We compare an SFTPRO1 field in SFT-only supercycle vs SFT+LHC supercycle from MBI Dataset - SFTx3___SFT_LHC1. We also subtract the eddy current contributions, as modeled on the B-Train. The results are similar, and the main effect is seen on the flat top.
When MD1 is not used, we see much larger field offsets, as well as effects caused by eddy current decay. This is from the extensive dataset acquired during Dedicated MD 2025-04-02.
So it seems that when MD1 is not pulsed, the field decreases by 1e-4 T at injection, which translates to an increased orbit (below transition). This is in agreement with what was observed during the Spontaneous MD 2025-07-08.
On the other hand, it was observed in the same MD that the field without an MD1 (in an LHC-type supercycle increased), which is not compatible with the previous observation.
The green, purple, gray and red lines is with a correction applied by the hysteresis compensation to try and reach field (though not reached successfully).
Debugging the B-Train
One hypothesis is that the marker is not applied.
According to the B-train the marker is applied at C1635 +/- 150 ms, wheras SFTPRO crosses 1100 G at C1707. Conversely the MD1 marks at C80 +/- 150 ms, and crosses around 100 ms.
I double check the marker by increasing the threshold on the MD1, which causes the field to drift more by SFTPRO injection, and one sees a large field jump (~1 G) at injection.