Skip to main content ↑ ↓
Topic: Logic Sniffer v1.03 manufacturing update (Read 7217 times) previous topic - next topic

Logic Sniffer v1.03 manufacturing update

PCBs are attached for a minor manufacturing update of the OLS. It includes these updates:

*SSOP PIC
*ADC routed to 1v2 supply for expanded self test
*4mm tact switches can be P&P for faster assembly
*Added one additional supply cap (C28) to the buffer (IC4)
*Remade the buffer (IC4) footprint according to the datasheet (there are problems with messy buffer chip soldering)

Maybe it would be good to add footprints on the bottom of the PCB for more filter caps, even if we can't place them on the top of the PCB.

Any thoughts or requests for a minor revision update? These will probably go out for prototyping soon.
Got a question? Please ask in the forum for the fastest answers.

Re: Logic Sniffer v1.03 manufacturing update

Reply #1 –
The silkscreen has arm, trig, reset and update upside down compared to the other textes on the silkscreen.

About the decoupling capicitors I would say the more the better ;)

Keep up the good work!

Re: Logic Sniffer v1.03 manufacturing update

Reply #2 –
[quote author="ian"]Any thoughts or requests for a minor revision update? These will probably go out for prototyping soon.[/quote]
What about connecting the shield of the USB connector, and maybe adding some inductors inline for VUSB and GND to reduce noise? I'm not completely familiar with all of the failure modes for USB in various situations, but I have seen some fairly complex circuits. I'm thinking in lines of the circuits that you see on the Microchip USB Evaluation boards, which have lots more filtering. If you're interested, I can share schematics.

Re: Logic Sniffer v1.03 manufacturing update

Reply #3 –
Thanks guys, great feedback. I'll try to:
*Add more caps on the top, or bottom (worst case)
[s:]*Add a bead to the power supply[/s:]
*Shottky diode (i.e. MBR0520 like Microchip) for polarity protection and current limitation between VUSB(USB/UART) and the MIC VRs.
[s:]*RC4/D- has a single right-angle turn in it which should probably be a couple of 45-degree turns like all the rest. Also, you might consider increasing the width of the D+,D- traces to be almost as wide as the PIC pads.[/s:]


There isn't much room on this revision for a more complex power supply filter, but it should be something to put on the next version for sure.
Got a question? Please ask in the forum for the fastest answers.

Re: Logic Sniffer v1.03 manufacturing update

Reply #4 –
Nice polishing job ...

I agree with Sjaak, it would be nice to have the text on the buttons, the ROM-ISP header and the ACT and PWR LED oriented the same way as all the other text. The small text (on the edge) next to the ACT and PWR LED is not required as the LEDs are clearly marked on the inside.

Like rsdio suggested, connecting the 4 standoffs/pads (shield) of the USB connector to GND would be be a good idea. In general, I'd follow the schematics of the Microchip "MPLAB Starter Kit for PIC 18F" for the USB interface - and insert a Shottky diode (i.e. MBR0520 like Microchip) for polarity protection and current limitation between VUSB(USB/UART) and the MIC VRs.

For production: I suggest to populate the TRGI/CKI/CKO/TRGO header in the next production run ... far more important than populated UART and ROM-ISP headers, I think. ;)

The idea to add pads for filter caps on the bottom sounds good but space should be kept for "sticky rubber" feet on the corners - with caps on the bottom you really want to keep some distance between the table/work bench and the board - or maybe make the board a bit wider and add some mounting holes ... or just extend the edges (long sides) sligthly so the board could be mounted in a housing/case similar to ril3y's Synthetos Bus Pirate V3 Plexi Glas Case ...

Slightly modify the UART header so the signals/footprint will match FTDIs USB to UART cable (3.3V), Sparkfuns FTDI Basic Breakout Board - 3.3V and compatibles (6-pin .100 header, GND, CTS, VCC, TXD, RXD, RTS - with CTS and RTS unconnected).

Re: Logic Sniffer v1.03 manufacturing update

Reply #5 –
Quote
match FTDIs USB to UART cable (3.3V), Sparkfuns FTDI Basic Breakout Board - 3.3V and compatibles (6-pin .100 header, GND, CTS, VCC, TXD, RXD, RTS - with CTS and RTS unconnected).

The OLS UART header has a supply input pin that powers the board. Does the adapter VCC supply power, or draw it for the FTDI chip's output pins? A separate supply would be needed (USB cable?) if the adapter doesn't provide power.

If the adapter draws power from the VCC pin for native-level interfacing, then it would use the 5volt USB supply under the current configuration. That's OK because the PIC pins connected to the UART are 5volt tolerant.
Got a question? Please ask in the forum for the fastest answers.

Re: Logic Sniffer v1.03 manufacturing update

Reply #6 –
Since there's some interest in connecting the USB shield, I think I should explain more clearly what I've learned.

You don't actually want to connect the shield to ground, because that conducts radio and other EMI noise into the power circuit and the rest of the board.  Instead, what you really want to do is build a small RC filter, and connect the shield node to the ground node through that RC filter.  This keeps the ground quiet (no noise) but the shield is still loosely connected.  I think there might even be a note in the USB specifications that you are not supposed to connect shield to ground at the USB device. The most recent circuit that I "borrowed" had a 1MΩ resistor in parallel with a 0.1 uF cap between the two.

If you don't have room for the RC filter, leave the shield isolated, but still connect all four solder tabs together. What I do in either case is draw a polygon in Eagle under the USB connector, on the bottom copper layer, which extends the shield all the way around the data lines, basically continuing the USB cable shield onto the PCB. This is easier with the through-hole USB-B connectors, but you'd have to do it here with one to four vias. In this specific case, it might force you to move some traces on the bottom layer to make room for the shield polygon.

Re: Logic Sniffer v1.03 manufacturing update

Reply #7 –
The USB UART cable/adapters (FTDI, Sparkfun) provide power on the UART side, they draw power from the USB port they are connected to.

Quote
+5V or +3.3V output (only 5V on the FTDI cables) allows external logic to be powered from the USB port.

TTL-232R USB - TTL Level Serial Converter (5V)
TTL-232R-3V3 USB - TTL Level Serial Converter
Sparkfun Basic Breakout 5V
Sparkfun Basic Breakout 3.3V

However, the spec states a limit of max. 75mA/5V for the device connected on the TTL-UART side.
For the FTDI cables Vcc is 5V (4.25 .. 5.25V) for both the 5V and the 3.3V version, as they draw Vcc directly from the USB host port.

The main difference between the FTDI cables and the Sparkfun Basic breakout board is that the FTDI cables bring out RTS on pin 6
while Sparkfun has DTR on pin 6. This makes the Sparkfun adapter Arduino compatible (reset via DTR). On the OLS there is no
need for either pins so pin 6 can be left away. If space allows pin 6 should be mounted for mechanical reasons and to prevent
confusion.

The other difference is that Vcc is 3.3V for Sparkfuns 3.3V breakout board, provided by the on-chip voltage regulator of the FT232R
while on the Sparkfun 5V breakout board Vcc is 5V (4.25V ... 5.25V), drawn from the USB port like on the FTDI cables.

@rsdio, my bad ... very bad ... shielding of signal lines should be "grounded" on one side only, where applicable on the powered side BUT should never be connected (directly) to GND of the logic circuitry ... funny thing is the Microchip MPLAB Starter Kit for PIC18F schematics shows the 4 shielding pads of the USB mini connector for the application MCU (PIC18F46J50 - big 44-pin brother of our PIC18F24J50) directly connected to GND (Figure A-2). ;)

Quote
6.8 USB Grounding
The shield must be terminated to the connector plug for completed assemblies. The shield and chassis are
bonded together. The user selected grounding scheme for USB devices, and cables must be consistent with
accepted industry practices and regulatory agency standards for safety and EMI/ESD/RFI.

Re: Logic Sniffer v1.03 manufacturing update

Reply #8 –
Forgot one point - I consider rather important:

The probe / I/O headers should be clearly marked - i.e. IO1 and IO2 or IOB (buffered) and IOW (wing header - unbuffered) for easier and clearer reference in documentations!

After looking over the silkscreen of the v1.01 board I finally get the idea why you put some labels upside down, some even in both directions - to make them easily readable from both sides ... still I think the small upside down labels close to the edges are not required and uniform labeling would make the design look "cleaner" ... more appealing ... artistic freedom, I guess :)

Re: Logic Sniffer v1.03 manufacturing update

Reply #9 –
[quote author="IPenguin"]
@rsdio, my bad ... very bad ... shielding of signal lines should be "grounded" on one side only, where applicable on the powered side BUT should never be connected (directly) to GND of the logic circuitry ... funny thing is the Microchip MPLAB Starter Kit for PIC18F schematics shows the 4 shielding pads of the USB mini connector for the application MCU (PIC18F46J50 - big 44-pin brother of our PIC18F24J50) directly connected to GND (Figure A-2). ;)

Quote
6.8 USB Grounding
The shield must be terminated to the connector plug for completed assemblies. The shield and chassis are
bonded together. The user selected grounding scheme for USB devices, and cables must be consistent with
accepted industry practices and regulatory agency standards for safety and EMI/ESD/RFI.
[/quote]
No worries.

This issue (USB shield) is one where I have found it very hard to find clear recommendations. In fact, almost everything about the USB connector is difficult to be certain about. Power and ground need filtering and static discharge protection. Differential data needs controlled impedance - i.e. traces should stay on the same side of the board, follow each other at a constant distance, and not make any right-angle turns for best results. (*) And, of course, the shield needs to be dealt with in a way that doesn't cause problems.

I've seen example circuits from chip makers which clearly violate the rules. It's taken a long time to learn enough from the USB specs to know when to ignore a part of an example circuit which is clearly wrong, versus knowing when to use the ideas in some other example circuit.

(*) On that note, the differential data traces look great on the v1rev3 PCB except for one thing: RC4/D- has a single right-angle turn in it which should probably be a couple of 45-degree turns like all the rest. Also, you might consider increasing the width of the D+,D- traces to be almost as wide as the PIC pads. It's actually difficult to know the impedance of the traces without coordination with the PCB fab house, but generally "wider" is better. My general rule is to go about as wide as the SMD pads and not worry about going beyond that.

Re: Logic Sniffer v1.03 manufacturing update

Reply #10 –
How about adding pull-up/down resistors to the Mx configuration pins with "solderable bridges" to pull the Mx pins to the other side ?  This would allow us to load the bit-stream through USB. Normal manufacturing would not need the bridges to be soldered and would not change manufacturing costs at all :)

Related thread: http://dangerousprototypes.com/forum/in ... opic=594.0

Logic Sniffer v1.03 - MOVED: right-angel bends and propagation delay discussion

Reply #11 –
alm, rsdio & all others, I moved all posts in this thread that pertained to the effect of right-angel bends in microstrips and to propagation delay in microstrips to Project development, ideas, and suggestions.

http://dangerousprototypes.com/forum/index.php?topic=614.0

This is a far more general subject that does not pertain to the OLS project in specific. In particular the question of the influence of right-angel bends in the USB traces on signals in the OLS design was found to be insignificant lateron in the discussion.

I think the discussion is very interesting so it deserves it's own thread and at the same time it keeps this thread on subject - Logic Sniffer v1.03 manufacturing update (!)