Morning Purge Air Checks 5-20-25

Dry air skid checks, water pump, kobelco, drying towers all nominal.

Dew point measurement at HAM1 purge inlet port, prior to pumpdown, -43.7 °C. 

The HAM1 purge exhaust dewpoint at top of chamber was measured to be -40.3C just before turning off the purge and starting pumpdown.

HAM1 pump down, monitor vacuum pressure with temporary EPICS IOC

Jonathan, Patrick, Janos, Ryan, Dave:

The HAM1 pressure gauge PT100 is temporarily only available as a raw voltage signal (H0:VAC-LY_X0_PT100B_PRESS_VOLTS).

I wrote an EPICS IOC which reads this voltage signal, converts it to a pressure (Torr) and exports it as a "H1" version of the PT100B cold cathode signal. The new EPICS channel is:

H1:VAC-LY_X0_PT100B_PRESS_TORR

The new IOC is currently running in a tmux session on opslogin0. It was started at 11:31 Tue20may2025, at which time HAM1 was still at atmospheric pressure (10.0V, 750Torr)

WP12558 was opened to add this new channel to the DAQ so that this week's HAM1 pump down could be trended.

We restarted the DAQ 0-leg at 12:16, the EDC at 12:17 and the 1-leg at 12:22

I took the opportunity to complete WP12523 and add the new CDS rack power strip EPICS channels to the EDC.

EDC details:

Addition of H1EPICS_VACHAM1.ini (3 channels) and H1EPICS_PWRSTRIP.ini (36 channels) for a total additon of 39 channels, which give a new total of 59,311 chans in the EDC.

I added the new HAM1 pressure channel to the vacuum overview (see attached).

The pump down was already on its way when we added it to the DAQ, attached trend shows data is available from 12:22 onwards, at which time the pressure was at 200Torr

Tue CP1 Fill

Tue May 20 10:05:48 2025 INFO: Fill completed in 5min 44secs

 

Loss of some h1digivideo5 cameras on Friday 16th May 2025

FRS34130

On Friday 16th May 2025 several cameras whose processes are running on one of the new AMD EPYC 9124 servers stopped working. Their MEDMs went "white-box" and their camera images went "blue-screen". The cameras and their last good data points are:

FC2 04:02

FC_TRANS_B 09:00

ETMX 10:03

ETMY 11:01

Note ETMX h1cam25 was not pingable, all the others were.

Yesterday I went to EX and EY to investigate. I power cycled both end station cameras by disconnecting their ethernet from the POE injector. After restarting their processes they came back online.

On the way back I did the same for the FC_TRANS_B from the CDS mezzanine in the FCES.

This morning I found that FC2 had also stopped (more about that laater) and got it running again by just restarting its process, no camera power cycle needed.

Looking at the logs we found 

"WaitObject duplicate failed (0): Too many open files. Reached open files limit"

type errors around these times. Debian has a limit of 1024 open file descriptors per process, and scanning the /proc/<pid>/fd/ dir sizes for h1digivideo5 showed that FC2 had 1024 open, which is how we found that camera had stopped working at 04:02 Friday.

Two immediate action items:

1. install num_open_fd check in cds_report for camera processes to trend and alert on an hourly basis if we are getting close to the limit
2. investigate if pylon code is responsible and if so can it be fixed.

 

Ops Day Shift Start

TITLE: 05/20 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
OUTGOING OPERATOR: None
CURRENT ENVIRONMENT:
    SEI_ENV state: MAINTENANCE
    Wind: 4mph Gusts, 1mph 3min avg
    Primary useism: 0.03 μm/s
    Secondary useism: 0.28 μm/s
QUICK SUMMARY: Pumpdown of HAM1 starts today as the main activity. The LVEA is currently Laser SAFE.

Workstations updated

Workstations were updated and rebooted.  This was an OS packages update. Conda packages were not updated.

VACSTAT single alarm on BSC3 gauge glitch

VACSTAT went into sensitive mode after BSC3 PT132_MOD2 glitched, spending 2 seconds at 3.2e-07Torr before returning to 1.5e-07.

I have restarted VACSTAT to clear this glitch.

HAM1 VAC works

Gerardo, Jordan, Travis, Randy, Janos

Today between 10 am and 2 pm both the Y+ and Y- HAM1 doors have been installed. The installation went smoothly, no issues were found.
A valve on the new interlock gauge-assembly was opened, and the HAM1 volume will be purged overnight. This is to clean the in-vacuum surfaces before pumpdown, resulting in lower gas load and therefore faster pumpdown.
After the door installation, the pumpdown of the joint HAM1/HAM2 Annulus volumes have been started. 2 turbo/pumping cart systems are pumping the volume, they are standing next to HAM1 and HAM2 - see attached photo.

Please do not move the pumping carts, as it can result in immediate damage in the turbo pumps. If you still need to do any activities which need these pumping carts to be moved, please talk to vacuum team members first!

Ops Day Shift End

TITLE: 05/19 Day Shift: 1430-2330 UTC (0730-1630 PST), all times posted in UTC
STATE of H1: Planned Engineering
INCOMING OPERATOR: None
SHIFT SUMMARY:

Doors are on HAM1 and things are being wrapped up in order to start pumping down!
LOG:

Start Time	System	Name	Location	Lazer_Haz	Task	Time End
14:15	FAC	Kim, Nellie	LVEA	n	Tech clean	14:58
15:20	VAC	Jordan	LVEA	n	Turning down purge air on HAM1	15:27
15:51		Oli	LVEA	YES	Transitioning LVEA to laser hazard	16:06
15:59	-	Corey	EX	n	Wallet hunt	16:19
16:00	FAC	Kim, Nellie	EX	n	Tech clean	17:08
16:11	OPS	RyanC	LVEA	Y	Restart DM5	16:11
16:00		Camilla	LVEA	YES	Turning down power	16:37
16:00	SUS	Jeff, Rahul	LVEA	YES	Checking RM pointing (Jeff out 16:40)	16:43
16:00	EE	Fil	CER	YES	Checking if PM1 is jumpered	16:14
16:39	EE	Fil	LVEA	n	Running cables	18:21
16:40	EX	Eric	EX	n	Looking for hot breakers	17:59
16:40		Camilla	LVEA	YES	Transitioning us back to laser safe	16:48
16:44	ISC	Betsy	LVEA	n	Last minute checks on HAM1 table	17:40
16:57	SUS	Jeff	CER	n	Checking SUS coil drivers	17:00
17:04	VAC	Travis	LVEA	n	Prepping for doors	18:34
17:09	FAC	Kim, Nellie	EY	n	Tech clean	17:50
17:09	PCAL	Tony	PCAL Lab	y(local)	Packing up standards	18:28
17:24	VAC	Jordan, Janos	LVEA	n	HAM1 doors	18:41
17:26	VAC	Gerardo	LVEA	n	HAM1 Doors	18:41
17:44	FAC	Chris	MY	n	Checking noisy bearing	18:24
17:45	EE	Jackie	LVEA	n	Helping Fil with cables	18:21
17:53	SUS	RyanC	CR	n	OpLev charge meas for ends	19:21
18:16		Elenna	LVEA	n	Watching doors get put on	18:21
18:22	EE	Fil, Jackie	CER	n	ISC cabling for HAM1	19:36
18:34	ISC	Camilla, Betsy	LVEA	YES	Transitioning to Laser Hazard and checking beams at ports	19:07
18:57	FAC	Nellie	LVEA CERish	YES	Restocking garb	19:26
19:24	VAC	Jordan	LVEA	N	Door work	21:51
19:33	VAC	Janos, Gerardo, Travis, Randy	LVEA	n	Putting on second door (Janos, Travis, Randy out 21:30)	21:51
19:44	FAC	Eric	EY	n	Checking breaker panel temps	21:16
20:15	CDS	Jonathan, Dave	MSR	n	Swapping frame writer power supply	20:47
20:46		Jeff	LVEA	n	Returning laptop	20:48
20:47	TCS	Camilla	LVEA	n	Checking CO2X table parts (laser is off)	21:13
20:48	CDS	Dave	EY	n	Checking cameras	21:56
21:13	PCAL	Tony	PCAL Lab	y(local)	Starting measurement	22:46
21:40	FAC	Randy	LVEA	n	Helping with doors	21:51
21:46	EE	Fil	LVEA	n	Finishing up HAM1 cables	23:34
21:46		Richard	LVEA	n	Trying to not help out	21:55
21:56	CDS	Dave	EX	n	Checking other cameras	22:58
22:26	SEI	Jim, Mitchell	LVEA	n	Moving Class A stuff out of HAM3 cleanroom	22:43
22:41	VAC	Jordan, Gerardo	LVEA	n	Prepping HAM1 for pumpdown	23:34
23:17	TCS	Camilla	LVEA	n	Measuring a distance and looking in the TCS cabinets	23:34

TCS Chiller Sock Filters Replaced

FAMIS 31406

The sock filters had slight discolorization, but there were floating string-like particulate in both reservoirs (pic1 & pic2). This isn't a new finding, but I was hoping after the last flush in October that it would have reduced the amount of floaters over time as they were filtered out, but this doesn't seem to be the case.

HAM1 POP and REFL power budgets

Sheila and I used the thorlabs power meters to do a quick power budget along the POP and REFL paths. Just as a note, this alog from Craig and others in 2022 still holds and is probably the best authority for the REFL path.

Attached is a handy cartoon with the naming convention for HAM1, see D1000313.

POP path measurements:

POP path measurements were taken in "single bounce" 10 W input. So PRM misaligned, ITMX aligned. 9.4 W was incident on IM4 trans.

Expected power to HAM1 = 9.4 W * 0.977 (IFI transmission) * 0.031 (PRM transmission) * 0.5 * 0.5 (two passes of the BS) * 229e-6 (PR2 transmission) = 16.3 uW (this assumes the reflection off the ITM is 1)

The POP beam then sees a 90/10 split at M12 (air/vac) and a 50/50 split at M15 (LSC/ASC).

Location	Measured	Expected (how calculated)	Ratio Measured/Expected
After periscope	15 uW	16.3 uW (9.4 * 0.977 * 0.031 * 0.5 * 0.5 * 229e-6)	0.92
reflection of M12 (90/10 splitter for air/vac)	14 uW	14.7 uW (9.4 * 0.977 *0.031 * 0.5 * 0.5 * 229e-6 * 0.9)	0.95
transmission of M15 (just before POP ASC diode)	0.65 uW	0.81 uW (9.4 * 0.977 * 0.031 * 0.5 * 0.5 * 229e-6 * 0.1 * 0.5)	0.8

The beam is hard to see until after the lens so these were the only measurements possible.

REFL path measurements:

These measurements were taken in two sets, with the first set at 10 W input with PRM misaligned and the second set at 2 W input with PRM aligned and ITMX misaligned.

Set 1 is single bounce with PRM misaligned and 9.4 W on IM4 trans.

Expected power to HAM1 = 9.4 W * 0.977 * 0.031 * 0.5 *0.5 * 0.031 * 0.977 (double pass of beam splitter, PRM and IFI) = 2.15 mW (this assumes that the backward throughput of the IFI is the same as the forward throughput)

Location	Measured	Expected (how calculated)	Ratio Measured/Expected
Before M14 (beam just coming to HAM1)	2.05 mW	2.15 mW (9.4 * 0.977 *0.031 *0.5 * 0.5 * 0.031 * 0.977)	0.95
reflection of M14 (90/10 splitter)	1.9 mW	1.94 mW (9.4 * 0.977 * 0.031 *0.5 * 0.5 * 0.031 * 0.977 * 0.9)	0.98
transmission of M17 (first 50/50)	0.1 mW	0.11 mW (9.4 * 0.977 * 0.031 *0.5 * 0.5 * 0.031 * 0.977 * 0.1 * 0.5)	0.91
transmission of M1 (second 50/50)	40 uW	53.9 uW (9.4 * 0.977 * 0.031 *0.5 * 0.5 * 0.031 * 0.977 * 0.1 * 0.5 * 0.5)	0.74

Set 2 is single bounce with PRM aligned and 2 W input, so 1.9 W on IM4 trans.

Expected power to HAM1 = 1.9 W * 0.977 (IFI throughput) * 0.97 (PRM reflection) * 0.977 (IFI throughput)= 1.76 mW

Location	Measured	Expected (how calculated)	Ratio Measured/Expected
before M2 (third 50/50 splitter)	64 mW**	44 mW (1.9 W * 0.977 * 0.97 * 0.977 * 0.1 * 0.5 * 0.5)	1.45
before M6 (LSC/ASC splitter)	14 mW	22 mW (1.9 W * 0.977 * 0.97 * 0.977 * 0.1 * 0.5 * 0.5 * 0.5)	0.64
transmission M6 (50/50 LSC/ASC splitter)	7 mW	11 mW (1.9 W * 0.977 * 0.97 * 0.977 * 0.1 * 0.5 * 0.5 * 0.5 * 0.5)	0.64
reflection of M18 (50/50 LSC A/B splitter)	3.5 mW	5.5 mW (1.9 W * 0.977 * 0.97 * 0.977 * 0.1 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5)	0.64
transmission of M18 (50/50 LSC A/B splitter)	3.5 mW	5.5 mW (1.9 W * 0.977 * 0.97 * 0.977 * 0.1 * 0.5 * 0.5 * 0.5 * 0.5 * 0.5)	0.64

** we made this first measurement as a quick check to see what power was going to the other side of the table, but clearly something is wrong with it! Maybe I transposed a number?

While there is clearly a discrepancy between what light we measure vs expect at M6, all the splitting ratios after seems to make sense. Craig calculated what the true splitting ratios for M14, M17 and M1 in alog 63510. None of these optics have changed. The splitting ratios are as follows:

M14 trans = 0.0756

M17 trans = 0.5072

M1 trans = 0.5295

If I update the calculations of the REFL path using those numbers, I get the following, with bolded values noting which ones have changed:

Location	Measured	Expected (how calculated)	Ratio Measured/Expected
Before M14 (beam just coming to HAM1)	2.05 mW	2.15 mW (9.4 * 0.977 * 0.031 *0.5 * 0.5 * 0.031 * 0.977)	0.95
reflection of M14 (90/10 splitter)	1.9 mW	1.99 mW (9.4 * 0.977* 0.031 *0.5 * 0.5 * 0.031 * 0.977 * (1-0.0756))	0.95
transmission of M17 (first 50/50)	0.1 mW	0.083 mW (9.4 * 0.977 * 0.031 *0.5 * 0.5 * 0.031 * 0.977 * 0.0756 * 0.5072)	1.2
transmission of M1 (second 50/50)	40 uW	43.8 uW (9.4 * 0.977 * 0.031 *0.5 * 0.5 * 0.031 * 0.977 * 0.0756 * 0.5072 * 0.5295)	0.91

 

Location	Measured	Expected (how calculated)	Ratio Measured/Expected
before M2 (third 50/50 splitter)	64 mW**	35.7 mW (1.9 * 0.977 * 0.97 *0.977 * (0.0756) * 0.5072 * 0.5295)	1.8
before M6 (LSC/ASC splitter)	14 mW	17.8 mW (1.9 * 0.977 * 0.97 *0.977 * (0.0756) * 0.5072 * 0.5295 * 0.5)	0.79
transmission M6 (50/50 LSC/ASC splitter)	7 mW	8.9 mW (1.9 * 0.977 * 0.97 *0.977 * (0.0756) * 0.5072 * 0.5295 * 0.5 * 0.5)	0.79
reflection of M18 (50/50 LSC A/B splitter)	3.5 mW	4.46 mW (1.9 * 0.977 * 0.97 *0.977 * (0.0756) * 0.5072 * 0.5295 * 0.5 * 0.5 * 0.5)	0.78
transmission of M18 (50/50 LSC A/B splitter)	3.5 mW	4.46 mW (1.9 * 0.977 * 0.97 *0.977 * (0.0756) * 0.5072 * 0.5295 * 0.5 * 0.5 * 0.5)	0.78

 

Checking ASDs of OSEMINF_OUTs on RMs

Jeff, Oli

Following up on one of Jeff's followup tasks about the signs of RM1 and RM2 (84462):

"(II) Measure the amplitude spectral density of each individual OSEMs well above the data rate we typically store them; look to see if there's no giant noise features. If there is -- debug that (usually a reboot of the satamp is the first step.)"

I ran ASD measurements for RM1 and RM2 at OSEMINF_OUT for each OSEM (with the comb60 filter turned off), and we verified they all look good(attachment1). To further check and compare them, we reran the same template (and filter) setup but for OM1 and OM2(attachment2). Since doors are currently being put on and the HAM1 ISI isn't isolated yet, we can see that the RM OSEM spectrum are very noisy below 30 Hz due to seismic noise, and other noises caused by the immediate environment are causing peaks at several different frequencies for thre RMs as compared to the OMs, but overall all four suspensions' OSEMs asymptote the same way and there aren't any noise features that seem to be of concern.

WP 12455 Replace failed power supply on h1daqframes-1

As per WP 12455 Dave and Jonathan replaced the power supply on h1daqframes-1.  Due to the available space in the rack we decided to power off the system so that we could move it safely and replace the power supply.

We did some double checking to verify which machines where hooked together.  We followed the physical connections and cross checked them with the names and numbers for interfaces on each machine.  This was good as the labels on the front of the system where wrong.  This is the frame disk associated with h1daqfw1.

We powered down h1daqfw1, verified that the link to h1daqframes-1 had gone dark.  After that we powered off h1daqframes-1 pushed it forwared enought to replace the power supply, pushed it back in th place and restarted it.

After the disk server was back we restarted h1daqfw1 as well.

Dave will fix the labeling.

Physical Check of Actuation Sign for RM1 and RM2 as they Currently Stand; OK for doors, but we've got work to do.

R. Kumar, C. Compton, J. Kissel
(Using Cartoon from other files of D1000313-v19)
(Following sign conventions from T1200015)

Given the sign confusion of the RMs (LHO:84289), I wanted a physical, with-our-eyeballs, confirmation of how the RMs deflect the beam under different alignment offsets given the current state of the system while we still had doors off HAM1. Using the REFL beam as our optical lever, we drove +/-20000 [ct] alignment offsets into each suspensions H1:SUS-RM[1,2]_M1_OPTICALIGN_[P,Y]_OFFSET, and watched the displaced beam downstream of the optic. Rahul was in chamber with the card and his eyeballs, Camilla was chamberside in view of the card as a second set of eyeballs to confirm, and I was driving with a CDS laptop and taking notes. Detailed blow-by-blow and raw convention-free data can be found in the attached text file.

Conclusions (all statements made with "Today," preceding the statement):
(0) We checked that 
  On the actuation side,
    (a) H1:SUS-RM[1,2]_M1_OPTICALIGN_[P,Y]_GAINs are all positive, at +1.0;
    (b) The sign of the pitch and yaw elements of the EUL2OSEM matrices for both RM1 and RM2 match convention;
    (c) There's no sneaky sign flop in the *filters* of the COILOUTF banks, i.e, in none of the "LPM1" or "AntiLPM1" filter modules (see (1) below for statement on sign of gains)
  On the sensor side
    (d) The raw ADC counts for both RM1 and RM2 are positive, as expected after the cable swap, and per a normal OSEM sensor;
    (e) The sign of the OSEMINF bank gains are all positive values close to +1.0, as expected for a standard OSEM sensor chain;
    (f) There's no sneaky sign flip in the *filters* of the OSEMINF banks, i.e. in none of the "10:0.4", "to_um," or "comb60" filter modules
    (g) The sign of pitch and yaw elements of the OSEM2EUL matrices are the correct transpose (in the mathematical sense) of the EUL2OSEM matrices, so they match convention

(1) the COILOUTF gains are set as they were reverted to their "been like this forever value" on May 6th LHO:84289, with 
    (a) RM1 obeying convention (UL, LL, UR, LR = +, -, -, +) and 
    (b) RM2 disobeying convention (UL, LL, UR, LR = +, -, -, +).

(2) RM1 displaces the beam per convention, with a
    (a) positive pitch offset displacing the downstream beam down and 
    (b) positive yaw offset displacing the beam in +Y (IFO Cartesian coordinates) i.e. rotating RM1 in +RZ, or +yaw,
    at RM2,

(3) RM2 displaces the beam against convention, 
    (a) positive pitch offset displacing the downstream beam up and 
    (b) positive yaw offset displacing the beam also +Y (IFO Cartesian coordinates) i.e. rotating RM2 in -RZ, or -yaw,
    at M5, 

(4) Under these displacements, 
    (a) the RM1 OSEM sensor sign agrees with the physical displacement: 
        (i) positive requested pitch shows up as more positive pitch OSEM sensor value, 
        (ii) positive requested yaw shows up as more positive yaw OSEM sensor value, 
    (b) the RM2 OSEM sensor signs disagrees with the physical displacement: 
        (i) positive requested pitch displaces the beam up (negative pitch), so the optic has physically tilted up in negative pitch, but the OSEM reads this as positive pitch
        (ii) positive request yaw displaces the beam in -RZ (negative yaw), so the optic has physically rotated in negative yaw, but the OSEM reads this as positive yaw.

In each of the attach trends of the test -- RM1 PIT, RM1 YAW, RM2 PIT, and RM2 YAW -- that show
    - The digitally requested alignment offset
    - The OSEMINF raw ADC channels (just to show that these are all positive now, after the cable work discussed in LHO:84027; you can't actually see the displacement)
    - The projected Euler basis OSEM sensors
    - The REFL WFS DC QPD signals (these didn't turn out to be that useful because we didn't take care to let the beam settle at each alignment, and predicting what the sign of the measured beam spot downstream of the WFS lens is not straight-forward)

All of the above conclude that
There are no sign issues in the sensor or actuator chains for RM1
Given the combination of (1)(b) and (3)(a) + (3)(b), we conclude that
There is *no* magnet polarity problems, and we *could* make RM2 actuators obey actuation convention without any hardware change.

Instead, from (4)(b)
There remains a sign issue with the OSEM *sensor* chain for RM2.

And yet, even in the face of the *difference* in RM sensor sign, (3)(a) vs. (3)(b), 
With what *very* low level "does it work or not" MEDM screen viewing that we've done, *both* RMs "need" the same *positive*, +1, damping loop gain which disobeys convention.
So -- 
   - Why do the RM2 sensors readout opposite from physical displacement? 
and
   - Why don't RM1's damping loops work with a -1.0 damping gain?

From here we need to:
(I) Compare the *phase* of now vs. old "damping loop OFF" "health check" transfer functions. If these show a sign change after the cable swap changed the OSEM sensor sign, good.
(II) Measure the amplitude spectral density of each individual OSEMs well above the data rate we typically store them; look to see if there's no giant noise features. If there is -- debug that (usually a reboot of the satamp is the first step.)
(III) If/once there's no gross electronics noise, after the dust settles in the chamber, with doors on, turn on the damping loops with +1 gain. 
(IV) Measure the open loop gain transfer functions (drive M1_DAMP_EXC, measure TF of IN1/IN2), and confirm stability and appropriate amount of gain; they should look like they did when I improved the damping loop filters back in 2022. LHO:63656. Debug further if they don't.

SR3 heater on overnight

TJ, Camilla, Sheila

TJ and Camilla got the HWSs working, and now we turned on the SR3 heater at 15:54 UTC, 2W requested power.  This is to do a check of the SR3 heater calibration and range similar to 27262

HAM1 POP path beam profile and distance measurements

location number on drawing	distance	horizontal 13.5% diameter [um]	vertical 13.5% diameter [um]	photo of profiler location	photo of profiles	photo of beam scan measurements
1	52 mm from dichroic M10	6243	64040	here and here	9022	9023
2	147 mm from dichroic M10	6202	6365	9028	9026	9027
3	119 mm from HR of 50/50 BS M15, also 295mm from center of lens L2	870	874	9031	9030	9029
4	153 mm from HR of 50/50 BS, approximate location of LSC diode	279	284	9035	9032	9033
5	128 mm from HR of 50/50 BS	721	726	9036	9037	9038
6	139 mm from HR of 50/50	483	490	9041	9040	9039
7	353 mm from HR of 50/50 BS	3280	3372	9042	9043	9044

Camilla made these measurements with 20W input power into the IMC, PRM and ITMY misaligned single bounce beam.  We didn't place the 90/10 BS M12 in the pop path yet so that we would have about 35uW to measure beam profiles. There's a rough pen sketch of where these locations are in this photo.

Camilla also made ruler measurements of some distances:

• PM1 to 50/50 BS M15 is 405mm
• PM1 to center of L2 is 229 mm
• center of L2 to M15 is 178 mm
• dichroic M10 to PM1 is 360 mm
• bottom of periscope  (M8 or M9) to dichroic M10 482mm
• HR side of M15 (50/50 BS) to front of LSC diode box is 158 mm, diode itself is 3 mm proud of diode box.