Was the connection between the feedthrough (atmosphere side) and the connector on the optical table confirmed to be OK?
We had a similar situation for the TT1. We found that we were using the wrong feedthrough connector (see TT1 elog).
The major problem that Manasa and I found was that we weren't getting voltage along the cable between the rack and the chamber (all out-of-vac stuff). We used a function generator to put voltage across 2 pins, then a DMM to try to measure that voltage on the other end of the cable. No go. Jamie and I will look at it again today.
Everything was fine. Apparently these guys just forgot that the cable from the rack to the chamber flips it's pins. There was also a small problem with the patch cable from the coil driver that had flipped pins. This was fixed. The coil driver signals are now getting to the TTs.
Investigating why the pitch/yaw seems to be flipped...
The PMC PZT voltage slider seemed sticky. First it would not do anything, than after moving slider back an forth a few times, it had a range of 60V and later it had full range and it locked
We started off to try and get TT2 working. We used the cables Jamie had already prepared while working on TT1 and used them to connect TT to the channels in 1Y3.
There were sma cable connectors already running between the channels 5-8 on the board to the UL,LL,UR and LR. Triggering the UL LL UR LR matrix on epics did not show any analog voltage at the output analog channels on the board. Talking to Jamie over phone, we inferred that the SMA cables that were already left connected corresponded to channels assigned for TT4 in epics. So we set the connections right and could see analog voltage outputs corresponding to epics triggers.
We connected the ribbon cables running from the board to the TT. But changing pitch and yaw did not do anything to the TT2 mirror. We opened the BS door and checked if the tt cables were connected to the post. We beeped the cable running from the board to TT (we also traced the cable's trail through the cable rack pile from 1Y3 to BSC). Using a function generator at the board end of the cable, we could not observe anything at the TT end of the cable.
We ran out of options on what can be done next and closed the doors. We hope Jamie can fix the problem once he returns next week.
I came in this morning to see that the PMC was down. The PZT voltage had drifted to below 50V. I adjusted the FSS slow controls to 0V and PZT was back at 126V.
PMC and IMC could eventually be locked.
History of PZT voltage behaviour in dataviewer over the last 24 hours shows it has been drifting everytime after it has been fixed.
FSS was saturating. Fixed.
[Bob, Manasa, Jenne]
We opened the ITMX heavy door. Before getting too far, we realized that we had to do the fancy pin swapping before we can activate TT2. So....
We followed the instructions in elog 7869, and the associated Picasa album, and swapped the pins for the in-vac connector that will go to TT2. Pretty easy, since the procedure was already well documented.
We then looked at the beam location on PR2, and the beam is ~2 inches up and to the left (as viewed from the front) from the center of the optic. This is very easily correctable with the actuators, so we're leaving TT2 as it is.
I have created a wiki page linked here with all details about the endtable upgrade.
The page has links to the new drawing for doubler post to hold the tilt-aligner that holds it. The Faradays will also be mounted similarly on tilt aligners placed on these posts. The bulk mounts will be made of aluminium similar to the colorful cylindrical mounts (images of which can be seen in the archived layouts on the wiki) that hold the He-Ne lasers and few faradays now.
* TT1 is in place, aligned so beam hits center of TT1, hits center of MMT1 (used pitch biases to finish pitch).
I had asked Q to write this down on a piece of paper, but then I forgot to transcribe it into the elog....
The TT screen matrix, at least for TT1, is flipped or something. When Eric moved the pit slider, the optic moved in yaw, and vice versa.
We need to fix this, but for now, here's the situation when TT1 was pointed correctly at MMT1:
TT1 Pit slider | 1000 1000 | ---> 700 UL
0 | -1000 1000 | ---> 700 LL
TT1 Yaw slider | 1000 -1000 | ---> -700 UR
0.7 | -1000 -1000 | ---> -700 LR
The confusing thing is that Koji and I confirmed (by plugging in the correct cable to the correct sensor) that "UL" on the screen goes to the UL coil, and the same for the other 3 coils. This needs investigation / fixing.
Today I have tested the MC transmission-end RF photodiode PDA255 for transimpedance and dark noise using Jenne's Laser and AG4395A network/spectrum analyzer. The dark noise voltage distribution for the transmission and reflection PDs of MC and the analyzer has been compared.
I am to do the input mode cleaner cavity mode scan. The electronic and shot noise of the components used , particularly photodiode noise, will affect the peak position of the modes, indicating the uncertainty in the measured frequencies of the modes. That will in turn give the uncertainty in the measured change of radius of curvature of the mirrors in presence of the laser beam, from which we will be able to calculate the uncertainty in the mirror-absorption value.
For PD transimpedance measurement I used Jenne's laser along with AG4395 network analyzer. The RF out signal of AG4395A had been divided by splitter with one output of the splitter going to R channel of the network analyzer and the other to the laser. The splitted laser beams - splitted with beam splitter - fall on two photodiodes - one reference(Newfocus1617? PD, the DC and RF transimpedance values were taken from its datasheet ) and the other on PDA255. The outputs of these two photodiodes go to channel B and A respectively of the network analyzer. The measured transimpedance data had been collected using the GPIB connection. It had been ensured that the PD under test is not going to saturation, for that the source power level was kept to -40dBm. transimpedance measurements were compensated by the ratio of DC photocurrent.
For dark noise measurement the output of the PD was connected to the A channel of the AG4395A, when there was no light falling on it. The response is collected using GPIB. The attenuation of channel A was made 0dB. ( AG4395A was kept in Spectrum analyzer mode in Noise Format).
The plots corresponding to the measurements are attached.
The comparison for the dark noise voltage levels of the MC transmission PD (PDA255) with MC REFL PD has been made with analyzer dark noise voltage. It is shown in the attachment (I will upload the dark noise current comparison too....since the output darknoise depends on the gain of the circuit, it is important to divide this voltage spectra by transimpedances.)
PZT2 was removed from the BS table, and packed away in a foil-lined plastic box.
PRM oplev path has been altered, including installation of a 3rd mirror, to accommodate TT2, which is larger than PZT2.
* Unfortunately, PR3 is a few mm more north than is indicated in the CAD drawing, so I wasn't able to place the oplev mirrors exactly as Manasa indicated in elog 7815.
* We came up with a different layout. Photos were taken. We will need to confirm that the IPPOS, AS, and GreenX beams all clear past the oplev mirrors, but by imagining straight lines between mirrors for those beams, I think we should be okay. but we must confirm when we have real beams.
TT2 was installed, according to the placement in the diagram. Dogged down just as TT1 - one dog for the riser, 3 dogs for the TT base which also squish the riser. You should be able to see this in the photos. Without having installed the PRM target, it looks like the input beam is hitting pretty close to the PRM's center. Tomorrow Jamie The Tall can install the PRM target for us so we can confirm.
Photos - I'm posting them on Picasa here. The new camera, and the fact that you can rotate the viewfinder, is amazing for overhead in-chamber photos. Seriously, it's much easier to take useful photos. It's great.
We remove the ITMX door first thing. If Steve isn't here, we'll ask Koji or Bob to help us with the crane.
First thing on the alignment list is to finalize TT2's pointing. Put a target in front of PR2, put on the PRM target, etc, etc. We're basically back to the same alignment procedure as we've been doing the last few vents.
Item for meditation:
Do we trust ourselves, or do we want to think about installing a 'bathroom mirror' so we can see the face of PR3 while we are pumped down?
- A new projector lamp installed.
- The old lamp lasted 8751 "equivalent lamp hours".
- The old lamp was found being shattered inside. It contains mercury.
So next time you hear the explosion sound of the lamp, establish the ventilation of the room and escape for an hour.
I don't know why (I'm just leaving the lab right now....) but BS, PRM, SRM all have no light on their oplev PDs. I have turned off the oplev servos for now, and will get back to them tomorrow, before redoing the BS table oplev layout.
[Jenne, EricQ, Nic, MattA]
* Riser installed, dogged down with 1 dog.
* TT1 sitting on top of riser, 3 dogs holding TT to table, with riser squished in between.
* IOO table leveled.
* Almost all of the weights on the IOO table were just sitting there, not screwed down! One didn't even have a screw, 3 had screws, but they were totally loose. 2 of those screws were in as far as they could go, but they weren't holding the weight. This means the screw was too long, and should have been replaced (which I did today). Just because the existing screw was too long, doesn't mean it should be left as-is. Everything in the chambers must be tightly clamped down, as soon as work on that item is complete! Anyhow, after finalizing the leveling, I tightened down all of the weights on the IOO table.
* MMT1 tweaked so beam hits center of MMT2.
* MMT2 tweaked so beam hits center of PZT2.
* Light access connector installed.
* I dropped a Class B golden-colored 3/16 allen key to the bottom of the IOO chamber. I can't see it, but Nic thinks he might be able to see it with a mirror, from the BS chamber. We should look for it when we look for the IR card that is still down there.
* We have an ant in the IOO chamber. Unfortunately my hands were on the TT1 optic holder ring when I saw it, so I couldn't dash quickly enough to grab it. I saw it run over the side of the table, and down, but looked under the table and couldn't find him. Not so good, but I don't know what to do about it right now. If anyone sees it, get it out please.
You have to correct this transimpedance ratio by correcting for the different levels of DC photocurrent in the two devices.
For the dark noise, you must always include a trace showing the noise of the measurements device (i.e. the analyzer noise must be less than the dark PD noise) with the same input attenuation setting.
The correction for different levels of DC photocurrent in the two devices had been taken care by one MATLAB code, the code that originally was made by Koji.
The analyzer noise I had not recorded; today I am going to record it.
Here is the data for AG4395A network/spectrum analyzer noise data. I collected the data by putting 50ohm terminator on channel A with same input attenuation setting (0dB attenuation).
Here I upload the plots corresponding to my last day's measurements.
Also, we still need to get a 32GB SD card for the new camera. It only has an 8GB one.
[Jenne, with backup from Koji and Steve]
TT1 was installed without a riser, optic is too low, riser we have doesn't fit, cannot proceed with alignment. Sadface.
I had gotten to the point of checking that the beam coming out of the Faraday was hitting the center of TT1, when I realized that we had forgotten to install the risers. The TTs are designed for 4" beam height, but we have a beam height of 5.5" in-vac. This means that the beam out of the Faraday was hitting the top of the optic / the optic holder.
Steve showed me where all of the active TT equipment is stored (down the X arm, almost all the way to the flow bench...there is a plastic tub full of baked items (individually wrapped and bagged)), and I got one of the 1.5" risers.
Upon opening the riser package, and comparing it with the base plate of the active tip tilt, the screw holes don't match!
It looks like for the passive tip tilts, we had holes machined at the far corners of the base plate, then had these risers made. You can see in the photo of SR3 below that the original holes are there, but we are using 1/4-20 holes at the far corners of the base plate.
Unfortunately, without checking the base plate, I had asked Steve to get 4 more of the same risers we used for the passive tip tilts. So, now the base plate holes and the riser holes don't match up. In a perfect world, we would have installed the risers on the TTs as soon as they were baked and ready, and would have discovered this a while ago....but we don't live in that world.
The reason we had originally chosen to put the new 1/4-20 holes on the corners of the passive tip tilts was so that when we tightened the screws, we wouldn't bend the base plate, due to the groove at the bottom of the base plate being directly under the screws. Since the new aLIGO TT base plates have the groove underneath going the opposite direction, we didn't need to move the holes to the corners.
Also, you can't really see this from the photos, but the active TT base plate is slightly longer (in the beamline direction) than the riser, but only by a little bit. Koji is currently trying to measure by how much from the CAD drawings.
Also, also, because of the way TT1 will hang off the table, I'm concerned about the underneath groove on the riser being the direction it is. I'm concerned that the grooved part will be what wants to touch down on the back corner of the table, such that either the TT is insufficiently supported, or it is tilting backwards. Neither of these will be acceptable.
I propose that we re-make the risers quickly. We will have the holes match the active TT base plate, the size of the riser should match the size of the active TT base plate, and the underneath groove should be perpendicular to the way it is in the current version.
List of things to do, in order:
* Remove BS heavy door. Steve, please remove the BS door as soon as you have enough people to do so. I will be a little late, since I have a dentist appointment, but please don't wait for me. Jamie and Manasa can help you. Put on a light door.
* Remove MC light doors, make aluminum foil tube (not light access connector, yet).
* Open laser shutter, lock PMC. (Required slight tweaking of input steering.) Confirm power level into vacuum <100mW.
* Lock MC and check spot positions of MC (quickly. this shouldn't take all day, hopefully).
------------------------------- End of work for Monday. See following elog ------------------------------------------------
* Move TT1 to be as close as possible to the location indicated on the diagram, then align it.
* Make sure beam out of Faraday is hitting the center of the optic.
* Make sure beam reflected off of TT1 hits center of PZT2. Only use actuators for the final alignment, then confirm that they aren't close to the edge of their ranges.
* Lock down TT1 with dog clamps.
* Put light access connector on MC.
* Swap PZT2 out with TT2. Should be at correct spot, according to diagram, and beam should be hitting center of optic. Alignment only to the ~few degree point here.
* Re-level BS table.
* Fix oplevs that need fixing. (Manasa should have the plan on one of the diagrams).
* Put target on PRM cage.
* Align TT2 so that beam goes through PRM target.
* Open ITMX heavy door. (Probably Tuesday morning).
* Place freestanding target in front of PR2. Ensure TT2 is aligned to go through PRM target, and hit center of PR2. Again, save actuators for fine-tuning.
At this point, I think we should (temporarily) install one of the G&H mirrors as a flat mirror facing the PRM, and see if we can lock that cavity using REFL. We will want to have already created a model for this case, to compare our observations to. Or we could align the full PRMI, and try to lock that in air.
D - UL
B - UR
A - LR
C - LL
Today I have measured the transimpedance and dark-noise of the MC-REFL PD.
For transimpedance measurement I first collected the data of the reference Newfocus PD connecting it at channel B of Network-analyzer using the set-up of Jenne's laser. The data for the MC-REFL PD had been collected by connecting it to the A channel of Network Analyzer. To do that I shifted the Jenne's Laser to the table of MC-REFL PD, I moved the laser output on the table and fixed a lens and a mirror on the table. Taking the ratio of the two sets of datas I got the required trans-impedance.
Dark-noise readings were taken keeping the laser off.
I will upload the corresponding plots tomorrow.
Today I found 3 power cables in the orange Pomona cable tray, put in so that the cables were damaged and therefore dangerous.
Please think about what you are doing before doing it. Damaging these things because your are in a hurry or frustrated will just waste our time and damage our interferometer.
For reference, we only use the thick blue Pomona racks for power cables. We use the orange and black ones for thinner cables. Pay attention and keep the cables organized.
Cable Rack Selection
Immediate things to do include finishing installation of new TTs and re-routing of oplev paths in the BS chamber, but after all that, we should retry in-air locking.
The last time we (I) tried in-air locking, MICH wouldn't lock since there was only ~ 6uW of light on AS55 (see elog 7355). That was before we increased the power into the MC by a factor of 10 (see elog 7410), so we should have tens of microwatts on the PD now. At that time, we could barely see some PDH signal hidden in the noise of the PD, so with a factor of 10 optical gain, we should be able to lock MICH.
REFL should also have plenty of power - about 1.5 times the power incident on the PRM, so we should be able to lock PRCL.
Even if we put a flat G&H mirror after the PRM to make a mini-cavity, and we lose power due to poor mode matching, we'll still have plenty of power at the REFL port to lock the mini-cavity.
For reference, I calculate that at full power, POX and POY see ~13uW when the arms are locked.
POX/POY power = [ (P_inc on ITM) + (P_circ in arm)*(T_itm) ] * (pickoff fraction of ITM ~ 100ppm)
REFL power = (P_inc on PRM) + (P_circ in PRCL)*(T_prm) =~ 1.5*(P_inc on PRM)
[Koji, Rana, Nic, Steve]
I recalled that we used an optical lever to check if the TT is moving or not.
We used a laser pointer on a tripod, which was prepared by Steve.
I should also note that we stepped back the eddy current dampers from the magnets
in order to enhance the motion of the suspension. They should be restored in the end.
The mini-D connectors on the OSEMs are loosened so that we can plug the cables in.
This requires a specific metric allen key that is stored in a clean tool box with an aluminum foil.
I have repeated the transimpedance measurement of PDA10CF. Also made the dark current noise measurement by connecting the PDA10CF output to the A channel of network analyzer. The results are as follows. I I started to take the reading for shot noise intercept current using a light bulb in front of the PD, changing the current through the bulb, but at higher current the bulb filament got broken, so the experiment is incomplete.
[Koji, Rana, Nic, Steve]
We went to the 25-pin D cable which connects to the TT1 quadropus and succeeded eventually in swapping pins 12/24 into the 13/25 positions.
Accu-Glass is closed till Jan 2, 2013
Atm2 would be the ideal solution Gender Adaptor Special with male - female sides and vented D-sub. How many are we getting 2 or 6 ?
We need short cables that mirror the pins:
The male side will plug into the 25-pin female on the stack-top bracket. The tip-tilt quadrapus cable will plug into the female side. This will match up pin 1 on the tip-tilt cable, which is connected to it's shield, to pin 13 on the bracket, which is the shield of the cable that runs to the stack.
They need to be vacuum compatible. Shorter length is preferred, and there is no minimum length (something like an all-in-one gender changer would be ideal, but probably expensive to have made).
So this is obviously a general problem for all the TTs. Our in-vacuum wiring is unfortunately mirrored relative to that of aLIGO, or at least:
And again, the problem is that pin 13 on the TT quadrapus cables is the one of the coil pins for one of the OSEMs.
I think the right solution is to make mirroring adapter cables for the TTs. Modifying the pins on the stack-top brackets for just the TTs would leave us with a bunch of brackets that are different than all the rest, which I think is a bad idea. Therefore we leave all the feed-thru-->bracket wiring the same, and make adapters. I'll describe the adapters in a follow-up post.
The silver lining to this whole thing, if there is one, is that I wired the polarity on the out-of-vac adapter cable at the coil driver in such a way that the drive/send signal went to the grounded in-vac pin. If I had by chance wired the polarity oppositely everything would probably have worked, except that the return for one of the coils would have gone through the cable shield and the chamber, rather than the return pin to the coil driver. I'll let you image the problems that would have caused.
Making new adapters will take a little while, but I think we can proceed with the installation and alignment with a temporary setup in the mean time by taking advantage of the polarity I mention above. We can temporarily swap the polarity so that we can drive current to the coil using pin 13. This will allow us to complete the installation and do all the alignment. Once the in-vac adapter cable arrives, we just put it in, fix the out-of-vac polarity, check that everything works as expected, and button up.
We'll pick up all this when we're back on Jan 7. Steve will put in an order for the in-vac adapter cables ASAP.
I take full responsibility for this fuck up. We've been unable to find any in-vac wiring diagrams, but I should have checked all of the wiring during the last vent so that we could have prepared for this ahead of time. Sorry.
Nic and I discovered a problem with the in-vac wiring from the feed-thru to the top of the table. Pin 13 at the top of the stack, which is one of the coil pins on the tip-tilt quadrapus cables, is *the* shield braid on the cable that goes to the feed-thru. This effectively shorts one of our coil signals.
There are three solutions as we see it:
* swap pin 13 for something else at the top of the stack, and then swap it back somewhere else outside of the vacuum.
* swap *all* the pins at the top of the table to be the mirror. We would then need to mirror our cables on the outside, but that's less of an issue.
* make a mirror adapter that sits at the top. This would obviously need to be cleaned/baked.
None of these solutions is particularly good or fast.
We've been trying to figure out the connector for the TTs. Since, we found the cables were plugged in wrong in TT1; when triggered in pitch, the mirror moves in yaw and viceversa.
Referring to the cabling diagram, D1000234-v10, we infer that connectors go as J2 - LR, J3 - UR, J4 - LL, J5 - UL and the connections are made looking at the mirror from behind.
The view looking at the optic from the back:
First Contact Training with Margot
Tip Tilt, serial number ### (Manasa will get the serial number and put it in the elog) was taken out of the cleanroom, for use as TT1.
Depending on how far things get tonight, Jamie and Manasa may ask Steve to help them remove the BS door, so they can get started on replacing PZT2 with TT2.
I have fixed TT1 close to what it's position looks like in the CAD drawing. Only 2/3 of TT1 rests on the table...so we need to be extra careful when we will move it for alignment.
Serial Number: SN 027
dcc number: D1001450-V2
We are still in the process of removing the quadrupus from the TT, figuring out which connector goes where, putting it on correctly, and re-testing.
We closed the IMC chamber with light doors calling it a day!
[Manasa, Jamie, Jenne]
PZT1 has been removed, and is wrapped in foil and stored in a (labeled) plastic box.
We beeped the cable between the cable holder bracket on the in-vac table, and the outside of the feedthrough. Things are mirrored, so pins 1,14 (one edge on the feedthrough) go to pins 13,25 on the in-vac cable bracket.
We checked the epics controls from the TT screen that Jamie made a while back (accessible from the ASC tab on the sitemap) to the output of the AI board. Things were very weird, but Jamie fixed them up in the model, then rebuilt and restarted the ASS model so that now the epics channel corresponding to, say, UL actually actuates on the UL output of the boards.
We tested the cables from the rack to the feedthrough, and discovered that they are also mirrored, to undo the mirroring between the feedthrough and the in-vac bracket.
Jamie made an adapter cable to take the pinout of the coil driver boards correctly to the pinout of the quadrupus cable, through this double-mirroring (i.e. no net mirror effect).
We set up a laser pointer on a tripod outside the door of the MC chamber (where the access connector usually is), and pointed it at the back of the TT. Den or whomever put the cable on the TT didn't follow the diagram (or something got messed up somewhere), because when we actuate in pitch (+ on the uppers, - on the lowers), we see the TT move in yaw, and vice versa.
We are in the process of removing the quadrupus from the TT, figuring out which connector goes where, putting it on correctly, and re-testing.
Koji, Manasa, Jenne, Jamie, Bob and Steve
Access connector removed this morning and work has began in the IOO chamber. BE WARE OF ANTS !
Just in case we want to retrofit the Tip/Tils with Beryllium Copper wire, here are links to a few sources which have a supply of the right composition and temper:
I don't think its worth it to do something to modify them unless we get a real reduction in the hysteresis - need a benchtop test setup ASAP.
Be Copper in the lab is from Ca Fine Wire : alloy 10 CDA 17 in sizes .008" & 0.002" There are other sus wire choices in the Drever lab
Today I measured the dark current of the PDA10CF. The output of the PD was connected to the A channel of the network analyzer, when there was no light falling on it. The response is collected using GPIB.
I will upload the result shortly.
[Jenne, Manasa, Jamie]
Now that we're up to air we relocked the mode cleaner, tweaked up the alignment, and looked at the spot positions:
The measurements from yesterday were made before the input power was lowered. It appears that things have not moved by that much, which is pretty good.
We turned on the PZT1 voltages and set them back to their nominal values as recorded before shut-down yesterday. Jenne had centered IPPOS before shutdown (IPANG was unfortunately not coming out of the vacuum). Now we're at the following value: (-0.63, 0.66). We need to calibrate this to get a sense of how much motion this actually is, but this is not insignificant.
I took the 2 G&H mirrors that we de-installed from PR3 and SR3 over to GariLynn to measure their phase maps. Data is in the same place as before, http://www.ligo.caltech.edu/~coreopt/40MCOC/Oct24-2012/ . Optic "A" is SN 0864, and optic "B" is SN 0884, however I'm not sure which one came from which tip tilt. It's hard to tell from what photos we have on picasa.
Both are astigmatic, although not lined up with the axes defined by where the arrow marks the HR side. Both have RoCs of -600 or -700m. RMS of ~10nm.
At least, we don't want to use Al-coated mirrors. We should use multilayer dielectric mirrors.
I popped into the cleanroom earlier today, and all 4 active TTs have dielectric coatings. I'm not sure why the mirror in this photo looks funny, but the actual mirrors installed are correct, at least in type of coating.
I'm not sure if Den wrote down what mirrors are actually in there, and I didn't look carefully - I don't know if they are G&H, CVI, other mystery company?
The 40m vacuum envelope is at atm It took 4.5 hours at ~3 Torr/min speed. Atm1
What I did:
Turned PZT voltages off: 84.1V at OMC-rack location
- 99V at 1Y3 location as PZT_pitch,
56.6V KEPCO at 1Y3 location as PZT_yaw
Took screen shuts of SUS_SUMMARY, yesterday, Atm3
Turned off OPLEV _SERVOs, today , Atm2
Closed PSL_SHUTTER and green shutters at the ends
Set vacuum valves for vent position and opened VM2 for the RGA
12/19: First thing in the morning we take off the access connector ONLY. The access connector is all we'll need to replace PZT1. Put light door on the OMC chamber side immediately, since we won't need to get in there at all. We won't need the light access connector.
The access connector is all we'll need to replace PZT1.
Really? Can you be sure of the input pointing this way?
...Nah. The projector is pretty new (t<1yr) and this is the first time to have the lamp busted after the installation last year in Jan.
We just should purchase two bulbs.