August 2023: Re-Testing Magnet Probe Signals at CRYRNG Extraction

Table of Contents

TL;DR

Actual kicks can be detected without problem. Getting precise timing on the rising edge is only possible at low energy. Realistic tests at high energy were not possible, most likely due to an hardware issue at the kicker itself.

Introduction

We repeated the tests from June 2023 in more detail.

Measurements

Testing the kicker was done at the following settings

• element hydrogen-like gold @ 4 MeV / u
• 511 kHz revolution frequency, T_rev is 1.956 us
• kicker length 'kick dauer' set to 0.8 * T_rev, ~1.565 us
• extraction kicker deflection at 4 and 11 mrad
• 4 mrad correspond to about 3239 V 'high voltage' and 124V 'low' voltage'
• 11 mrad corresponds to about 7786 V 'high voltage' and 233V 'low voltage'
• pattern CRYRING_demo_2023_02_esr_prepost_fast

As described here, the b2b system uses dedicated technical system to tap the magnet probe signal at CRYRING (see here). The main new feature is a modification to the comparator module that allows to set the comparator level remotely. The main focus here is on understanding the setting of the comparator level. The procedure has been the following.

1. set comparator level
2. measure delay of probe signal at comparator output relative to CMD_B2B_TRIGGEREXT
3. repeat this for different values of the comparator level

Please note, that due to the 'air-coiled' tap, the signal to the comparator is very weak can not be compared to the signal to the digitizer. Moreover, the electronics of this comparator dedicated to CRYIRNG only allows to measure the first rising edge of the probe signal.

Figure: Top: Signal 'through tap' to Digitizer. Bottom: 'Tapped signal' to comparator. Measurements using an oscilloscope in the CRYRING cave in June 2023.

Low Energy

Here, the kicker was operated at 4mrad, which corresponds to about 3.2kV at one power supply.

A measurement of the comparator response was done. For each comparator level, about five extractions from CRYRING were done and the values of the measured delays averaged. The total measurement time was about half an hour.

delay [ns]	comparator level [%]
1209	42.00
1234	42.25
1237	42.50
1242	42.75
1246	43.00
1252	43.25
1254	43.50
1258	43.75
1262	44.00
1273	44.25
1282	44.50
1291	44.75
1315	45.00
1320	45.25
1358	45.50
1457	45.75

Table: Measured delay as a function of comparator level.

The result was plotted and compared to the probe signal acquired by the Digitizer YRCD002 (YR06MK1:CurrentTriggered@Raw).

Figure: Signal of digitizer (blue) and data of the comparator (red). Top: Complete signal. Bottom: Zoom at rising edge. The x-axis shows the delay after the trigger signal CMD_B2B_TRIGGEREXT. Left y-axis scale: Signal of comparator [a.u.]. Right y-axis scale: Comparator level [%] (the comparator level can be set within a 12bits range).

The figure above shows the signal acquired by the digitizer and the data measured by the b2b system. First of all it is remarkable, that the data acquired by the comparator yields a meaningful curve. Second, the comparator seems to be about 100 ns 'faster' than the digitizer. An obvious explanation of the latter effect are different cable lengths. The electronics of the b2b system is integrated into the rack of the kicker electronics, while the digitizer is located at different location in the CRYRING cave, connected by a long copper cable (a delay of 100 ns corresponds to about 20 m of copper).

Looking at the data in the figures above, a safe working point at this kicker energy is a comparator level of about 44% (out of the 12bit range).

Figure: Histogram of 245 extractions at CRYRING. x-axis: offset from the b2b kicker-trigger event,y-axis: number of occurrences. Details see text.

To test magnet probe signals can been detected reliably, the comparator level was set to a fixed value of 44%. The 2-sigma (95% CL) value is 2ns only and the window defined by the maximum outliers is 6 ns only.

High Energy

Here, the kicker was operated at 11 mrad. Here, the power supply configuration is a different one.

delay [ns]	comparator level [%]	remark
1282	42.00	only occasional comparator values
2216	42.50	unstable, delay jumping
2800	43.00	unstable, delay jumping
3001	43.50	unstable, delay jumping
3510	44.00	stable
3535	44.50	stable
3561	45.00	stable
3589	45.50	stable
3605	46.00	stable
3678	46.50	only occasional comparator values

Table: Measured delay as a function of comparator level.

Figure: Signals of digitizer (blue) and data of the comparator (red). The cross-iconed comparator values indicate very unstable conditions from shot-to-shot.

In June 2023, this measurement at high-energy has been a difficult one, see here. Again, this data by the B2B system was not able to successfully detect the rising edge of the magnet probe signal. Although, the figure above was kind of 'glued together', the data by the b2b seem to be even delayed with respect to the digitizer signal.

As the setup for high energy is exactly the same as for low energy (no modification), the explanation is not obvious.

Moreover the signal acquired by the digitizer is looking strange. Instead of a flat-top, it can be seen that the first rising edge does not go up to full energy but is 'interrupted' at about half the desired energy; the flat-top looks more like a steep rising ramp.

However, during a visit in the CRYRING cave yielded the audible observation of a 'sparking' noise (like a discharge) from the kicker cabinet which might be due to a hardware defect. This would explain the signal shape by the digitizer easily. As a speculation, the comparator electronics by the b2b system might not have analyzed the magnet probe signal from the 'tap' but the electronic noise from the spark/discharge which happened in the same rack just about 50 centimeters away.

Looking back at the data from June, the problem might have occurred in June already. The intention is to repeat this measurement after the kicker hardware has been checked.

Kicker Issues

Pulse Shape at High Energy

When operated at high energy, the shape of the magnet probe signal did show a steep ramp instead of a flat-top. Moreover a 'sparking' noise was heard inside the kicker cabinet. This needs to be investigated.

Pulse Length

In ParamModi Tab 'Extraction' the value of 'Kick Dauer' was set to 0.8, which is supposed to result in a pulse length of 0.8 * T_rev which should be approximately 1.6 us. Instead, the length of the pulse shown by the Digitizer App was about twice as long, 3 us. There are two possibilities,

• either the kick is too long (model, LSA makerule, FESA, kicker electronics),
• or there is problem of data acquisition (wrong timescale at picoscope, ... )

Conclusion

The new setup implemented to detect the magnet probe signals at the CRYRING kicker has been tested again. The results at lower energy are nice, the first rising edge is safely detected. The results at high energy might have been hampered by an hardware issue of the kicker and shall be repeated.

-- DietrichBeck - 07 Sep 2023