Summary of BCU Sensor Data Range and Calibration Parameters

References: nspotr_BCU-test_BI.doc nspotr_BCU-test.doc nspotr_BCU_operation_Manual_to_RU.doc nspotr_BCU_Cmd.doc nspotr_bcu_DataFmt.doc nspotr_bcu_elec_connect.doc

Affirmed on 5 January , 2008, revised 12 January, 03 July , 2009 , 20110529

by Prof. T.L. Yeh, Co-PI, Engineering, , 2008

Approved on , 2008

By Prof. H.C. Yeh, PI, , 2008

> From: Mon, 22 Dec 2008 18:29:34 +0800
The BCU is only the payload, which did not pass the final functional test at the factory.
TA: 1. Numerical values and dynamic ranges of variations of the BCU parameters;
TA: 2. Key numbers of calibration of the BCU sensors;
TA: 3. Calibration graphs for each sensor
** tly: Please do send us any data log from BCU for examination.
** tly: Please let me know what is the difficulty the Russian Team encounters in making judgement on the functionality of BCU.
** tly: Let me have / see what they get.
** tly: we can always provide more info to assist, Please Do Show Me A Sample of What You Need.
>tly:?!? > alex: Prof. Yeh can not add any more.
>tly: To run BCU test through BI, check this manual nspotr_BCU-test_BI97b11.doc
> alex: nspotr_BCU_operation_Manual_to_RU_morphous080413tly429.doc - that is completed Operation Manual (инструкция по эксплуатации). The description of functional tests is presented in Section 2.3.3.2.1 Concise Workability Check (for the Russian Team)
> alex: nspotr_BCU-test97b10.doc - that is manual for functional test.
---- rev 20110529 Subject: nspotr mrm std畫圖
To: nspotr_papermvmc, nspotr_ss, nspotr_sso
Bcc: edupaper, edupapersh, edupaperother, family0, family0_, familyc0
major revision: for your information nspotr_bcu_sensor_data.doc
http://mvmc.me.ncu.edu.tw/nspotr/nspotr_bcu_sensor_data.doc

I.  INTRODUCTION

The sensor data from the flight unit BCU are generated by its BCU Main Board labeled with "BCU V1.1, ETP V1.0, and MRM V1.3".

Ref: nspotr_bcu_DataFmt.doc

1)  Each Sequence of Sampling Data Contains 15 words: {Sequence Type and Calendar Info in 2 words, 3 pairs of VTEL (ETP-OUT) and VF of 1 word each for the 3 ETP phases, 2 sets of 3 axes MRM data of 1 word each after set and reset pulse, and 1 word of VOUT1_T1 of TMP36 data}.

2)  Each Sequence of System Data Contains 9 words: {Sequence Type and Calendar Info in 2 words , 1 word of system parameters, and 6 TMP36 data of 1 word each}

Sensor data, namely {VTEL (i.e. ETP-OUT), VF, MRM-x-set, MRM-y-set, MRM-z-set, MRM-x-reset, MRM-y-reset, MRM-z-reset, TEMP1, TEMP2, TEMP3, TEMP4, TEMP5, TEMP6}, are represented in signed 16 bit integer with value range in [–32768, +32767] representing analog signal of [0, 2] volt. All signals are scaled and offset to fit in this range. The ADC has 14 bit resolution and the ADC result is left justified in the 16 bit data word.

ADC channel 0 to 7 corresponding to data MRM-x, MRM-y, MRM-z, VTEL, Temp1 or 2, Temp3 or 4, Temp5 or 6, and VF respectively. Temp1 3 5 and Temp2 4 6 are multiplexed by 0 and 1 on ADC channel 4, 5, 6. However, in the flight unit, TMP36 sensors #1 #3 #4 and #2 #5 #6 are connected to Temp 1 3 5 and 2 4 6 locations instead. Ref: Fig 1.a 1.b and 2.c in nspotr_bcu_elec_connect.doc

II.  CALIBRATION PARAMETERS of THE SENSORS

II.1 ETP-OUT VTEL and ETP-VF:

The flight data of ETP produces ETP-OUT VTEL and ETP-VF in volt. Post data processing incorporating other orbital information is necessary to obtain accurate Electron Temperature Data. Therefore, in this document we provide only calibration curve to convert ADC data to voltage difference between the sine wave driven probe ETP#1 and the reference probe ETP#2 as ETP-OUT VTEL, and the voltage at the reference probe as ETP-VF. (Ref: )

1)  Calibration Curve to Obtain ETP-VF Voltage from A7T0 Data:

(Note: The A7T0 data may vary over the full range of a singed 16bit integer [-32768, +32767].)LSE parameters obtained by merging both data sensrep976192316A7T0.txt and sensrep976261909A7T0.txt :

VF = (data_A7T0 + 23260 ) / 13816 +-0.0411

Ref: [sensrep976261909A7T0.txt note0712.txt a7261909.m],
VF = (data_A7T0 + (22886 ) ) / (13602) +-0.03
/ Ref: [sensrep976192316A7T0.txt a7192316.m],
VF = (data_A7T0 + (23474) ) / (13942) +-0.0407

2)  Effect of Common Mode Voltage of ETP#1 and ETP#2 on ETP-OUT VTEL A3T0 Data:

Ref: [sensrep976191503A3T0.txt a3191503.m],

(data_A3T0 + l5022 +-591.05 ) = (-24.4872) * VTEL_VF_Comm

3)  Calibration Curve to Obtain ETP-OUT VTEL Voltage Difference from A3T0 Data:

The voltage difference between the two ETP probes (ETP#1_VTEL – ETP#2_VF) due to induced electron current flowing into ETP#1 of input impedence (110M +-1%) ohm is estimated by the following regression formula:

(Note: The A3T0 data may vary over the full range of a singed 16bit integer [-32768, +32767].)

Ref: [sensrep976191513A3T0.txt a3191513.m],

(VTEL-VF) = (data_A3T0 - ((VTEL+VF)/2)*(-24.4872) + 14873 ) / (-136340) +-0.0041

In summary, evaluating VTEL@ETP#1, VF@ETP#2, and the difference dV=(ETP#1-ETP#2):

a) VF =0.7238e-4*data_A7T0 +1.684 +-0.041,

b) VTEL = 0.7237e-4*data_A7T0 -0.7334e-5*data_A3T0 +1.574 +-0.045,

c) dV@(110M +-1%) ohm = -0.1300e-7*data_A7T0 -0.7334e-5*data_A3T0 -0.1094 +-0.0041

note 99202: VTEL dV -0.35~0.13V @110Mohm增益-0.73e-5V/count 偏移-14900 標準差 0.004V, VF –0.68~4.05V 增益0.72e-4V/count 偏移-23300 標準差 0.04V, by “solve({dv=-0.1300e-7*data_A7T0 -0.7334e-5*data_A3T0 -0.1094,VF=0,VF=0.7238e-4*data_A7T0 +1.684,data_A3T0=-32767},{dv,data_A7T0,data_A3T0,VF});”

Ref: BCU_V11_ckt_pic\DSP_MRM_ETP_V11\ETP\PreAmp¥ Pre_amp_input.gif .jpg BCU_V11_ETP_PreAmp.jpg, mvmc-nspotr¥nspotr_BCU_sensor_data.mws

rev 20110505 Note for 3) Calibration Curve to Obtain ETP-OUT VTEL Voltage Difference from A3T0 Data : deriving etemp from Vtel and VF:

* Vtel is driven by carrier since wave voltage signal in 3 phases sequentially of 0.15 sec width each: #1 1Vpp 0.5Vpp and 0Vpp. In each phase induced electron current flow into ETP#1 and the input resistance 100Mohm +-1% to generate differential voltage. It is filtered and measured as the averaged offset to reflect the electron current in response to the driving voltage dV1 dV2 dV3: (dV1-dV3), (dV2-dV3), and ratio (dV1-dV3)/(dV2-dV3) can all be used to calculate etemp. If the electron energy distribution is Maxwellian, their results should be consistent.

= phases timing of BCU-Tatiana2: ... %old nspotr98b24rep¥nspotr98b24rep.ppt , nspotr99531TripRuSatBCU.vsd _timing.gif , nspotr96727rep.vsd , nspotr-asc¥OverallSampleTiming.vsd ...lefteye080820tly.gif

= vTel changes after the switching of carrier amplitude at each phase D:\temp\mvmc\mvmc-nspotr\morphous\Test_Calibration\ETP_Test_Result\100M_10M¥ { TEL_OUT.jpg (1st_Phase_1Vpp.JPG, 2nd_Phase_05Vpp.JPG, 3rd_Phase_0Vpp.JPG) VF.JPG}

= data analysis program by Dr Kakinami: etemp {te1 te2 teR} are calculated based on {Vfa1 = ETP3dv-ETP2dv; % 0.25 V , Vfa2 = ETP3dv-ETP1dv; % 0.5 V , R = Vfa2./Vfa1; } ; ref: calTeB.m in http://mvmc.me.ncu.edu.tw/nspotr/science/nspotr99504scienceDataCleanKakinami/nspotr99504tatiana_data_programsKakinami.zip D:\temp\mvmc\mvmc-nspotr\nspotr98a19TereshkovKaluga\orbit-jhk482001\nspotr99504tatiana_data_programsKakinami.zip ; ref: http://mvmc.me.ncu.edu.tw/nspotr/kakinami_diry.htm d:¥temp¥mvmc¥mvmc-nspotr/kakinami_diry.htm

= sample BCU-Tatiana2 in-flight data: VF1 VF2 VF3 Vtel1 Vtel2 Vtel3 dV1 dV2 dV3 MRMx y z http://mvmc.me.ncu.edu.tw/nspotr/orbit-jhk482001/bcu2009-12-20_10-55-54-113b000latlonPhys121908-12ut.gif D:\temp\mvmc\mvmc-nspotr\nspotr98a19TereshkovKaluga\orbit-jhk482001¥bcu2009-12-20_10-55-54-113b000latlonPhys121908-12ut.gif

= sample BCU-Tatiana2 in-flight data: dV1 dV2 dV3 and etmp1 etemp2 etemp_ratio calculated by Dr Kakinami http://mvmc.me.ncu.edu.tw/nspotr/orbit-jhk482001/bcu2009-12-20_10-55-54-113b000ETPtempk121908-12ut99126kakinamitly.gif D:\temp\mvmc\mvmc-nspotr\nspotr98a19TereshkovKaluga\orbit-jhk482001¥bcu2009-12-20_10-55-54-113b000ETPtempk121908-12ut99126kakinamitly.gif

II.2 MRM:

The magnetic field (flux density) component along an MR sensing element is measured by the difference between the sensor read out after the set pulse and that after the reset pulse applied to the sensor, e.g. (MRMx_Set – MRMx_Reset). The middle point between the two read outs is not necessarily zero. If either read out became close to the boundary of the ADC dynamic range, saturation occurs. The difference did no longer reflect the magnitude of the magnetic component due to saturation.

(Note: The A0T0, A1T0, and A2T0 data may vary over the full range of a singed 16bit integer [-32768, +32767].)

The MRM-x, MRM-y, and MRM-z correspond to the ADC channel A0T0, A1T0, and A2T0 of BCU respectively.

The MRM sensing elements have functional dynamic range of +- 2 Gauss.

The orientation of the three MRM sensing elements on the BCU Main Board with respect to the BCU box is described in "Fig 1.b The Assembling and the Coordinate System of BCU – Perspective View" (Ref: NCU_IF1_0.doc nspotr_BCU_assemble.vsd .gif) of nspotr_bcu_elec_connect97501.doc .

Calibration Function of MRM:

The calibration curve of the MRM on the flight unit BCU V1.1A is described below (Ref: BCU_MRCali_NSPO_071210_測試結果¥V11A_RESULT¥total¥MRM_analysis11a.m nspoMRcalib¥nspoMRcalib071210¥bcu_v11a_mrm_cali.doc).

Magnetic field components in nT along the 3 BCU axes shown in the lower left corner of the figure above can be obtained by data from the three MRM sensor components on the BCU main board. The calibration function makes use of the difference between the data after set and reset pulse on each MRM sensing element. MRM_x, y, z data is obtained from ADC channel 0, 1, 2 respectively.

rev 20110529: The following calibration was obtained by comparing the averaged sensor raw data counts with the magnetic field setting of the Helmholtz coil. For each test field value there are large number of data measured and averaged, therefore, this calibration has minimum std representing the linearity deviation over the full testing range [-0.5,+0.5]*1e5nT. Thus, the nominal linearity error in this dynamic range is +-0.026%.

X / -1.8477 / -0.0294 / -0.0676 / (Xset – Xreset) / -877.4915 / 25
B_ / Y / = / +0.0240 / -1.9008 / -0.0149 / * data_ / (Yset – Yreset) / + / -426.0594 / +-std / 26 / nT
Z / -0.1145 / +0.0248 / +1.8770 / (Zset – Zreset) / +493.3449 / 23

rev 20110529 Note: 你發現的 "原來 mophies作的 std 是 把 在太空中心測試時 同一環境磁場下

所有的數據 作成一個平均值 (其 N->inf, var(mean)->0)

再把 在所有測試條件下的 量測平均值與 設定磁場值 之間的 誤差 所做的 std 26nT

" 因此 是代表了 整個測試範圍內的

linearity deviation -> 因此是 線性度的量測: 線性度的比例值

會是 這個 std 26nT / 全測試域的範圍 (+0.5 - (-0.5))*e5nT

而 這個 26nT 不是 單一量測數據的 std - 精度,

----

作者 pondaniel (NICE~~彭彭!!!) Fri May 27 16:31:56 2011 看板 G_MVMC

標題 [討論] nspotr台俄校正數據

02_計畫相關資料¥060401_NSPO_台俄衛星¥nspotr¥20110527_test_calibrated_data¥total

1. 先run mrm_analysis11a.m

會把所需要的raw data 放到matlab

2. 再run test_gg.m

會把raw data 拿來運算

rev 20110529 Note: 單一量測數據的 std - 精度,

這個值應該是你用 每個數據與校正值 的差異 做出來的 std

或者是 你用 所有數據作 calibration matrix LSE 得到的

所有 residual 的 std 105.3 120.5 141.0 nT for x y z axis)

?...TBD...?

To: nspotr_papermvmc { }

Subject: nspotr mrm std畫圖

!! 請把 用所有原始數據 做出來的 整個 calibration matrix LSE 的結果的

校正 矩陣公式 與 std 估算值 都給我 以便完整納入 校正紀錄中

(格式請參考 原來用 一個平均值得到的:

nspotr_bcu_sense_data.doc [Calibration Function of MRM])

---

但是 單一量測的 std 與 std極限值(即linearity std)間的

漸近關係 是 我們現在想看看的 可以驗證 我們的量測值

裡面除了 persistent 的 quantization error

還是還有其他的 persistent 電路干擾 -- 那就不美了

(那是無法用 平均 趨近於零的 會一直上下擺盪)

這件事情 我會寫進 nspotr_BCU_sensor_data.doc 中 做為改正

請你一定要找 立業 跟他講一下 ,

不知道 他要投稿的論文裡面 有沒有要相對修改的地方

From: "Yuan-Lung,Peng" <> Sat, 28 May 2011 23:07:52 +0800

Subject: std畫圖

老師你說的是這個網頁講的嗎??

http://en.wikipedia.org/wiki/Standard_deviation#Relationship_between_standard_deviation_and_mean

--- tly:

你找到的 reference 很棒,

我希望看到的就是 std_mean(N) 與 1/sqrt(N) 的作圖

註: 網頁裡的 std 用 symbol font 的 s 來表示 (小寫 sigma)

意義是 "large number theorem": 若 亂數是 independent random

則 用他們的N個數做出來的平均數mean 的變異量var(mean)

與原來亂數的變異量var(X) 間的關係 會有 1/N的 關係

若真是如此 則 我們可以很準確的估計 需要多少精度

就需要作多高倍數的 over sampling 來得到 適當精度的 mean

做為我們的量測值

現在 需要驗證 我們的原始數據的統計特性 的確 有這種表現

才能在後續 做這種規劃

--- EO rev 20110529

II.3 TEMP:

The typical calibration curve is shown below. According to TMP36 data sheet (ref: Analog Devices TMP35_36_37.pdf), its working range is [-40, +150] deg C. Therefore, the dynamic range of TEMP1 … TEMP6 can be calculated as ( [-40, +150] + [-2.22, +2.22] –3.81 -50 ) * (31598/100) = [-30344, +31096].

Note: TMP36 Rout = ( Vout_max=2V / Iout_max=50uA ) = 40KW, in direct measurement by a volt meter, loading effect by the volt meter must be considered if RV is not larger than 5MW.

Note: 2.3.3.2.2.B) Calibration Data and Dynamic Range of Sensor Responses in nspotr_BCU_operation_Manual_to_RU.doc (_morphous080413tly429) , nspotr97616¥sensrept.m based on sensrep97617*A{4|5|6}T1.txt (nspotrnote97617.txt).

degC = (T_data/30414)*100 + 50 +[offset for ADC channel A6:4.84 A5:3.41 A4:6.24] +-2.38

Combining offset variations across all sensors:

T_data = (degC – 50 –4.83 +-(2.38+1.4)) * 304.14 = 304*degc – 16676 +-1150 .

therefore, typically for all TEMP’s data T_data_25degC = –9076 +-1150 .

---- double checking is in progress in nspotr_BCU_sensor_data_temp.doc

---- appendix memo

?? mismatch between nspotr97616 sensrept.m & 溫循071214_BCU_Temperature_Cali¥BCU_Temperature_Cali071214.doc 需要討論/解決 詳見 nspotr_BCU_sensor_data_temp.doc

...

Revising 98107...:

---98112 still undergoing double check ...?

Ref: nspotr_BCU_sensor_data_temp.doc rev 98207, based on V1.2B sensrep982051723A4T0.txt a4t0_205.m

Calibration formula is

( data_count + (14760 +-54) ) / (225.0 +- 1.1) = temp_degC +-1.7

Calibration Curve

Note: 比較 溫循得到的公式:

? ADC數據 = 溫度_deg_C * 15.9 -1006.1 +-17.56 ;

? 似乎 約等於 data 要放大 16倍 而非 4倍 => 16*ADC數據 = 溫度_deg_C * 254 -16098 +-281

?? 有什麼合理的解釋 ? 哪個公式才對?

Note: Alternative calibration curve: {morphous\ | 060401_NSPO_台俄衛星¥}之

071214_BCU_Temperature_Cali¥BCU_Temperature_Cali071214.doc

從溫循數據直接得到的 校正線與公式 為

"data = deg C * 15.9 – 1006 +-17.6"

(似乎這個公式的 右側 需要 x16 才能近似 從nspotr97616的數據得到的公式,

而不是單純的 shift 2bit x4.)

nspotr_BCU_sensor_data.doc 10 97c29 tly c31 98112 211 703 99202 03:29@TW 20110505 0529