Improvements after design and layout review
General
- Have you tried to see how a typical filter that is used already
would fit? This will show immediately how practical the board is, or
what is missing.
- They are relatively small and would fit even on smaller space,
but one needs space for handling and wiring.
- It seems that the input and output connectors are really distant,
while for HF everything should be as short as possible.
- It's less a HF design than more a low noise design. The
frequencies we will be dealing with will be rather low (the
sinusoid will be <100MHz) where wiring is not that critical.
But inductances and stability (mechanical and electrical) are
important.
- Could a standard board have been “good enough” for the prototyping
of a few filters? Often HF specialists prefer just a copper plate to
assemble things.
- Yes, would be possible, but the build-up would not be as
mechanically stable and might confund the results.
- Could it have been better technically to have made a dedicated board
optimised for say 3 or 4 typical filters that you would plan to use
in this project?
- That's actually a valid approach. But using this kind of board
will allow us to be faster.
- I thought the filters to use were passive and do not need a power
supply. Usually when OPAMP are used, positive and negative supplies
are needed.
- The filter (or rather resonant circuit) is passive, yes, but
the pulse former needs power. This circuits needs power between
GND and VCC.
- As it is a generic and independent board, it could get its own OHR
project.
Schematics
- Add one LED for the power supply
- Unfortunately, as the power supplies may change, it would be
difficult to put one LED which lights up when the power is set
without adding its own regulator.
- Why not using the PG pin ?
- The PG pin is an open collector output. As the power supply can
vary from 5V to 12V, it is difficult to power the LED without
adding one regulator.
- With the two instructions “Do not solder” R1-R3, the board is not
going to work. Remove instruction (the values are OK, aren’t they?)
so that you do not have to update the schematics later.
- On this board the jack needs 5V, while on the ADC board it should be
12V. Risk of interchanging and blowing up this board.
- Actually the maximum power supported by the regulators is
around 20V.
- Why is the input supply not available to the breadboard area? Add at
least some holes so that a wire can be pulled to that area.
- Added one power supply line.
- Replace single 10uF by two 4.7uF to reduce BOM
- Add URL of ohwr project on schematics (like is already on PCB)
- Mark that connectors are SMA
- “Matrice of pads” -> Matrix
- As the 5V is going into the regulators (and not used directly), the
input voltage range is actually much higher (maybe 12V is OK?). Mark
the full range on schematics and PCB.
- Why has this board not the same power connectors as the main board
(DIN 4-pin)?
- The power supply of this board is not protected for reversal of the
input voltage. The main ADC board is. Why the difference?
- Why decoupling of 1V8 and 3V3 different (either 4u7 or 100nF, never
both). Put both C’s in parallel on both.
- One of the regulators needs to drive higher capacitance than
the other. Actually, the higher this capacitance, the greater
the stability. We changed the previous 100nF by multiples
4,7uF.
- Can you use an LM1117 also for P1V8? Saves one BOM type to order.
- The LM1117 would only be used if we need another power supply.
This regulator is not especially low noise. Placing two LT3042
would be too expensive.
- Put the ferrite beads on the input side of the LDOs infront of the
blocking capacitors.
- Why not putting termination resistors on the board ?
- As the board has to be as generic as possible, we do not put
the termination directly.
PCB Layout
- Make that also DIN 4-pin (or 2-pin) ‘domino’ can be inserted.
- Make clear on silkscreen where plus and minus are of CON1.
- There seems to be a special system with the ‘arrow-like’ pads. Where
is it documented on how to use? Could it be on the silkscreen?
- Three strips and areas in prototype area called LT.., GND and
LM1117. Shouldn’t they be called P1V8, GND and P3V3?
- Actually no. As these voltage values will change, we cannot put
them directly on the silkscreen. We changed them by SV
(Secondary Voltage) and LNV (Low Noise Voltage) to be as generic
as possible.
- If decide to have a dedicated OHR project, update the URL.
- Place the resistors/capacitors around the LM1117 infront of it and
not to the side. This makes later modifications easier.
- Put the formula on the silkscreen
- Change the label for the matrix second "A" => "AA"
- Change the label for the power supply names
BOM
- Typo in filename shown in BOM file (FilterPtro…)