|
Hey all! Got my QMX+ rev 3 last week and have had a blast assembling it over the past few days. Finished it last night, booted just fine, and was able to tune some CW on 10m with a big grin on my face ?
?
The thing of it is - after a few minutes, it froze. The LCD contrast went a bit darker, audio output stopped, and all front panel controls had no effect. Unable to power off by holding the VOL encoder, I had to cut power at the source. A short time later I was able to boot back up, browse some bands, tune some CW, and then it freezes again. Same symptoms: LCD goes higher contrast, audio stops, all controls stop, have to cut power.
?
I'm pretty new to debugging PCBs but figured I'd dive in as best I could to gather some info, and here's what I've been able to piece together:
- I wired it for 12V, and based on measurements my battery was delivering consistent 12.13V
- Contemporaneous with each instance of the board freezing was a drop from 12.13V down to 12.0V
- I checked a number of the components around the PWR_ON and PWR_HOLD circuits, and no blown components
- However, the measured voltage across R105 went from ~3V (normal operation) to ~4V (frozen)
- No smells or popping sounds like components faulting/failing
- No changes to the LCD indicating on-board protections or anything like that
- Tried running 9V in case there were unobserved fluctuations in my power source, but it was the same behavior - a few minutes of normal operation, and then freezing
- measured voltage during operation was around 7-8V
I've tried connecting to the serial terminal to check the on-board diagnostics, and that seems to be the most consistent way to trigger a freeze. Without trying to connect a terminal, I can usually get about 2-5 minutes of normal operation between power cycles, but as soon as I attempt to connect the terminal, it freezes.
?
Like I said, I'm new to a lot of this but eager to learn so forgive me if there's obvious debugging steps I've missed ? ??. But yeah, not quite sure where to go from here! Would greatly appreciate any help or pointers!
|
Hi Michael,
?
Please describe the power source in detail.?
Including DC plug.?
?
73,
George?
K3GK
|
Hi George,
?
The 12V power source is a 6000mAh Talentcell rechargeable battery. To connect power to the radio, I'm using the included DC plug that I got as part of my kit, which I've joined to one of the barrel connector wires that came with the battery. So the battery's barrel connector connects to the battery, and the QMX+'s barrel connector connects to the QMX+.
?
For the 9V, I just used an off-the-shelf 9V battery with a 9V battery-clip-to-2.1mm-x-5.5mm barrel plug.
?
73,
Michael
N7WTM
|
I have a rev.3 I didn't burn in yet. So let me see if it is working HiHi. You made me think ??
|
Do you have enough time to transmit into a dummy load on 20?? Does it kick out, immediately??
|
Also is there time to do a reset before the pro analysts see this post? That will give you a headstart! I did notice a little quirky behavior on mine, right after I first powered it up.?
|
Just tried it - the board gave me enough uptime that I could send a few dahs. It looks like I'm barely putting out any power, but I did see the FWD meter on my ATU bump a little bit. I didn't wait long enough for the radio to freeze out though, so I was at least able to give it a gentle shutdown this time ?
|
Will try doing a reset next.
|
Alright, did the reset, correctly selected the QMX+ version from initial selection, and had about 2 minutes of solid listening time to a QSO on 20m. Thought maybe the reset was going to fix it until I went into the menu to change my keyer setting and it froze on me ?
|
Also replug the front panel and look at the display connection and while it's out check for a tiny bridge or a whisker that may be in there. That will answer a lot of questions before real assistance comes!
|
On Tue, Apr 1, 2025 at 07:54 PM, Michael Shiplet wrote:
However, the measured voltage across R105 went from ~3V (normal operation) to ~4V (frozen)
Michael, where did you connect the GND probe of your Meter to the QMX GND?
3 V ... 3.3 V would be ok. This signal comes direct from a controller I/O-pin and the controller runs from VDD 3.3 V. 4 V is wrong. Maybe VDD is going wrong after a few minutes. This would be dangerous for some IC especially the controller. Could you measure VDD. You may use JP105 to contact VDD. But TAKE CARE not to connect VDD to "+12V". This would destroy some important elements.
Could you measure the supply current from the beginning until the QMX stops?
Could you check temperature of IC101? But please take care it could be hot.
?
I would avoid to transmit in the moment.
?
73 Ludwig
|
Also do you have the internal GPS? if so unplug it.
Also is it in a case? If so remove it.?
|
Ok real help has arrived! CUL.
|
On Tue, Apr 1, 2025 at 03:43 PM, @Ludwig_DH8WN wrote:
Michael, where did you connect the GND probe of your Meter to the QMX GND?
Could you measure VDD. You may use JP105 to contact VDD.
Could you measure the supply current from the beginning until the QMX stops?
Could you check temperature of IC101? But please take care it could be hot.
Thanks Ludwig! To answer the above questions:
- When connecting to GND I used either a GND port on one of the JP nodes, or the bar on the edge of the board
- Measuring through VDD showed the same results as when I measured R105 (which I had previously measured directly, one probe on either side of the resistor)
- during normal/healthy operation, it showed ~3.3V, and during the frozen state it showed ~4.5V
- I don't have an inline ammeter to measure supply current yet, but I have one arriving either today or tomorrow and can provide those readings at that point
- Temperature of IC101 was cool enough that I could touch it with my fingertip and and it felt about the same as room temperature, so about 17°C / 63°F
73, Michael
|
Did you use The Talentcell for 9 V and 12 volts??
Possibly flaky BMS??
|
On Tue, Apr 1, 2025 at 05:52 PM, George Korper wrote:
Did you use The Talentcell for 9 V and 12 volts??
Possibly flaky BMS?
No, the 9V was a standard Duracell like you'd put in a smoke detector. I wasn't expecting much, just to see if the lower voltage helped or resulted in the same. Ultimately I think it was about the same.
?
Also Ludwig! I've got my ammeter measurements now as well:
?
- battery plugged in, radio turned off: 0.00A, ~12.08V
- radio turned on, normal operation: 0.06-0.10A, ~12.08V
- radio turned on, frozen: 0.18-0.20A
- there's an initial drop in voltage to around ~11.9V, but then it seems to stabilize back at ~12.04V
?
So it looks like the supply current effectively doubles when the radio's in the frozen state, in addition to the voltage jump on VDD.
|
A measurement of Vcc would also be useful, and how it changes.
Maybe also touch D109 - I expect it will be getting hot as Vcc goes up above 3.3V.
I'm wondering if you have a temperature-dependent issue that causes the 5V supply to get shorted to Vdd, pulling Vdd up and Vcc down.
Maybe also take a resistance measurement (with power off) between Vdd and Vcc.
?
When you said you could reproduce the problem by trying to attach the terminal, did it happen when you connected the cable, or when you hit enter in the terminal window to bring up the menu?
|
Michael, please don't switch on your QMX+ in the moment. I will send more information later today.
73 Ludwig
|
On Wed, Apr 2, 2025 at 01:24 AM, @Ludwig_DH8WN wrote:
Michael, please don't switch on your QMX+ in the moment. I will send more information later today.
73 Ludwig
RR! Will hold off on any more tests/debugging until I hear from you ?
|
On Wed, Apr 2, 2025 at 05:32 AM, Michael Shiplet wrote:
- battery plugged in, radio turned off: 0.00A, ~12.08V
- radio turned on, normal operation: 0.06-0.10A, ~12.08V
Hello Michael, that's fine.
- radio turned on, frozen: 0.18-0.20A
A bit much but no heavy short or so.
?
This case is interesting and the situation is cloudy. Hard to debug from afar. And debugging maybe dangerous for the QMX. There are two reason:
- VDD = 4.5 V is higher than the rating and heavy stress for most of the IC connected to VDD. The controller, the heart of the QMX, is connected to VDD.
- VDD = 4.5 V and a higher current looks like the zener diode D109 is in action. With regard to the characteristic the diode current could be 120 ... 160 mA and the power 540 ... 720 mW. This diode is rated 500 mW and may heat up to much from to much power. If this diode will break the resistance could go to a low value or a high value. For first case VDD would be lower and no problem for the controller and for the second case vice versa bringing more danger for the STM32 and others.
So staying (for a while) in the failure status is dangerous but how to debug not being in this status?
?
And I see one more problem. I don't know is the controller stopping because VDD is to high or is VDD raising because the controller stops? Without a proper running controller no proper running SMPS for VDD and VCC.
?
Could you check please R110 and R112 (good soldered and not damaged). Via this voltage divider a feed back signal from VDD is produced to control VDD.
?
There are some good points in posts before:
?
George wrote:
Also replug the front panel and look at the display connection and while it's out check for a tiny bridge or a whisker that may be in there. That will answer a lot of questions?
And please make a careful visual inspection using a strong magnifier. Look for solder bridges, blobs and whiskers.
?
George wrote:
Also do you have the internal GPS? if so unplug it.
Also is it in a case? If so remove it.?
And when you test for success and you reach the faulty state again please remove the supply immediately.
?
Stan wrote:
I'm wondering if you have a temperature-dependent issue that causes the 5V supply to get shorted to Vdd, pulling Vdd up and Vcc down.
Maybe also take a resistance measurement (with power off) between Vdd and Vcc.
Stan wrote:
When you said you could reproduce the problem by trying to attach the terminal, did it happen when you connected the cable, or when you hit enter in the terminal window to bring up the menu?
I'm curious about your answers.
?
73 Ludwig
?
?
|
Michael Shiplet