Linac Marx Modulator Update Matt Gardner Trevor Butler

Linac Marx Modulator Update Matt Gardner & Trevor Butler PIP General Meeting 23 March 2016

54 Cell Prototype • Marx Cabinet & Cells – Internal assembly of cell support structure completed. – Output circuits / hardware still need to be installed. – Cells are being assembled. • Capacitor assembly and frame/heat sync assemblies are complete for all cells. The two assemblies are being put together. • • Charging Supplies – All 24 supplies have been delivered, with 17 tested so far without any failures. – Charging supply cabinet was delivered. – All internal wiring needs to be designed and installed – Summing plate still needs to be designed and built. Controls – Controls crate has been assembled with all cards except the FPGA control board. – FPGA control board is still being designed. – Crate needs to be wired, a few days worth of work. – Crate needs a power supply. • The 28 cell prototype controls crate borrowed a supply from the EE Support kicker group, which is highly over spec’d for this use. • A new supply must either be built or purchased 2 Presenter | Presentation Title 1/15/2022

54 Cell Prototype Cabinet 3 Presenter | Presentation Title 1/15/2022

New Cabinet for Capacitor Charging Power Supplies 4 Trevor Butler | Linac Modulator Testing 1/15/2022

LRF 1 Marx Modulator Testing all 4 cavity regulation schemes • Digital Feedforward (Open Loop) – – • Digital Learning (Pulse to Pulse) – – – • Compares actual cavity to desired cavity signal after going through lead out network Works along side Analog Feed-Forward to flatten cavity fields Also activated on entire flattop Digital Beam Control (Feedforward and Learning) – – 5 Voltage minus Current is proportional to cavity gradient error Set gain to minimize the cavity pulse to pulse error Activates on flattop and held right before beam enters cavity since this effects the response of the tube Plan to convert this to a Howie defined “JITR” algorithm (Just In Time Regulation) which will use the cavity gradient error sampled early in the pulse to ramp up the PWM values to compensate for pulse to pulse variations. (54 cell design) Analog Feed Back (Closed Loop) – – – • The pulse to pulse wave shaping the is used to flatten the cavity fields Will shape the rising edge of the waveform and slope of the PWM to keep the cavity fields flat Will not account for pulse to pulse changes which can be as large at a few percent on LRF 1 without addition control Analog Feed-Forward (Closed Loop) – – • Implement R & L or Proportional and Differential response assuming a specific gain of the 7835 (R / Proportional) and response of the cavity (L or Differential) Looking into the best way to make changes to these parameters after a tube change out occurs without calling in experts Presently used the LLRF Feed-Forward command to switch in Beam Pulse from Marx Controls Turns on special cell to learned voltage Adds additional signal to PWM to learn wave shaping required to flatten the gradient Ideadlly we could use beam start time and width information to replace the present Trevor Butler | Linac Modulator Testing 1/15/2022

LRF 1 Marx Modulator Progress since last report • • • Goal since last meeting – Get the Marx Modulator system operational for long term testing Resolved Driver Low Problem – We think it is the fact that the modulator was not shutting off after a cavity spark, causing instabilities in the 7835 output power, which feedback into the driver power level. Resolved problem when going from short to long pulses where the learning would not update correctly. Installed new Ross Voltage Divider and interfaced with Marx Controls Installed HV Ross Relay on output of modulator before the RG 220 Output cable – Unless we can assure limit of 1000 mrem/hr of xrays that could be emitted from the cavity , we need to implement a two independent methods of turning limiting power out of the modulator – Worked with safety to implement a tunnel access interlock that both grounds out the HV capacitor charging supplies and the output of the Marx modulator when activated. • • 6 Wired in HV On, which is used to turn on the Driver Anode HV power supply, in order to trip off Marx modulator when Driver or other subsystem trip off. Installed new Humidity Sensors on 7835 and 4616 Cavities Testing Interlocks between Marx and Linac control system Added wiring for Driver Start Pulse, Driver Stop Pulse, & LLRF Gating Pulse to provide signals required to operate the drive without the A 5 controls. Trevor Butler | Linac Modulator Testing 1/15/2022

LRF 1 Marx Modulator Latest Results • • 7 Ch. 1 LLRF Beam FF Signal Ch. 2 Modulator Voltage (5 k. V/V) Ch. 3 Gradient error Ch. 4 Unused Trevor Butler | Linac Modulator Testing • • Ch. 1 Modulator Voltage (5 k. V/V) Ch. 2 Feed forward error Ch. 3 Gradient error Ch. 4 Beam timing pulse 1/15/2022

LRF 1 Marx Modulator Pulse to Pulse Stability • • 8 Decreased the standard deviation of pulse to pulse variations from. 16% to about 0. 04% We also measured an overall gradient variation using an oscilloscope of around 1% on tube modulator and about 0. 25% on Marx modulator. Trevor Butler | Linac Modulator Testing 1/15/2022

LRF 1 Marx Modulator Next Steps to Full Operation • • • 9 Steps remaining to finalize switchover on LRF 1 Run for the next two weeks with the present system to look for any failure modes that may occur in the modulator. Ideally, resolve any issues without switching back. Build analog reference circuit into analog comparator board that could be used to regulate the capacitor charging supply voltage. Update code to reflect changes in beam learning during turnoff advance Document present setting on both waveform and system timing Finalize switch over to driver On/Off controls to enable replacement of present Modulator control system, which is being partially hijacked to enable operation of the Marx using its interlock and driver level commands Work on building spare parts, control cards, power supplies, etc. to enable us to have an system can could run in this mode and be able to repair parts instead of switching right back. Need to determine best way to remove requirements of keeping the old modulator parts, such as Crowbar Cabinets, power up to make up the required interlocks Will need new LOTO procedure written and new interlock configuration explained. Continued training of Linac and Operational personnel of system functionality. Trevor Butler | Linac Modulator Testing 1/15/2022

End – Start of running 10 Trevor Butler | Linac Modulator Testing 1/15/2022