excitation_curve problem [SOLVED]

[jonathan dumas]

Dear Didier,

I use the excitation_curve command for the "qpole" field map that is in the "example" folder of TraceWin with the following .dat file:

DRIFT 0 150
;;;;;;;;;;;;;;;;;; First quad with excitation_curve command
EXCITATION_CURVE -1 -1
FIELD_MAP 70 480 0 200 10.0 0.0 0 0 qpole
DRIFT 100 150
;;;;;;;;;;;;;;;;;; Second quad with excitation_curve command
EXCITATION_CURVE 8888 8888
FIELD_MAP 70 480 0 200 10.0 0.0 0 0 qpole
DRIFT 100 150
end

there are 2 things that I would like to check:
1) When I use negative values for the excitation_curve command, instead of getting 10A in the DATA tab of TraceWin like I asked, I get 12A. And the field is indeed increased by 20%. I can put a negative current instead but, is there a reason for this?

2) The first command overrides every other excitation_curve commands if I use this field map again. For example, on the second qpole, the current is also going to be 12A even though I put different values for the command. I understand the benefits of this, because in principle you should have the same conversion current->field for a given field map. But if you don't, then it requires more complicated tricks. If I want different values, I need to copy my field map, give it a different name and only then the second excitation_curve command will be taken into account. I would prefer having to write several times the command rather than having to make several copies of the same field map. Is it possible to do something about it?

Thanks,

Jonathan

[Didier]

Dear Jonathan,

1) You enjoy moving off the beaten track . It seemed natural and obvious to me that the G(I) curve should be positive. Don't make any distinction according to the direction of connection of the power supply. So -1/-1 is forbiden.

2) Here, I don't understand this request, why change the G(I) if it is the same quad.
Technically speaking, reading the field maps and then calculating the modes to know what it is (quad, sex, hoctupole...) and then calculating the field integrals... takes time and is therefore only done once per type of quad. And it is at this time, that I can calibrate the G(I) with the obtained gradient. So I didn't plan that.

Regards,

Didier

[jonathan dumas]

Thank you for your answers Didier.
Fine! I'll use it as implemented!
Cheers,
Jonathan