#u3s #synth 27MHz Crystal on Si5351 module failure? U3S just showing boxes... #u3s #synth


Al Holt
 

I did a little work on the Si5351A synth board, removed and reseated the 27MHz crystal and now I'm getting only the top line of boxes on the U3S when powering up. If I remove the synth module the U3S powers up OK and the display informs me the synth module is absent.

I checked the board for other obvious signs of trouble, it's getting +5 Volts and the regulator is making 3.3V. I haven't tried checking the module away from the U3S on the bench powered separately for clocking signals, yet. Any other troubleshooting ideas I might try?

I'm wondering how susceptible the crystal is to heat damage? 

Thanks!

--Al
WD4AH


Daniel Ekman SA2KNG
 

The 27MHz crystal has nothing to do with the mcu, so the boxes tells you that the lcd is not initialized. It requires a certain sequence from the mcu to start up and clear.
One pin is shared between the lcd and the synth, and if it's shorted to either rail it will certainly mess up the lcd init.
Synth pin 14, mcu pin 4, lcd pin 13. Look for problems on the synth as you say this is the one worked on last, ohm with a dmm perhaps.
Not sure what symptoms to expect if synth module is put in backwards...


Al Holt
 

Thanks Daniel, I'll look into it.
Are you saying if the 27MHz crystal is not oscillating it will not prevent the synth module from keeping the U3S from initializing? (Is that a double negative? :-) )

I understand the crystal on the synth module doesn't have anything to do with the MCU, but the synth module is needed to get the mcu to start up.

I should have asked this in my OP, what should I expect from the synth module if I test it on the bench and not connected to the U3S? Does it oscillate or free run at a certain frequency? Does it need some sort of I^2c commands to get it to work?

--Al


Daniel Ekman SA2KNG
 

If I'm not mistaken, the lcd display is initialized before the mcu even speaks to the synth over i2c. So the two rows should disappear (and probably display text) before the i2c synth stuff begins. So the 27MHz crystal is irrelevant during this phase. The mcu has it's own 20MHz crystal to run, so it does not need anything from the synth, as is clear when you power up the board without the synth.
Search on the synth board for the problem first. I suspect there's something wrong with the level shifters (Q1,Q2) or the i2c pins (4,5) on the Si-chip (measure at R3,R4). They should both idle at 3.3V at the chip side and 5V at the pinheader to the QCU board.
I'm not so familiar with the Si-chip as to know what it does when it's powered up without a controller talking to it, perhaps in the datasheet. Still this does not help too much if the problem lies in the i2c pins and level shifters.


David
 

If the synth module is not even plugged in you still get text -- "QRP Labs U3S ver 3.12 etc....."until an error message is shown.


On 03/04/2018 21:45, Daniel Ekman SA2KNG wrote:
If I'm not mistaken, the lcd display is initialized before the mcu even speaks to the synth over i2c. So the two rows should disappear (and probably display text) before the i2c synth stuff begins. So the 27MHz crystal is irrelevant during this phase. The mcu has it's own 20MHz crystal to run, so it does not need anything from the synth, as is clear when you power up the board without the synth.
Search on the synth board for the problem first. I suspect there's something wrong with the level shifters (Q1,Q2) or the i2c pins (4,5) on the Si-chip (measure at R3,R4). They should both idle at 3.3V at the chip side and 5V at the pinheader to the QCU board.
I'm not so familiar with the Si-chip as to know what it does when it's powered up without a controller talking to it, perhaps in the datasheet. Still this does not help too much if the problem lies in the i2c pins and level shifters.


Al Holt
 

I've given the board a close review and there's not much I can tell with the board out of circuit. My oscilloscope (Tek T922) does show high frequency signal on the crystal pins so I'll assume it's oscillating. 

There's no clock activity on any of those pins (CLK0-1-2) that I can see and the SDA and SCL pins are 'high' at ~3.3v. Although this might be what you supposed to get when no connected to a I2C bus.

I did pull one of the BS170's and it tests good on my low-cost (but very, very handy) Chinese "Component Tester."

I'll probably pull the other BS170 and test it, try a couple of fresh ones. I'd better not overlook the obvious issues of cold solder joints and do some continuity checks.

Otherwise, it looks like the board is FUBAR and time to order another as well as the OCXO version that I've been meaning to try.

Thanks for the help and suggestions!

--Al


holden.dave@...
 

Al,
Sorry. Missed your earlier post which said you were seeing the Flash Screen on bootup without the synth.
David


Hans Summers
 

All correct, Daniel

FYI the standard Si5351A-B-GTR does nothing on power-up, no clock outputs. It requires programming via I2C to make it do anything.

Custom versions are possible, which have default power-up configurations programmed in. I imagine you have to make a huge order to interest SiLabs in making a custom version! 

73 Hans G0UPL 

On Tue, Apr 3, 2018, 23:45 Daniel Ekman SA2KNG <knegge@...> wrote:
If I'm not mistaken, the lcd display is initialized before the mcu even speaks to the synth over i2c. So the two rows should disappear (and probably display text) before the i2c synth stuff begins. So the 27MHz crystal is irrelevant during this phase. The mcu has it's own 20MHz crystal to run, so it does not need anything from the synth, as is clear when you power up the board without the synth.
Search on the synth board for the problem first. I suspect there's something wrong with the level shifters (Q1,Q2) or the i2c pins (4,5) on the Si-chip (measure at R3,R4). They should both idle at 3.3V at the chip side and 5V at the pinheader to the QCU board.
I'm not so familiar with the Si-chip as to know what it does when it's powered up without a controller talking to it, perhaps in the datasheet. Still this does not help too much if the problem lies in the i2c pins and level shifters.


Al Holt
 

Thanks Hans for clearing up the module's behavior when 'out of circuit.' As I alluded to in the other thread I mistakenly created, the module is currently working and my U3S has been running overnight transmitting WSPR and QRSS on 40 meters. The repair probably boiled down to inspecting it with a 'fine-toothed comb' as well as reheating connections and connectors. Probing the Si5351A chip is a challenge and I found a small sewing needle to be a great help.

I used a 'proto' type of breadboard to mount and power the Si5351A module to make it easier to troubleshoot away from the U3S. Not that I could get a lot of testing done with it, but it's got me thinking about doing some experimenting with the I2C bus structure. Maybe an Arduino sketch to simulate some commands...or just build some jumper header extenders.

But, away from the 'mother ship' the little module stood up pretty well; "took a licking and kept on ticking" as John Cameron Swayze used to say about Timex watches.

Thanks again!

--Al