference list. Check p0015 first before investigating the individual parameters.