==> Next meeting: tuesday, june, 2nd, 1998, 14:00 h
 

Presents:  G. Bianchetti   CERN
                H. Breuker      CERN
                I. Crotty       CERN & CALIFORNIA
                G. Dumont       CERN
                C. Fabjan       CERN
                P. Farthouat    CERN
                A. Fucci        CERN
                Z. Hajduk       CERN & CRACOW
                B. Hallgren     CERN
                R. Hammarstrom  CERN
                M. Hass         CERN & WEIZMANN
                P. Jarron       CERN
                M. Letheren     CERN
                M. Newcomer     PENNSYLVANIA
                H. Reithler     AACHEN & CERN
                R. Richter      CERN & MPI
                G. Stefanini    CERN
 

R. Richter: Sources for n-irrad.: reactors, accel., dedicated n-sources

For the simulation of irradiation conditions in the LHC-halls, reactors are producing too many gammas together with the neutrons, due to thermal
neutron capture.

N-irrad. at CERN is bases on neutrons leaking from shielding with a relative low dose rate.- Access is ony possible on scheduled accelerator stops.

In contrast to this the facility PROSPERO in France is a n-source specially designed for testing electronics. Only fast n's are produced in this facility which is operating like a fast-breeder of very low power (3 kW). As there is no n-moderation, gammas are only coming from the fission process, which gives a low dose of 8 krad together with 10**14 neutrons.

As this correspond closely to the situation in the LHC-halls, irrad. should be done at this facility.- The facility also provides dosimetry, easy access to the irrad. area with cables at about 20 m distance from the control room.

One day is sufficient to collect 10**14 n/cm**2. Lower dose rates can be chosen, as in general there is only one user at the same time. Cost of about 35 kFF per day are charged to the users.

It was decided to plan for a first irradiation still before the summer break and to start with a test of commercial DC/DC converters. Converters have to be powered during the test, voltage and current have to be constantly monitored.- Other devices can be tested at the same time, e.g.: RASNIK CCDs, because absence of gammas is crucial for n-sensitivity tests (see last meeting), CAN-bus components, etc.

A number of participants agreed to participate in preparation and execution of the program.

Main tasks are: selection & procurement of DC/DC converters, pre- and post-rad tests, preparation of control equipment for the test. A. Fucci said that the electr.-pool can participate in selection, tests and preparation. P. Jarron, G. Stefanini and R. Hammarstrom would participate in the procurement of PSs. This list is not exhaustive.

C. Fabjan commented that a consistent n-irrad. program was important for the qualification of many components to be used in the LHC-halls and therefore should start as soon as possible. Given the relevance of the results for overall planning and costing of electrical services, the projected expenses are justified and necessary. The cost charged by the PROSPERO-facility seems to be reasonable, given the level of service offered.
 

P. Jarron: Progress on COTS. Results from rad-tol voltage regulators

A number of CMOS families have been qualified as rad-tolerant.

Tests of the VICOR DC-DC converters showed failure below 100 krad, therefore not suited for ECAL. However, CEA offers rad-tol converter. It
seems to be crucial that opto-couplers be avoided in the feedback loop.

Annealing technics for CMOS devices were discussed and first impressions were communicated from the ESA database on COTS. Data of the DB are
earlier than 1994, but an update is to come soon. The DB can be installed on every PC.

COTS selection strategies from NASA were also discussed.

LHC strategy: design total system for failure robustness: failure of one device should not DAMAGE others. E.g.: Power supplies may and will
fail, but are not allowed to give fatal overvoltage.

Rad-tol voltage regulators:
There seems to be good progress on the rad-tol voltage regulator program with SGS-Thomson. There is now a bipolar process with vertical NPN and PNP
transistors which stand more than 10**13 n/cm**2 and more than 1 Mrad. In the next phase practical devices will be designed. Prototypes will be
available in nov. 98.

The specifications address 4 issues at LHC:
        - hardness in total dose, neutrons and SEE
        - low dropout voltage
        - remote sensing
        - interface to slow control system

Specification details:
        - output curr.: 1 A and 3 A
        - output volt.: + - fixed and adjustable:
                - fixed: 2.5V to 8V
                - adjust.: 1.5 to 11 V
        - dropout: 0.5 V @ 1 A
        - overcurrent digital output, inhibit digital input
        - remote sensing
        - reasonable target price

Pierre also showed a number of rad-tol instruments offered by LETI, e.g. a CCD camera.
 

G. Stefanini: power supplies and FE electronics in the CMS cavern

Giorgio focussed on VMEbus crate power supplies and DC-DC converters. DC-DC converters operate at sw. freq. of 100 kHz to 2 MHz, cost goes from
1 $/W up to 100 $/W (space grade). Key element is frequently a PWM IC - e.g. Unitrode. Useful information can be found on:
http://www.unitrode.com/products/powsup/powsup-sg.htm

The technic of the feedback control loop can be crucial for rad-tol: it can be done by passive transformer, isol. ampl. modulation or opto-couplers.

Six commercial DC-DC converters were compared in more detail: 3 from industry-, 3 from space-standard. A list of other vendors was mentioned.

There was a comparision between switching vs. linear PSs. Switching PSs usually need external filters to get output ripple below 50 mV. Radiated
EMI in addition needs good passive shielding. Cost for sw. PSs with filtering is typically 2.2 $/W.

Other interesting web-sites on DC-DC are:
http://flick.gsfc.nasa.gov/radhome/papers/b013098.pdf

A detailed discussion was devoted to the test reports on the Cassini spacecraft. Here electronics has to be tolerant against the on-board radio-isotope thermoelectric generator. Opto-couplers proved very sensitive to radiation, which showed up in a degradation of the CTR (current transfer ratio). Current supply as well as output voltage were strongly influenced by displacement damage (n-irrad.). SEE were also discussed.

Giorgio concluded that rad-risks in the caverns are very real. Degradation will most likely show up after several years, when considerable luminosity is accumulated. The original components will then be no more available.- Careful screening of devices and circuits is necessary. The risk of SEE is substantial and events with high ionisation can be predicted by simulation software. CMS simulation experts are looking into this.

Giorgio undelined the importance of in-built protection circuits and load protection against overvoltage. Industrial standards should be respected w.r.t. transient voltage suppression and crowbar switches.

Discussion on SEE (Single Event Effects):
There was some debate, on how much highly-ionizing events can be expected from the n-spectrum, which shows a peak at about 1 MeV and a second, 10 times smaller peak, at 100 MeV. Only the latter class of n's seems to have sufficient energy for a nuclear reaction with MANY CHARGED particles in the outgoing channel. There are few energetic n's,  however, and cross-sections are small. Therefore few SEE should happen in the cavern during the LHC lifetime.

For these few events it would be sufficient to prevent damage to the load. Destruction of the circuit could be tolerable at very low rate, if the circuit is accessible.

Would experts who disagree please comment!
 

R. Hammarstrom: sensitivity of DC-DC converters to rad. and mag. field

Robert described options for the internal structure of DC-DC converters in conjunction with linear post-regulation (for lowest EMI) and discussed the resulting power efficiency. He showed that normally opto-couplers are used in the design of the feed-back loop and in the coupling to DCS. A degradation of the CTR (current-transfer-ratio) in the feed-back loop may lead to a loss of control over the output voltage and may damage the load. An alternative design was shown, using a "differential" opto-coupler, which could considerably reduce this risk.

Robert pointed out that new ferrit-types saturate at much higher magnetic fields and therefore would work without problem in the < 1 kG fields in the cavern. Frequency, however, is limited to < 100 kHz.

Robert finally gave a summary of tests for full cavern qualification of DC-DC converters. He also gave a detailed list of active and passive components to be qualified, including e.g. safety equipment on mechanical installations.
 

B. Hallgren: Progress Report on TCC2 test status

Bjorn described the TCC2 radiation area in LAB II, where the accelerator division has started to test CERN recommended field-buses. The group is
chaired by Raymond Rausch SL/CO. A typical dose of 50 Gy (5 krad) and 5 * 10**10 n/cm**2 can be obtained in a 6 week running period.

Bjorn said that the first irrad. period was over, during which a dose of about 15 Gy has been accumulated. No damage has yet been observed on the CAN-bus equipment (cf. minutes of last meeting). There will be more irrad. periods during this year, the next one starting june, 15th.

Bjorn gave a list of DCS-components to be tested for rad-tol, including voltage references, opto-couplers and complex ICs. He described a new low-power, low-cost IC for temperature sensing and a new self-calibrating OP-amp from TI, which both may be interesting for application in DCS devices.
 

==> Next meeting: tuesday, june, 2nd, 1998, 14:00h
 

Robert Richter