INCAA looked at it and wrote the text below.
Sundance will do an evaluation of the proposed alternative CTX10-1-R.
Part
|
Parallel open circuit
inductance [uH]
|
Parallel full load
current
[A (dc)]
|
Parallel DC resistance
[Ohm (max)]
|
Series open circuit
inductance
[uH]
|
Series full load current
[A(dc)]
|
Series DC resistance [Ohm
(max)]
|
Height
[mm]
|
CTX10-1A-R
|
10.58
|
2.5
|
0.046
|
42.30
|
1.25
|
0.183
|
4.19
|
CTX10-2-R
|
10.58
|
2.5
|
0.031
|
42.30
|
1.25
|
0.125
|
5.97
|
CTX10-1-R
|
10.00
|
1.9
|
0.044
|
40.00
|
0.95
|
0.176
|
4.19
|
CTX10-1P-R
|
9.62
|
1.8
|
0.052
|
38.48
|
0.85
|
0.227
|
4.19
|
The big difference between the CTX10-1A-R and the CTX10-2-R, CTX10-1-R or CTX10-1P-R is the used core material. De CTX10-1A-R is an OCTA-PAC PLUS, the CTX10-2-R and CTX10-1-R are an OCTA-PAC, and the CTX10-1P-R is a ECONO-PAC…
OCTA-PAC’s are designed around high frequency, low loss core material.
ECONO-PAC’s are a lower cost version of OCTAPAC’s offering high saturation flux density, Iron powder core material.
OCTA-PAC PLUS’s offer higher current ratings and higher saturation flux densities than OCTAPAC and ECONO-PAC, Amorphous metal core material.
The advantage of the ‘higher current ratings and higher saturation flux densities’ for the originally used CTX10-1A-R in a lower size device can be replaced by the higher size CTX10-2-R device which has about the same specifications. Het problem of the CTX10-2-R is however its height. The FMC specification states that the height of the part located in component envelope (see ANSI/VITA 57.1 FPGA Mezzanine Card (FMC) Standard, Figure 3) can only be 4.7mm high (see Figure 9), so the CTX10-2-R is not allowed, as it is situated on the EDA-02063 Fmc Adc 100M 14b 4cha within that area.
The next best is the CTX10-1-R which has a slightly lower inductance and a slightly lower full load current specification.
The power circuit of the EDA-02063 Fmc Adc 100M 14b 4cha, using the inductor as L1 and the LT1931ES5 (IC21) as switching regulator, can only provide 350mA. The LT1931 datasheet states: “When using coupled inductors, choose one that can handle at least 1A of current without saturating, and ensure that the inductor has a low DCR (copper-wire resistance) to minimize I2R power losses.” The datasheet even suggests the CTX10-1.
So you could ask yourself why in the original design the CTX10-1A-R was chosen and not the CTX10-1-R. This seems to have been a choice based on tests. The EDA-02063 Fmc Adc 100M 14b 4cha V1.0 used a SDQ12-4R7-R which is 4.7uH. If a 10uH equivalent should have been the better value, a SDQ12-100-R could have been a drop in replacement. As of the V2.0 version of the design the CTX10-1A-R is used and not the CTX10-1-R. There are no design review documents available on the ohwr site regarding this design change to V2.0, so it is unknown why this part is changed.
A change from the CTX10-1A-R to the CTX10-1-R for the V5.0 version could have some potential influence on the characteristics of the LT1931ES5 circuit, which could be shown in the form of a difference noise floor or a different amplitude of spurious frequencies in the measured signals. This can only be checked by an investigation by means of measurements.