/dev/null

[sphinx]

/dev/null

[phil from seattle]

I like your approach.  Using the chip (rather than a Pico 2) makes a lot of sense. Especially for laying out the length tuned lines - you get a lot more space to work with. I very much like the idea of a single board - connectors can be problematic.

You know, even a 16 bit LA based on this would be pretty sweet. 40+ seems a bit like overkill. I have a couple of 32 channel LAs and have never used all 32. Although, easier to DNP for a smaller version than to maintain two different layouts. And, the cost difference between the 235xA and 235xB is pretty minimal. I'd rather have the extra lines for signal/pattern outputs though don't recall if his sw supports that.

Gusman is a bit brittle about his baby. I asked about sigrok support, big mistake.

Do you have a github or similar repo?

[phil from seattle]

I did not realize you were looking for input on the schematic/pcb.  I have done several boards based on the RP2350B and RP2040.  I don't see anything obviously bad.

If you are referring to the +5V switching MOSFET, unneeded if you are only powering from USB.

I did not use the onchip 1.1V regulator but rather a TI LDO.  Main reason -  There was a lot of chatter that putting a buck converter on IC was a bad idea.  A polarized inductor is needed and only abracom was qualified at the time (9 month ago) - the TI LDO was cheaper than the abracom part.

Crystal selection is straightforward.  Either use the specified part or select one with similar specs at about 1/4 the cost.

You may not need a separate plane for the 1.1V.  I did my designs on a 2 layer board. You might get by with a 4L board with 1.1, 3.3 and 5 on the same plane. 1.1 will only be near the mpu and 5V could be pushed around outer edge of the board. You could keep a solid 3.3V or gnd plane under the signal ltraces. The screenshot below shows  the 1.1V (Dvdd) ring around the ground slug with the 3.3V (IOvdd) ring outside that. The Pico and Pico 2 are both 2L boards, btw.

You might want to consider an adjustable Dvdd LDO if you want to overclock the RP2354.  I have seen a lot of reports of success at much higher clock rates.

I think you have a typo, TXU0401 rather than TXU0104. And, BTW, converting from 5V to 3.3V, the max data rate is 125 MHz, not 200.

1206 IS pretty big.  I use 0402 but have my boards assembled on a PnP production line.  For manual placement, 805 are very easy. The RP2354, A or B, is going to be a bigger challenge to manually place. I often do the first turn of a board with 805s for hand rework and then reduce to 603 or 402 for full production.

[Renate]

Why not some space between the two connectors if someone wants to use shrouded headers?
Yeah, I know it will mess up the lengths.

[Renate]

Well, I don't get it.
Were you planning on using ribbon cable to break this out?
With one plug or two?
If you're using ribbon cable wouldn't it make more sense to have it go in order by conductor? (Ping pong from left to right).
Or you're going to do all that mapping in software?

A single 54 to 60 pin connector is getting a bit wide.
Why not two (or even 3) connectors with similar pinouts?