Damping Ring RF Upgrade Review

Ron Akre, Patrick Krejcik




Details from Ron Akre … SLAC Memorandum

Klystron/Microwave Department Nov. 5, 1997

Electronic/Microwave Engineering Group (08D)

To: Marc Ross MS 66

From: R. Akre

Subject: Damping Ring RF Reliability


Below is my categorization of cater data supplied to me from Wayne for the damping ring caused down time between June 1, 1997 and Oct. 29, 1997.

Total items 23

Total days 151

Total hours 3624 SLD = 1384 hrs

Total Beam Time Lost 62.2 hours

Availability 98.3%

Klystron & Magnet 25 hrs

Cavity 590 14.8 hrs

S Band feedback 9.8 hrs

Rate change problems 1.9 hrs still not resolved

VCO 1.5 hrs

Pi Cavity Tune 1.4 hrs

Various Other 6.2 hrs

Wrong Shop 1.6 hrs

Lost Beam Time Hours
SLD OP YR / hrs Total RF KLY Cavity Low Level Stability Other % Available
97 1384 52 31 6 10 5 0 96.2
96 1417 48 24 4 2 18 0 96.6
94 2830 65 43 2 15 1 5 97.7
93 3026 57 23 8 21 2 4 98.1
92 1662 43 6 16 7 2 6 97.4


Current Inventory October 4, 1999 http://www.slac.stanford.edu/grp/kly/amrf/drrf/drhist.html
Tube Type
Vac Ion Pump
C7-08 VA-955H Varian VA-1955H S/N 0101R. NDR Socket Online No No Installed in NDR socket 10-22-97.
H4-159 VKP-7555S Varian VA-1955H S/N 0157 NDR vault Dressed Spare No No. Water leak repaired, magnet rebuilt
J7-11 VA-955H Varian VA-1955H S/N E7-011R SDR Socket Online No No Installed in SDR socket 11-29-98
J6-101 VKP-7855 Premier VA-1955H S/N 0123 SDR Vault Dressed spare Yes No  
C8-507 VKP-7855 None Sector 1 cage, in crate New tube Yes Until April 2000 Delivered to Sector 1 cage 4-1-98


Damping Ring Klystron and Klystron Magnet Failures

Jan 21, 1995 Tube C6-14 Water Leak SDR

Jan 23, 1995 Tube H7-11 Gun arc NDR

Mar 22, 1995 Magnet E7-011R Short SDR

Dec 11, 1995 Tube C6-14 Downtime Damage SDR

Jan 08, 1996 Tube C1-19 No emission, Collector Leak SDR

Jan 13, 1997 Magnet 0123 Interlock failed NDR

May 29, 1997 Tube C6-14 Collector Leak NDR

Jun 13, 1997 Magnet 0157 Water Leak NDR

Jun 19, 1997 Magnet 101R Unknown Problem(Installation) NDR

Sep 03, 1997 Magnet E7-011R Water Leak SDR

Sep 09 1997 Magnet 0123 Water Leak SDR

Oct 22, 1997 Magnet 0157 Water Leak NDR

Feb 11, 1998 Tube H4-159 Body OC Faults SDR

Feb 19, 1998 Magnet 0123 Water Leak SDR

Jun 01, 1998 Magnet 0157 Short-Near Old Repair SDR

Nov 29, 1998 Tube H4-159 Water Leak SDR

Magnet E7-011R Mar 22, 1995 repair unknown.

Magnet 0123 was rebuilt by Premier Jun 19, 1997

Magnet 0157 Received New from Premier Microwave Nov 07, 1995

rebuilt by Stangenes Jun 04, 1999

Magnet overheating due to interlock failure partly motivated the interlock upgrade.

Magnet Summary

The interlock system is being upgraded to better protect the magnets. The SDR interlock system has been upgraded and the NDR interlock system will be upgraded in the near future. Failures due to shorting are expected to decrease.

Magnet water leaks are suspected to be vendor design related. The new vendor has redesigned the cooling system to eliminate the problem areas.

Klystron Summary

The tube failures occurred with beam voltages of about 24kV. The collector life is reduced at this power level. We have implemented the following three items to improve reliability and klystron lifetime.

  1. Interlock to shut off HV power when RF is off for 5 minutes. Reduces the time all the beam power is going to the collector.
  2. Tubes with over 20k hours of operation are no longer rebuilt.
  3. The beam voltage has been reduced to 21kV, SDR, and 22kV, NDR.
(This is not an option for SLD type current.) Remaining Problem Areas

Tube H4-159 needs to be run to prove reliability. The tube has been removed in the past due to body current trips. These may have been due to undetected water leak.

Magnets E7-011R and 0123 have an unreliable water cooling circuit.

The water cooling system for the klystron magnet, body, and collector has been a source of down time. The system is worn out and needs replacement.

The water system is integral to the cabinet, which could also use redesign for maintainability.

Replacement of Klystron with New Type of Tube


Typical IOT Operating Points from Thomson TH770 Data sheet.

Cathode V 34kV

Cathode I 0.7A RF off = 23.8kW Beam

Cathode I 2.8A 63kW out = 96kW Beam = 65% Efficient

Full Power(63kW) Gain = 23.5dB Drive = 300W

Typical IOT Operating Points from CPI K2D110W Data sheet.

Cathode V = 36kV I = 2.2 A (80kW) Collector 47kW

28kW RF out avg. 110kW RF out peak

Gain 20dB minimum RF Drive = 1.2kW

Cost Savings

NDR Klystron 135kW Beam Power 25kW Average RF power

SDR Klystron 120kW Beam Power 19kW Average RF Power

Assume IOT output of 25kW @ 40% efficient = 60kW Supply.

Power Savings of 75kW NDR + 60kW SDR =135kW

1 year operation 46 of 52 weeks = 7k Hrs/yr.

= 1GWhr / yr. @ $0.22 / kWh = $220k / yr.


Single Tube Transmitter (High voltage, 34kV, power supply, 1.2kW drive amplifier, interlock system, IOT, all other tube support). All we do is hook up 480VAC power and LCW and we have a high power amplifier.

$400k to $450k each


Leave as is. $10k

System will be refurbished as required during maintenance down times. (New weir, klystron cabinet rails) High risk of failure in the water cooling system The cabinet needs to be refurbished to insure safety when replacing klystrons. Klystron support rails are worn.
Not Recommended

New Water cooling system and cabinet $82k

The klystron cabinet and water system can be built as a unit during a one-year period and then installed in a two week, per ring, down time. In the future this unit could be moved upstairs. Separate Septum water pump will increase flow. Lower flow and pressure in klystron body and magnet will increase reliability. Smaller footprint will allow more workspace.

New system can be moved upstairs in the future.

New cabinet will reduce klystron replacement time by 1 hour and reduce likelihood of errors.

Minimum Recommendation


Move existing klystrons upstairs $200k

The klystron cabinet and water system can be built as a unit and cables run during a one-year period and then installed in a four week down time. All the controls and monitoring for a new transmitter would be upstairs.

Will reduce down time associated with klystron related problems. Will save about 4 hours in diagnosing and replacing a klystron.

Will reduce calibration time for direct feedback loop and allow access to components during running.

With the system upstairs one less spare klystron and magnet would be required. Magnet $22k, Klystron $80k.

Preferred Recommendation


New transmitters - Single tube 30kW average $1.3M
Lead time about 1 year. Installation can be done upstairs during run. Change over would require 4 to 6 weeks depending on amount of upstairs prep. Total project time about 2 years. All the advantages of having the source upstairs Power savings of $200k/year. This is the reason klystrons have not been used in new designs for several years.

Future availability insured. Replaceable beam stick lowers replacement cost compared to klystron from $80k to $35k.

Tubes are water cooled – no steam system. There are more support components for an IOT than a klystron, 1.2kW drive amp and grid bias supply, however the TV industry requires high reliability.

Will need to review component protection from 130kW-peak power capability, or reduce the peak power capability.

Future Option




The reliability of the existing system has been greatly increased due to upgrades and lower power requirements. There is nothing currently in the future that will require higher power levels, 2 bunches at 5e10 e-/bunch.

Projections from Debbie Sharrah of CPI are that klystrons will be available for the next 5 years as a standard product and for the next 10 years as a special order.

Desired Plan

Redesign the klystron cabinet and water system. Near the end of the project determine if the manpower and down time is available to install the system upstairs, or if would be beneficial to install it downstairs.

Drill penetration into the NDR vault.

During the next several years we will keep updated on the TV transmitter market and review the reliability of the existing system. High power solid state drive amplifiers will improve during this time.


There are a number of refurbishment/replacement options