Logbook from 1 January 2003 to 25 July 2004

Logbook from 1 January 2003 to 25 July 2004

9-Jan-2003

Now SXR2 has three diodes, 16, 18 and 20.

22-Jan-2003

New amplifiers build by M.Reyfman (MST electronic workshop) and installed in the SXR2 (fixed version) probe; all 12 diodes connected. These amplifiers are the 3-gain amplifiers with the simple external toggle switch. These amplifiers were designed with a MFB, 3-pole low-pass Butterworth output filter, 100kHz cutoff; an additional 5.6pF is present in feedback when gain=10^8.
The amplifier gains are 10^8 for diodes #1,3,18,20 and 10^7 for diodes #5,7,9,10,11,12,14,16.

30-Jan-2003

All the Femto channels are working again (cable of diode 3 of SXR1 probe was shorted, now repaired).

21-Feb-2003

New amplifiers and Femto ones tested together: they seem very similar. Amplifiers performance can be checked looking at the MST data of 20-feb (lots of high current plasmas).

27-Mar-2003

Amplifiers of SXR2 disconnected. Work on encoder started.

3-Apr-2003

Amplifiers hooked up to SXR2 diodes. The encoder was ready but not enabled.

10-Apr-2003

Encoder on SXR2 signals enabled. This should allow to record the amplifier gain (for Mikhail's amplifiers) in the digitized signals. A circuit has been built that provides a pulse of different voltage (height) depending on the gain (e.g., 100, 200, and 300 mV for gains 10^6, 10^7, and 10^8). The pulse is short (500 microseconds), and it can be added to the amplified signals with each plasma (and before each plasma begins). Thus, one can tell for sure what gain on has set for each channel by looking at the digitized signal.
At Mikhail's request, the triggering setup for the SXR array digitizers has been changed (by D.Craig). In the past, it was triggered with the usual trigger pulse of TTL level, 5 microsecond square pulse. Mikhail was interested in "encoding" the gain setting of the amps with a 50 (or 500) microsecond square pulse that has a height which corresponds to the gain setting. The simplest way to do this was to trigger the digitizer with a 50 microsecond gate pulse, rather than the normal 5 microsecond one. Then the trigger signal is teed off and sent to his circuit which produces the gain setting waveform. So, if you want to change the trigger time for the digitizers, you must now change the 2 element array in the trigger node of the tree. The 2 elements are for start and stop time; now they are -5.0 and -4.5 ms.

11-Apr-2003

Arrived in Madison.

14-Apr-2003

The SXR2 probe has signals up to 4 times bigger than the SXR1 probe, and the brightness profile is asymmetric in a strange way. Checked all the electronics and layout of the probes. The SXR1 probe is in its correct position (inserted in the 75 degree porthole, with the VAT valve opened, with the pinhole 10 mm from the inner wall of MST); the probe was slightly rotated around its axis (about 10-15 degree, counterclockwise looking at the probe from behind, or at 11 o'clock), but this does not introduce big difference in the signals. The probe was rotated to its normal position (poloidal 0 degree, with the fan of lines of sight in a poloidal plane, diode 1 looking HFS, diode 20 looking LFS) after shot 1030414134. Cables were checked for both probes, and no problems were found. Both the probes has the same gain, 10^8 for diode 1,3,18 and 20, 10^7 for diode 5,...,16. The iso amplifiers areworking and were calibrated last 6 March, with values of the gain equal to 1.0 +/- 0.3%. All digitizers (Joerger TR612) have the same voltage range, both input and output one. All these tests should exclude an electronic problem.

15-Apr-2003

SXR1 probe disconnected from MST after shot 1030415038. The probe is in good condition, there are signs of arching but nothing serious. On the other hand, the pinhole is damaged, there are signs of erosion and it is wider on one side, along the toroidal direction (see picture; if you look at the probe, installed, from behind and standing on the second deck of MST, the pinhole is wider towards left, that is towards decreasing toroidal angles and the FIR). The distance of the IRD array from the edge of the open probe has been measured and is 10.45 mm (in July 2002, before inserting this new probe, with curved Be filter, it was 10.35 mm; we think that when taking the measurment the IRD was slightly pressed on its socket, and this should explain the 0.1 mm difference).
Be filter checked: it shows signs of heating, and maybe of coating; a few colored rings are present on the surface exposed to the plasma. The filter was measured using a micrometer (precision 0.0001"=0.00254 mm), and the thickness is >18 micron (when measured in Padova, July 2002, it was about 16 microns). This may be due to the coating. Be filter reinstalled in its frame.
Pinhole replaced with a new one, of 0.8 mm of poloidal width; this should have been used to test the difference in the tomographic reconstruction when using smaller pinhole (and so smaller cone of sight, and better spatial resolution), and was prepared in Padova two weeks before.
All these measurements have been collected in this page.
The diodes have been checked with a meter; all similar, no one is shorted.
SXR1 probe installed in MST, extracted, VAT closed, at the end of the shot session (after shot 1030415095).

16-Apr-2003

A few tests on the SXR1 signals: diode 1 and 3 seem to have different noise level; the VAT valve was still closed, and remained so all the day. The following is the list of the tests.
1030416013: reference shot
014: switched inputs of Femto of diodes 1 and 3, SXR1
015: switched cables of diodes 1 and 3, SXR1, on the flange of the probe
017: all connections are as usual, like shot 013
018: the short tube covering the SXR1 plastic flange and the copper shield from that tube to the metal box containing the Femto have been removed. 60 Hz present on the signals, diode 1,18 and 20 more, the others less evdent
019: switched inputs of Femto of diode 1and 3, SXR1
020: the cable connecting the inner copper shield of the SXR1 diagnostic to the flange, and thus to MST, has been removed. No changes observed
021-028: an alligator put the ground of the cables of the SXR1 diode to MST. Before doing this the ground of the diodes got to MST in a strange way, through the tube, the copper shield and the Femto metal box. It is possible that during past tests on this diagnostic the ground of the diodes remained disconnected. After this the 60 Hz disappeared
029: diode 20, SXR2 powered by batteries. No changes
030: diode 20, SXR2 powered with power supplies; input of this diode disconnected
031: inputs of Femto of diodes 1 and 3, SXR1, and diode 20, SXR2 disconnected
032-033: cables hooked up to amplifiers as usual
034: a BNC T is used to put the ground of diodes of SXR1 to MST. This T is hooked up to to diode 3 Fischer feedthrough in the flange; the cable to Femto is hooked up to one end of the T, the other end has a cable going to a screw of the flange (the BNC of this cable has not the centre). From this point on signals are grounded to MST using this cable
035: switched inputs of Femto of diode 1 and 3, SXR1
036: inputs Femto of diode 1 and 3, SXR1 restored. Now these two diodes have other cables (those used in this probe at the beginning, Nov 2001). The signals are now noisy, +/- 2 V
037-039: original cables used, no noise
040: replaced cables of diodes 1 and 3, SXR1, with new ones, BNC-BNC RG58, with an adapter BNC-Lemo 00. Noise level on both signals similar. Diode 16, SXR1 has greater noise than before.
041: moved the Lemo connector of diode 16, SXR1, on the flange. Now the noise level decreased to the usual value. This is an old problem with this kind of connector-fedthrough: sometimes they have bad contacts
042: replaced the BNC-Lemo cable of diode 16, SXR1 with another one: same noise. The cables have to be checked, and possibly not touched after that
043: the original cable has been hooked up for diode 16, SXR1. Diodes 1 and 3, SXR1 have Femto to 10^7: noise level decreased
044: Femto of diodes 1 and 3, SXR1 to 10^8, noise increased
045: original cables of diodes 1 and 3, SXR1 used (those used before shot 040). Diode 3, SXR1 has too low noise, perhaps is broken
046: replaced the cable of diode 3, SXR1 with another one, BNC-BNC plus Lemo adapter, now it is ok
047: gain of Femto of diodes 18 and 20, SXR1 to 10^7, nosise decreases
048-049: gain of Femto of diodes 18 and 20, SXR1 to 10^8. Diode 18, SXR1 has greater noise than before
050: moved the connector on the feedthrough of diode 18, SXR1, and noise level decreased to usual values. Maybe diode 20, SXR1 has too much noise
051: moved the connector on the feedthrough of diode 20, SXR1, but noise did not changed. This is the best configuration: do not touch it anymore!
052: VAT of SXR1 diagnostic opened, and probe inserted. No changes in noise level.
The conclusion of these tests is that when disconnecting or connecting cables on the flange of the SXR1 probe, a few checks of the noise level should be performed. This can be done directly looking at the overload led of the Femto: if, swiching the gain to 10^9, this led lights up and remains on, then move the corrispondent connectors on the flange until the led goes off. This should be repeated on all connectors.
Another results is that the old cables BNC-Fischer, with yellow plastic on the connectors, should not be used, since they have bad connection in the feedthrough.

17-Apr-2003

PPCD+pellets shots @ 200 kA. In general, for all shots SXR signals are low; a good example is shot 54.
PPCD started at shot 004; from shot 058 started the injection of the pellets, before PPCD (between 10 and 12 ms).
SXR1 and SXR2 gain: diodes 1,3,18 and 20 have 10^8, diodes 5,...16 have 10^7.
In shot 1030417007 diode 9 and 10, SXR2 behaviour bad, the signals seem clipped at voltage greater than 5.5 V.
The brightness profile of SXR1 is still asymmetric, but flipped horizontally: this is because the Be filter was mounted flipped, on 15 apr. Then the main problem in SXR1 probe is the Be filter! This is probably due to a coating. And maybe also the SXR2 filter has the same problem, apparently less important.

18-Apr-2003

PPCD+pellets shots @ 200 kA. Low SXR signals; SXR1 and SXR2 gains same as yesterday.
300-400 kA shots, no PPCD and no pellets from shot 035 to 038. A few PPCD @ 400 kA at the end of the day, starting from shot 039 to 048. Shot 045 is an example of QSH.
Several diodes of SXR1 and SXR2 saturate in shot 039 (a 400 kA PPCD shot, and the two probes had still the gains for 200 kA PPCD). From shot 040 gains are 10^7 for diodes 1,3,18 and 20, 10^6 for diodes 5 to 16, both SXR1 and SXR2 probes.
SXR1 probe dismounted and Be filter replaced with a new one. This is a 18 micron foil (measured in Padova and here at MST with a micrometer), circular; to get it rectangular, 10x16 mm, a cutter was used. No problem in doing this, and the thickness did not changed after this operation.
The distance between the IRD array and the edge of the open probe has been measured, and it is still 10.45 mm (the same as 15 apr). The coating in the front part of the probe has been estimated in 0.1 mm (measured the distance betwwen the surface of the top cover and the pinhole foil, before and after scratching it with a screwdriver). This is indicative of the order of magnitude of the coating in the Be foil. See here for further details.
SXR1 probe mounted on MST, extracted, VAT closed. Rough pumped all night.

19-Apr-2003

PPCD @ 400 kA.
Shot 1 and 2 have SXR1 extracted and the VAT closed. After shot 2 the VAT has been opened and the the probe inserted in its measuring position (137 mm).
From shot 3 to 20 the gains are 10^8 for diodes 1,3,18 and 20, and 10^7 for diodes 5,...,16, both for SXR1 and SXR2.
Diode 20, SXR1 did not worked from shot 3 to 6, due to the old problem of the Fischer connectors and feedthrough. After moving the connectors, starting from shot 7, diode 20, SXR1 worked fine.
Shot 1 to 20 are 400 kA sustained shots; good examples are shot 6,8,16 and 18. SXR1 brightness profile is in general greater than SXR2 one, and is not asymmetric anymore.
Form shot 20 the VAT valve of the three filter SXR diagnostic (for Te measurements) has been opened (before it was closed, but not to the end, to allow pumping).
Shot 21 to 175 are 400 kA PPCD shots. Nice ones are shot 34, 76, 87, 120, 128 and 129. From shot 140 a few problems with density control occured, and the quality of the PPCD shots decreased. From that shot up to the end several changes have been tried, but with no results.
From shot 20 the gains are 10^8 for diodes 1 and 20, 10^7 for diodes 3 and 18, and 10^6 for diodes 5,...,16, both SXR1 and SXR2. From shot 28 diode 1, SXR1 has 10^7.
Shot 141 to 146 have the prototype ELAD1 hooked up to diode1, SXR1 (gain=10^7=10^5x1x10x10), and the evaluation board of the logarithmic amplifier AD8304 of Analog Devices hooked up to diode 20, SXR1. Offset in diode 1 signal (it was not zeroed), but AD8304 seems not to work.
Shot 147 has the ELAD1 to diode 20, SXR1 (gain=10^8=10^5x10x10x10, offset zeroed) and AD8304 to diode 1, SXR1. Logarithmic amplifier still does not work; the ELAD signals seems noisy. Ok, more tests next time.
At the end of the session the SXR1 probe has been extracted and the VAT closed but not to the end, to allow pumping. This will protect the probe, and the Be foil, also during PDC overnight.

21-Apr-2003

Test on electronics of diodes 9 and 10 of SXR2. Often, when the signals are greater than about 5.5 V they are clipped. Diode 11, SXR2, which has a very similar signal to that of diode 10, does not show this behaviour. In particular, the signals, when greater than 5.5 V, are different from all the other diodes of SXR1 and SXR2 (and also the Be1 and Be2 SXR monitor signals). See shot 1030417007 for example, diode 9 and 10, SXR2, between 15.5 and 19.5 ms. This is why the brightness profile of SXR2 has, often, a "hole" in the centre, due to the smaller values of diode 9 and 10; that is not the Be foils, because the hole appears only when the signals are more than 5.5 V.
The amplifiers of diode 9 and10, SXR2 have been tested, and also the isoamplifiers: no problems found, with input signals from 0 to 10 V (divided by 10^7 with a precise resistor). Also, another Mikhail's amplifier have been tested on the bench, and the response was linear from 0 to 11 V, then it saturates. The encoder was checked on the bench (not the installed one, but another exactly similar), simulating also the triggering: again, no problems from 0 to 10-11 V. We do not think this is a problem with the Be foil, since the signals should be smaller all the time, while the signals are clipped when the output voltage is greater than 5,5 V. And this should not be the diodes, since on saturday (19 apr) the plasma currents was 400 kA, PPCD, and the currents from diodes were up to 8 times larger (a gain of 10^6 was used). This feature was not observed, and in fact the signals are smaller then 5.5 V. To check this problem the inputs of the amplifiers of diodes 10 and 11, SXR2 have been switched. This change is valid from shot 1030421001.
Other shots have been analysed, and it turned out that this problem on diode 9 and 10 started from 10 april 2003, when the encoder was installed in the SXR2 amplifiers. Shot 1030318054 is a 400 kA one, diode 10 and 11 have voltage greater than 6 V (gains are 10^7), but no clipping is observed. Instead, shot 1030410025 is a similar shot, the gain are the same, but clipping is present on diode 10, and not on diode 11. The encoder was installed on 3 april, but enabled on 10 april...Is that the problem?

22-Apr-2003

More tests trying to solve the clipping problem of the diodes of SXR2.
1030422044: all SXR2 amplifiers are off, except those of diodes 18 and 20.
045-050: power supply (of SXR2 amplifiers and encoder, built by Mikhail) replaced with a new one. All SXR2 amplifiers are on. With this power supply the signals are sligthly more noisy.
051: all SXR2 amplifiers are off
052: all SXR2 amplifiers are on
053-056: plasma shots, no clipping problems
057: diodes 9 and 10, SXR2 have the clipping problem, starting at 5.5 V. So this is not due to the power supply
058: now the clipping on diode 10, SXR2 occurs at 7.2 V
059: clipping at 9 V
060-062: clipping and saturation of the Joerger (you can recognise by the jump of the amplitude signals from 10 V to -10V)
063: inputs of amplifiers of diodes 10 and 11, SXR2 restored (they have been switched on 21-apr). To further investigate this problem the outputs of the iso-amps of diodes 10 and 11, SXR2, have been switched (this is the same as switching the inputs of the Joerger of the same diodes). No clipping is observed. The switched cables put the two Joerger of the SXR2 probe in "communication" (diode 1 to 10 are in one Joerger and diodes 11 to 20 are in another digitizer).
064-080: no clipping observed
The outputs of the iso-amps of diodes 10 and 11, SXR2, are left switched.

23-Apr-2003

Checked the cables from the iso-amps to Camac rack #1 BNC panel, and those from the Camac panel to the digitizers. All they seems ok, no one is shorted, and the polarity of these cables are all the same
The original power supply of SXR amps and encoder reinstalled; the clipping problem is not due to the power supply
Modified the voltage of the impulse of the encoder, now the correspondence gain - voltage is:

gain	voltage (mV)
10^6	700
10^7	300
10^8	500

A few check shots. In 1030423003 all diodes of SXR1 and SXR2 have 10^8; SXR1 extracted, VAT closed. From shot 1030423004 the gains of 400 kA shots have been set: diodes 1,3,18 and 20 have 10^8, diodes 5,...,16 have 10^7, both SXR1 and SXR2. All encoders seemed to work, apart from that of diode 3, SXR2, which showed a different gain (10^6 instead of 10^8)
Dumpers have been put under the rack with the SXR amps and the iso-amps: these should reduce the noise at the beginning of the discharge, which may be microfonic noise

24-Apr-2003

200-400 kA , PPCD+pellet shots. Started with 200 kA; diodes 1,3,18 and 20 have 10^8, diodes 5,...,16 have 10^7, both SXR1 and SXR2. 400 kA from shot 1030424053, diodes 1,20, SXR2 are 10^8, diodes 3,18, SXR2 are 10^7, diodes 5,...,16, SXR2 are 10^6; for SXR1, diodes 20 is 10^8, diodes 1,3,18 are 10^7 and diodes 5,...,16 are 10^6
Shot 055 is a nice PPCD; in shot 060 SXR (and also magnetic) data have not been stored, and problems with the pellet injector started
From shot 061 the output cables of the iso-amps of diodes 9,10, SXR2 have been switched back to original configuration (they were switched from 1030422063). No other plasma shots, the run session was closed
At the end of the session SXR1 diagnostic have been extracted, and the VAT valve closed, but not completely: the VAT valve is left cracked to shutter the fragile filters and to protect the pinhole and the probe

25-Apr-2003

400 kA PPCD+pellet shots. SXR1 inserted and VAT opened. diodes 1,3,18 and 20 have 10^8, diodes 5,...,16 have 10^7, both SXR1 and SXR2. High density at the beginning, started to be ok from shot 1030425025. A few problems with the interferometer and C1 capacitor bank, solved after shot 030
PPCD started from shot 033. In that shot, and also 034 to 037, the gain of the diodes have not been changed, in order to check the clipping problem in diodes 9,10, SXR2. These signals are greater than 5.5 V, up to 10 V (saturation) but NO clipping is observed. This problem apparently disappeared...
Gains have been changed from shot 038: diodes 1,20, SXR2 are 10^8, diodes 3,18, SXR2 are 10^7, diodes 5,...,16, SXR2 are 10^6; for SXR1, diodes 20 is 10^8, diodes 1,3,18 are 10^7 and diodes 5,...,16 are 10^6
Encoder of diode 3, SXR2 is still not working (a different voltage is showed). This has been checked between shots 046 and 047: the cable from that encoder to the amp, a bipolar one, showed wrong voltage in the three pins for the three gains
Pellet did not work, and the rest of the day has been 400 kA PPCD without any pellet. Nice example: shot 059
At the end of the session the SXR1 diagnostic has been removed from MST, and the pinhole has been replaced with a "blind" one (no hole) in order to check the noise component in the signals.Two stainless steel foils (0.1 mm thickness each) have been installed; the Be foil has been replaced with the coated one (to avoid any damage to the new foil, installed on 18 april); after 50 minutes the diagnostic has been installed in MST, extracted and with the VAT closed
A few check shots, from 114 to 119. Cables and gains checked. After 45 minutes of rough pumping the VAT has been opened and the SXR1 probe inserted. A shot has been taken, 120, to check if the probe is working; no problems found

26 and 27-Apr-2003

400 kA shots, OFCD. Even if the SXR1 diagnostic is blind (no pinhole), noise is still there, and its level seems not to change. This noise has not an electrostatic origin, instead we think it is definitely magnetic pick-up

28-Apr-2003

SXR1 probe still without pinhole. 450 kA LH shots
Clipping problem in diodes 9,10, SXR2 not anymore present. We still do not know what was going on
Looking at the noise component in the SXR1 signals, it seems that noise increases from diode 14,16 towards diode 1 (try to plot all SXR1 signals with the same Y scale). Diode 1,3,18 and 20 have 10^8, the others 10^7, and 10^8 has always twice the noise (see tests on 16 april). But still an increasing noise is observed moving from diode 16 to 3; and also moving to diode 1, since diodes 18,20 have smaller noise than diodes 1,3. The signals are grounded to MST using the BNC "T" on the flange; all shields of the cables are hooked up together in the socket of the array, and a wire goes to pin 12 of the socket, which is the common anode. Diode 14,16,18,20 are close to pin 12, diode 1,3,5 are on the opposite side. Maybe this increasing noise is due to the asymmetry in cable grounding configuration. We are using only one pin of common anode, instead of two (the other one is pin 1) because in february 2002 we realised that the noise on two channels grounded one in pin 1 and the other in pin 12 have opposite phase; this may be due to the fact that noise has magnetic origin (pick-up), and grounding in two different points produces noise with opposite sign since the magnetic flux concantenates with the circuits in opposite directions
In shots 1030428063-073 diode 20, SXR1 has one Mikhail's amp, gain=10^8, with 5 kohm input resistor (instead of the usual 50 ohm). The idea (of G.Fiksel, but not so very clear...) is that magnetic pick-up noise should decrease with higher input impedance. Instead, higher noise is observed
From shot 1030428074 the original Femto has been installed in diode 20, SXR1; also, all diodes of SXR1 have 10^7 (signals with 10^8 have bigger noise due to amp, and we are interested in measuring the pick-up noise). Diodes 18,20, SXR1 have now smaller noise, but still higher than diodes 14,16, SXR1, which seems the best ones
At the end of the run session the SXR1 diagnostic has been removed, and VAT valve closed. Tomorrow, the original pinhole will be reinstalled, and a different ground configuration will be used; also, the pinhole will be moved in order to get a pinhole-array distance of 20 mm (it was about 10 mm all the time); a ring will be placed between the probe and the top. This is to test a different spatial resolution of the probe (with increased distance the fan of lines of sight covers a smaller poloidal section, more concentrated in the plasma core)

29-Apr-2003

To try to reduce the noise in the SXR1 data, these signals have been grounded to MST directly in the array socket, inside the probe. To do this two wires from the cables shields (which are all hooked up together) have been passed through a hole in the socket and wrapped around the screws which hold the socket in the probe. This way the cables shields will be not left disconnected inside the diagnostic (they were grounded outside, in the flange, using the BNC "T"). A few _photos_ have been taken during this operation
10 mm ring added to the probe, to increase the distance between pinhole ad array from 10 to 20 mm. The ring dimensions, the positions of the socket and array have been carefully measured; see this _page_ as a reference for these kind of measurements
the blind pinhole has been replaced with the 0.8 mm wide pinhole and the 18 microns new Be foil (used from 18 april and removed for the tests with blind pinhole) has been reinstalled
SXR1 diagnostic installed in MST, extracted, VAT closed, and rough pumped. No problems found: shots from 1030429132 and on are dry ones (no plasma); gains: diodes 1,3,18 and 20 have 10^8, diodes 5,...,16 have 10^7, both SXR1 and SXR2. See 1030429138 for an example of the level of noise on signals. As usual, between shot 132 and 137 one or more diodes had bad signals, due to the bad Fischer feedethrough/plugs in the flange of the SXR1 diagnostic.
The BNC "T" is no longer use to ground the diodes signals to MST (this is done in the socket, inside the probe). Instead, this "T" is used to ground the inner copper shield in the diagnostic, which was, up to now, left floating. This copper shield is connected to the feedthrough named "A" in the flange of SXR1
Now that the probe has the 10 mm ring, it is no longer possible to close the VAT valve when the probe is fully extracted. To close the valve the probe must be manually extracted 15-20 mm; the _ring_ which blocks the tube to the moving part must be loosen and left so. To keep the probe extracted, use the small piece of white plastic which should be in the metal box of the Femto amps; or use a BNC "T" (this one is the exact distance you need to keep the probe inside the diagnostic,out of the VAT valve). See _photo_

30-Apr-2003

Encoder SXR2 amps checked, and diode 3,SXR2 repaired. Now they all work fine
One 300 kA shot to test light tightness. The probe was inserted to 127.0 mm (10.0 mm less than as usual) and the VAT opened before shot 10304030015. In this shot no problems are observed
After this shot the diagnostic has been extracted and the VAT valve closed (cracked, only light tightness)
During insertion one of the nuts of the inserting system dropped (one which was fixed in the long screw used to position the probe). Eventually it was find and reinstalled. The insertion system must be checked from time to time
Started to design small amplifiers to be inserted in the probe

1-May-2003

SXR1 probe inserted, 127 mm, VAT opened. 450 kA plasma, LH.
SXR1 signals improved: see for example shot 1030501005, diode 14,16,SXR1 have noise during the discharge comparable to that before and after the shot! Also, edge signals are bigger due to the modified geometry (now the fan is looking more in the plasma core, where SXR emission is larger). See for example shot 1030501031
Diodes 9,10,SXR2 have not the clipping problem observed a few days ago. The reason for that clipping is still unknown; now all channels seems to work properly (in the same shot 1030501031 diodes 9,10,SXR2 are more than 5.5 V and no problems are found)
4 _miniaturised preamplifiers_ was built and are ready to be installed inside the SXR1 probe. They have fixed gain of 10^6 and a bandwidth limited to 100 kHz (can be more than that, up to 1 MHz). The op-amp is the OPA655

2-May-2003

Worked on miniaturised board with 4 preamplifiers. The diodes 1, 10, 11 and 20 will be tested with the preamplifiers, the remaining 8 will be used as usual (with Femto amplifiers outside the probe). On the flange, feedthrough number 2 and 4 will be used as -5V and 5V, resp., for the preamps.
Ground of new SXR1: all cables, 8 from the diodes and 4 from the outputs of preamps, have all shields hooked up together; also the "0" of the power supply is hooked upto these shields. A wire connects the shields to pin 1 of the socket, that is the common anode (pin 12 of the socket was damaged and was not used). Also, two wires connect pin 1 and pin 12 (common anode) to the screws used to fix the socket in the probe: one wire from pin 1 to one screw (that next to diode 1) and another from pin 12 to the other screw (next to diode 20). These wires put the shields of all cables to MST (which seems the best configuration).
Cables and preamps checked on the bench: no problems found. Checked also the installation in the old probe (that used from november 2001 to july 2002). To insert cables, board and socket the probe has been internally modified.
Amps gains are 10^8 for diodes 1,3,18 and 20 and 10^7 for diodes 5,...,16, both SXR1 and SXR2 in all shots up to 1030502066, and from 1030502080; for shots 1030502067-079 diodes 1,20, SXR2 are 10^8, diodes 3,18, SXR2 are 10^7, diodes 5,...,16, SXR2 are 10^6; for SXR1, diodes 20 is 10^8, diodes 1,3,18 are 10^7 and diodes 5,...,16 are 10^6 (should be 400 kA PPCD shots, with F=0, but SXR signals are small)
End of day: SXR1 probe removed, VAT valve closed.

3-May-2003

New cables and preamps installed in SXR1. Measured the distance between the array of diodes and the edge of the open probe: now it is 10.20 mm next to diode 1 and 10.25 next to diode 20. The distance is smaller than before (10.45 mm) because all cables and the board behind the array are pushing on the socket, and one hole in the socket (next to pin 1) was larger than the other (it was modified to fit in this probe: in fact this is the socket used in the old version of probe, which had different distance between the holes); due to this, the socket is not parallel to the edge of the probe in toroidal direction (in poloidal direction it is ok since the measured distances are the same within 0.05 mm); see _photo_.
Installed the probe, with the 10 mm ring. The distance between the IRD array and the edge of the probe with the ring is 20.10 mm next to diode 1 and 20.15 mm next to diode 20 (which are the previous measured distances plus 9.90 mm, which is the lenght of the ring)
SXR1 diagnostic installed on MST, extracted and with VAT valve closed. Rough pumped, no problems found

4-May-2003

SXR1 probe still extracted, with VAT valve closed. The signals seems improved (regarding noise): compare the noise of SXR1 and SXR2 signals in shot 1030504091 between 40 and 45 ms. Today the shots are 400 kA, full PFN; this evening 200 kA PPCD, and the SXR1 will be inserted for that campaign
The preamps are working well; one way to check them is to look at the absorbed curents in the power supply: voltage should be 4.9-5.1 V, and current 80-90 mA

5-May-2003

Shot 1030505121-123 are "dry" shot (no plasma, only data acquisition) to check noise in SXR1 signals; gains were changed by changing the iso-amps amplifications: D1,20,SXR1 are 10^6x100, D10,11,SXR1 are 10^6x10. Actually D10,SXR1 was set to 10^6x2, and was put to 10^6x10 from shot 169
During all day there are plasma shots and boronization, so when SXR data are not present those are boronization shots
All SXR1 signals have less noise but have many spikes, in particular a -10/10 V spike at the beginning. Also, while all SXR2 signals have a definite spike at t=0, the big spike in SXR1 signals happens about 0.3 ms later
To understand if there are problems with power supply, we performed a few tests. In shot 294 the power supply of diodes 1,10,11,20,SXR1 have been switched off, and the + and - cables have been disconnected from the SXR1 flange. Spikes still there
From shot 316 the power supply has been switched on
Shots 327-330 have D1,10,11,20 off, all 4 cables and + and - cables are disconnected. Spikes come and go, and also the 0.3 ms delay sometimes is not present (for example shot 328). In any case these internal amplifiers did not further reduced the noise, so for the next campaign they will be removed and the original cables will be installed. In these shots both common anodes of the IRD are grounded to MST, using the two screws that hold the teflon socket, while previously only common anode of pin 12 was grounded to MST (pin 1 was left open)
SXR1 diagnostic extracted (but left attached to MST), VAT closed

6-May-2003

Shots 1030506001-114 are 400 kA, F=0; SXR signals small. From 115 we started a dedicated campaign with 400 kA PPCD and then 400 kA EC shots. PPCD started at shot 116, and up to shot 119 the gains were not changed, so a few channels hed saturation. From shot 120 gains are 10^8 for D1,20,SXR2, 10^7 for D3,18.SXR2 and 10^6 for D5...16,SXR2; for SXR1 diodes 1...7,10,11,14...20 are 10^7 and D9,12 are 10^6. Diodes 1,10,11,20,SXR1 are 10^6x10 (x10 due to iso-amp)
During all day (actually, evening) diodes 1,10,11,20,SXR1 showed a strange behviour: after 13-16 ms the signals went to 0. We think this was due to power supply. Look for example shot 1030506121, SXR1 signals after 15-16 ms
From shot 124 D1,10,11,20,SXR1 are 1x10^6. Signals went to zero the same
From shot 126 D1,20,SXR1 have been disconnected from the flange
From shot 132 D1,20,SXR1 have been hooked up to the flange, and D1,10,11,20,SXR1 are again 10^7=10x10^6. Power supply to 5.2 V instead of 4.9-5 V. No changes
In shot 134 the power supply has been switched off; no changes. From shot 135 this power supply was on again
From shot 153 D9,12,SXR1 are 10^7. The - wire has been disconnected from the power supply: no changes (the internal amplifiers have been actually switched off)
From shot 155 the - wire has been hooked up, and + wire has been disconnected; D1,10,11,20,SXR1 have no data (internal amps are off). From shot 160 the + wire has been hooked up
From shot 170 shots are EC, 400 kA, F=-0.4
From shot 173 gains are 10^8 for diodes 1,3,18 and 20 and 10^7 for diodes 5,...,16, both SXR1 and SXR2 (D1,20,SXR1 are 100x10^6, D10,11,SXR1=10x10^6)
From shot 176 D3,18,SXR1 are 10^7 and D1,20,SXR1 are 2x10^7=20x10^6 (iso-amps had x20). SXR2 signals have many "spikes" (could be not only noise) which saturate, but the "good" signals are less than 10 V
From shot 184 D3,18,SXR2 are 10^7 and D11,12,SXR2 are 10^6. This way it was possible to get also the "spikes" (possibly) due to heat-pulse propagation
At the end of the session the probe has been extracted and removed from MST. VAT closed @ 22:40
Note for the clipping problem for diodes 9,10,SXR2. We realised that one of the power cord of the MIK amplifiers had not the -15 V. This cord was originally hooked up to the amplifier of diode 3,SXR2, and that's way the voltage pulse proportional to the gain setting was incorrect. After a few modifications and tests on the encoders this cord was hooked up somewhere else (there are no lables on these wires). The same time we realised the clipping problem. Mikhail believes that one cable without -155 V could affect all other channels.

7-May-2003

Note on MST shutdown: the starting date has been delayed a week, not it will start on monday, 26 may 2003
Replaced the cables of the SXR1 probe with the original ones (used until 3 may), teflon cables, without any internal amplifier. Changed the ground layout. Each teflon cable has both the shields hooked up together (they are triaxial) and all these shields are connected; the cables are arranged in two bundles, 6 cables each, and a wire connect the shields of one bundle to the other. Then all the shields are hooked up to PIN12 of the IRD array, that is to common anode. Diodes 1,3,5,7,9,10,11,12,14,16,18,20 are used; the remaining 8 diodes are grounded to PIN12. Two wires are hooked up to PIN1 and PIN12 and go to the screws which hold the teflon socket, and thus to MST; PIN1 is connected to the screw near diode 1, PIN12 is grounded to the screw near diode 20.
No copper shield was installed (no major changes in using it)
Distance between the surface of the diodes and the edge of the open probe has been measured and is between 10.40 and 10.45 mm. The usual distance of 10.45 mm can then be used. The socket now is more stable than before (see notes for 3 may)
The 10.0 mm ring has been installed to improve the spatial resolution. The pinhole is still 0.8 mm wide poloidally
SXR1 diagnostic mounted on MST, extracted, VAT closed. Gain are are 10^8 for diodes 1,3,18 and 20 and 10^7 for diodes 5,...,16, both SXR1 and SXR2. Iso-amps gains are 1
Shot 1030507001-002 are "dry" shot to test the electronics: all SXR1 diodes have signals constant=0.4 V. After this check the VAT has been opened and the probe inserted (127 mm). Electronic layout checked: the iso-amps of SXR1 were off. Shot 003 is another test shot, iso-amps are on and all channels are ok
The internal amplifiers have been tested on the bench and no problems were found. With input currents of 5x10-7 A=500nA the output signals are about 0.5 V, and gain is 1x10^6 to a few percent; yesterday the problems of the signals put to zero presented at about this voltage, but on the bench no problems have been found. We tried also to install the IRD array on the socket and to illuminate it with a flashing diode but again, with output voltage up to 1 V we did not found any problem. The iso-amps have been calibrated, and the gains are 1,2,5,10,20,100,200 and 1000 within a few percent. We still think that there was a problem in power supply of the internal amplifiers
Due to problem to the primary winding of the MST transformer the run shedule has been suspended

8-May-2003

No shot

9-May-2003

No shot

The SXR1 electronic layout is now:

diode	Femto	Cable
1	1	11
3	2	12
5	3	13
7	4	14
9	5	20
10	6	15
11	7	21
12	8	16
14	9	17
16	10	18
18	11	19
20	12	10

so it is easy to change gain for the diodes (before the diodes and amplifiers were not connected sequentially). Next time the cables from SXR1 flange to Femto amplifiers and from Femto to iso-amps will be also changed...

10-May-2003

Back to Itay

15-May-2003

MST experiments restored
From 15 to 19 May 400 kA OFCD and OPCD experiments have been performed

20-May-2003

v x B dynamo data during QSH with IDS @ 300 kA. All channels of SXR1 and SXR2 have gain=10^8

24/25-May-2003

OFCD experiments @ 300 kA. SXR1 and SXR2 have the usual 400 kA gains, that is diodes 1,3,18 and 20 have 10^8, diodes 5,...,16 have 10^7. Low signals

26/27-May-2003

400 kA PPCD + pellet. The pellets have been also fired during the PPCD
From shot 001 to 015 diodes 1,3,18 and 20 have 10^8, diodes 5,...,16 have 10^7, both SXR1 and SXR2
From shot 016 SXR1 gains are all 10^7; in shots 016...018 D1,3,18,20 have been set by mistake to 10^5
From shot 016 SXR2 gains are: D1,20=10^8, D3,5,16,18=10^7 and D7...14=10^6
There are a few 200 kA PPCD shots at the end of the 27 May session (from shot 095)
After this session the gains have not been changed

2-Jun-2003

A few 400 to 530 kA shots. Nice SXR measurements.
MST vent started.

Sep-2003

LM and PP changed the pinhole and Be foil of the SXR2 probe. The SXR1 and SXR2 cover have been cleaned (to remove the carbon which we believe can contribute to the covering of the Be filter).

19-Dec-2003

SXR3 and SXR4 probes arrived at MST.

Jan-2004

End of the MST vent. First plasma shots.

3 to 17-Apr-2004

Visit to MST to install the four SXR probes. Tests of the amplifiers (Femto, Mikhail's modules and a prototype of a logarithmic amp). At the end of the visit the four SXR probes have been installed, and 5 diodes of the SXR2 have been connected; the other probes (manipulators) have been left extracted. The logarithmic amp was not suitable for our applications, and the Mikhail's amplifiers have been chosen as the amplifiers of the tomography. After the visit the construction of the new electronics (66 channels, each with a trnasimpedance amp, gain 10^6,7,8, and an isolated amp, gain 1,2,5) started.

top | back