|
Hi all
I completed the QMX kit build and got everything up and running OK including making some initial FT8 QSOs.
I built for 12V and have been running on 12V from my regulated (but not current limited) bench PSU with a total power on time so far of about 2 hours.
So today I was exploring the menus for beacon mode, when the magic smoke suddenly escaped - QMX was not transmitting at the time
It looks like the C107/Q108 failures reported by others here
I think the best approach here will be to rig up linear 3.3/5V supplies (I have plenty of linear regulators to hand) to test the main boards for collateral damage before repairing the PSU board
Any thoughts ?
Cheers
Rick
|
Be careful with the voltages. ICs 403/404 are max voltage of 5.5 V. I blew out IC403; probably because I accidentally put about 5.5 V on it. Processor continues to function well despite all the issues I had.?
73, Dan? NM3A
|
Rick wrote above: "I think the best approach here will be to rig up linear 3.3/5V supplies " Hello Rick and group, I have a similar tought on another thread in this group. Check for the message:
There already is a 3,3V linear regulator in the rig. One should only disable the "handover" from the linear regulator to the buck converter. I seriously intend to do so, at least until I understand what's up with all this happening that makes me feel rather unsure when handling the rig in so many experimental ways.
The rig is small and frail; my hands are big and clumsy... ;-)
For a 5V rail there is no such linear regulator in the rig, so one might add it somewhere where there is enough groundplane on the PCB to solder it to. I would refrain from using one of those scarry xx1117 5V LDO's needlessly having the output electrode on PIN 2 (middle pin) and the cooling tab. In case of a thermic failure, the input gets toasted directly to the output. Whoever designed this chip... (censored)
More civil linear regulators have the cooling tab connected to the ground electrode and if "melt-down" happens, the input gets connected to the ground and there might be hope that a fuse somewhere gets blown before the trace on the PCB burns out. But in most cases, the output stays safely isolated from the ugly happening.
And do not forget to connect the 3,3V regulator to the 5V rail, to avoid unnecessary loss. By doing so, there might be problems with the wake-up timing of the whole combo. As designed, the 3,3V rail feeding the main professor is alive and kicking long before the 5V rail is up. Reversing the order of wake-ups, some latch-up might occur that was not foreseen in the initial sequence. A big electrolitic cap here and there or even some sort of wise watch-dog reset-policy might be needed if wake-up gets bitchy.?
Let's try and please keep us informed of your progress. The more I think, the better I find the idea of replacing the buck converters with linear regulators, if not for good, at least for the testing period, until things settle down. But the more I think, the more I believe that, if I would succeed in replacing the buck converters with LDOs... I would just leave it stay put this way.?
It is also very refreshing to know that the buck converter boards can be tested without endangering the main professor, the one and only, the unreplaceable.
Yours friendly, Razvan dl2arl
|
Hello Razvan
The SMPS boards are not a risky proposition, my friend. There is a known issue with a proportion of the current batch, having the Q103/Q104 drain short, regrettably. However other than this, there are no known issues with the SMPS boards or the whole concept.?
There are certain short-circuits, via solder bridges, carelessly placed probes, wire clippings and so on, which the QMX will not tolerate or forgive. However this is always the case in?any radio with multiple power rails. Well in fact in almost any circuit! You never have the right to short any arbitrary parts of the schematic together and expect the rig to not get damaged.?
On Thu, Aug 24, 2023 at 12:45?AM DL2ARL < dl2arl@...> wrote: Rick wrote above:
"I think the best approach here will be to rig up linear 3.3/5V supplies "
Hello Rick and group,
I have a similar tought on another thread in this group. Check for the message:
There already is a 3,3V linear regulator in the rig. One should only disable the "handover" from the linear regulator to the buck converter. I seriously intend to do so, at least until I understand what's up with all this happening that makes me feel rather unsure when handling the rig in so many experimental ways.
The rig is small and frail; my hands are big and clumsy... ;-)
For a 5V rail there is no such linear regulator in the rig, so one might add it somewhere where there is enough groundplane on the PCB to solder it to. I would refrain from using one of those scarry xx1117 5V LDO's needlessly having the output electrode on PIN 2 (middle pin) and the cooling tab. In case of a thermic failure, the input gets toasted directly to the output. Whoever designed this chip... (censored)
More civil linear regulators have the cooling tab connected to the ground electrode and if "melt-down" happens, the input gets connected to the ground and there might be hope that a fuse somewhere gets blown before the trace on the PCB burns out. But in most cases, the output stays safely isolated from the ugly happening.
And do not forget to connect the 3,3V regulator to the 5V rail, to avoid unnecessary loss. By doing so, there might be problems with the wake-up timing of the whole combo. As designed, the 3,3V rail feeding the main professor is alive and kicking long before the 5V rail is up. Reversing the order of wake-ups, some latch-up might occur that was not foreseen in the initial sequence. A big electrolitic cap here and there or even some sort of wise watch-dog reset-policy might be needed if wake-up gets bitchy.?
Let's try and please keep us informed of your progress. The more I think, the better I find the idea of replacing the buck converters with linear regulators, if not for good, at least for the testing period, until things settle down. But the more I think, the more I believe that, if I would succeed in replacing the buck converters with LDOs... I would just leave it stay put this way.?
It is also very refreshing to know that the buck converter boards can be tested without endangering the main professor, the one and only, the unreplaceable.
Yours friendly, Razvan dl2arl
|
Hi Razvan Thanks for your input - I meant only to rig up temporary linear supplies on my bench to check through the main-board, not as a permanent change The plan is :
- Remove both SMPS boards
- Supply +3.3 from linear regulator and check if the CPU runs
- Replace the 3.3V SMPS/linear regulator board and supply +12V from linear supply, check CPU again
- If all good, repair or replace the 5V SMPS board
Regards Rick Regards Rick
|
On Thu, Aug 24, 2023 at 01:46 AM, Rick Crowther wrote:
I meant only to rig up temporary linear supplies on my bench
hello Rick, Hello Hans, this is exactly what I was aiming for myself: a temporary replacement of the smps boards until all other smoke tests have been succesfully passed. Hans: really no offence intended! I am of an older school and do not trust these new-age gadgets piloted by software, not because of paranoia, but because of the fact that they do not belong to my trusted analog world. I find the ideas standing behind the SMPS concept of the QMX most refreshing and innovative: this is the main reason I am glad of having purchased a QMX, without having been in need of yet another qrp toy. But now.. I am sort of discouraged by my own courage and would like to go one step behind into those quiet linear waters I am used to. Not for good, but at least for the first current intake of the fine rig. Afterwards, having thanks to Chris a test checklist for the SMPS boards outside the rig, without danger of damaging the main professor, I would maybe gain some courage and "replace the linear replacement" with it's new age microprofessor controlled switching PS. This is what I meant. You might help by telling me if this is doable or if I am to expect some latch-up phenomena during power-up due to the fact that the 5V rail would come alive before the 3,3V would do. But latch-ups are not harmfull, so if you do not have an answer to this question, just forget it: I will find it out by myself. Yours friendly, Razvan dl2arl?
|
Razvan said : "... the 5V rail would come alive before the 3,3V would do. But latch-ups are not harmfull, so if you do not have an answer to this question, just forget it: I will find it."
Dear Razvan,
The latch up phenomenon that CMOS components exhibit is much more dangerous than that and is to be avoided!
It's effects are unpredictable, ranging from the harmless and resettable condition that you mention to the full? destructive meltdown of the affected component, having made a massive short circuit between supply rail and ground!
Everything depends on the regenerative gain of the parasitic SCR that is intrinsic to the CMOS structure, and of the level of the current that is? being inappropriately injected into that SCR.
Modern CMOS circuits and processes have been engineered with latch up avoidance as an objective, and they have some forgiveness. Risks remain however. The processor in QMX for example has many pins that can "eat" an injection current of up to 5 mA without latch up or other damage. Some of the pins on that processor have zero tolerance for injected current!
Just an FYI, Razvan.
Wishing you well, JZ KJ4A?
On Thu, Aug 24, 2023, 5:38 PM DL2ARL < dl2arl@...> wrote: On Thu, Aug 24, 2023 at 01:46 AM, Rick Crowther wrote:
I meant only to rig up temporary linear supplies on my bench
hello Rick, Hello Hans,
this is exactly what I was aiming for myself: a temporary replacement of the smps boards until all other smoke tests have been succesfully passed.
Hans: really no offence intended! I am of an older school and do not trust these new-age gadgets piloted by software, not because of paranoia, but because of the fact that they do not belong to my trusted analog world. I find the ideas standing behind the SMPS concept of the QMX most refreshing and innovative: this is the main reason I am glad of having purchased a QMX, without having been in need of yet another qrp toy.
But now.. I am sort of discouraged by my own courage and would like to go one step behind into those quiet linear waters I am used to. Not for good, but at least for the first current intake of the fine rig. Afterwards, having thanks to Chris a test checklist for the SMPS boards outside the rig, without danger of damaging the main professor, I would maybe gain some courage and "replace the linear replacement" with it's new age microprofessor controlled switching PS.
This is what I meant. You might help by telling me if this is doable or if I am to expect some latch-up phenomena during power-up due to the fact that the 5V rail would come alive before the 3,3V would do. But latch-ups are not harmfull, so if you do not have an answer to this question, just forget it: I will find it out by myself.
Yours friendly, Razvan dl2arl?
|
Hello Razvan
If you use linear 3.3V and 5V power supplies it will work fine, and there is no harm in the 5V coming alive before 3.3V. In order to make the processor complete the boot process you will need to arrange the 3.3V and 5V ADC measurements to both be above 1.65 and 2.5V respectively; the easiest way to do this is simply connect them to the 3.3V supply rail.?
On Fri, Aug 25, 2023 at 12:38?AM DL2ARL < dl2arl@...> wrote: On Thu, Aug 24, 2023 at 01:46 AM, Rick Crowther wrote:
I meant only to rig up temporary linear supplies on my bench
hello Rick, Hello Hans,
this is exactly what I was aiming for myself: a temporary replacement of the smps boards until all other smoke tests have been succesfully passed.
Hans: really no offence intended! I am of an older school and do not trust these new-age gadgets piloted by software, not because of paranoia, but because of the fact that they do not belong to my trusted analog world. I find the ideas standing behind the SMPS concept of the QMX most refreshing and innovative: this is the main reason I am glad of having purchased a QMX, without having been in need of yet another qrp toy.
But now.. I am sort of discouraged by my own courage and would like to go one step behind into those quiet linear waters I am used to. Not for good, but at least for the first current intake of the fine rig. Afterwards, having thanks to Chris a test checklist for the SMPS boards outside the rig, without danger of damaging the main professor, I would maybe gain some courage and "replace the linear replacement" with it's new age microprofessor controlled switching PS.
This is what I meant. You might help by telling me if this is doable or if I am to expect some latch-up phenomena during power-up due to the fact that the 5V rail would come alive before the 3,3V would do. But latch-ups are not harmfull, so if you do not have an answer to this question, just forget it: I will find it out by myself.
Yours friendly, Razvan dl2arl?
|
Hi Hans Thanks for your help so far? So I checked the mainboard by
- supplying 3V3 from a linear regulator - CPU runs (can connect putty)
- installing the 3V3 SMPS board then applying 12V to the switched +12V line - runs OK so the 3V3 SMPS is good
- supplying 5V from linear regulator - with a voltage divider to feed ADC_5V - as you can see the measured 5V is looking good but the processor still detects an error - see screenshot
Ideas how to bypass that so I can complete checks ? Regards Rick
|
I have worked on about 20 different QMX rigs for fellow hams. ?I made up a set of linear regulated ?PS boards for this kind of testing. ?3.3V from the existing linear regulator and 5V from a bench supply (fed from a hacked and burned up SMPS board) that I can control and measure the current. ?Feeding the ADC_ lines as well. ?This all works well and is very helpful in isolating problems and quickly confirming the main circuitry. ?One of my best tools.
|
OK
No need to respond to my last message : the problem was that it needs to see 5V and I only had 4.93 ... so I adjusted the voltage divider a tad
Now up and running properly, so I know the only damage is on the 5V regulator board
Cheers
Rick
|
i have just had the same problem, (like the qcx mini had) i was pulling the power supply plug from the QMX when it was turned on, (as it wouldnt turn off). and smoke came from the SMPS board. Great this just gets better
Mike
|
I would still like to see what the power sequencing with a scope of VDD and VCC looks like during power ON and Power OFF whether using the pushbutton or pulling the V_IN power lead. Just the details of the 1st second after switching, once and with a sequence of ON/OFF several times. Both Plug-In PCBs installed in a working QMC.?
73 Kees K5BCQ
|
Kees,
Here you go. ?Measured on my working 9v QMX.
On:
Turn off with encoder press:
Hard turn off (in this case turning off the bench power supply, not actually pulling the plug):
Hope this helps,
Jonathan KN6LFB
|
Jonathan,
Thank you very much, let me look at it.
73 Kees K5BCQ
|
What are the horizontal and vertical scales ?
73 Kees K5BCQ
|
1V/division on the vertical scale, with a -3V offset
50ms/division on the horizontal scale
Not quite the full second you requested but there didn't seem to be anything going on further out.
Jonathan
|
On Mon, Aug 28, 2023 at 11:52 AM, Michael Warrington wrote:
i have just had the same problem, (like the qcx mini had) i was pulling the power supply plug from the QMX when it was turned on, (as it wouldnt turn off). and smoke came from the SMPS board. Great this just gets better
Mike
This is very similar to my release of smoke : ?Please refer to
/g/QRPLabs/message/108631
This looks like a?systematic problem!
|
Hi all
If the rig won't turn off, the question is, what is causing THAT to be the problem? Something must be wrong?already with the situation, for that to be the case. Until the radio is turning on and off properly I think it is prudent to run it at a lower voltage if possible.?
I have seen quite a few cases where there were shorts from the female pin header connector pins, to nearby SMD components. I think it will be worth adding to the build instructions, a step to cut 2mm off each pin of the connector just to reduce this probability.?
I don't know if this was the problem in your case; but SOMETHING is wrong, if it won't power down on a long left-knob press.?
I do not think removing power suddenly from the QMX should cause any issues. Here I do it routinely. Very often if I am doing hardware experiments (filters, etc) or tuning up a QMX, I do so with the LCD and the controls boards removed, so that I can easily access the LPF and BPF toroids. I have a little 1x2-pin male header in which I joined the pins at the short end, with a blob of solder. I just insert this into the 2x4-pin controls header at the appropriate point to effect a press of the left rotary button, to switch on. I commonly then just pull the banana plug out of the power supply, or unplug the 2.1mm power jack from the QMX. I have done this hundreds of times and never had any issues.?
On Mon, Aug 28, 2023 at 11:52 AM, Michael Warrington wrote:
i have just had the same problem, (like the qcx mini had) i was pulling the power supply plug from the QMX when it was turned on, (as it wouldnt turn off). and smoke came from the SMPS board. Great this just gets better
Mike
This is very similar to my release of smoke :
?Please refer to
/g/QRPLabs/message/108631
This looks like a?systematic problem!
|
Hans said the sequencing was set for about 250ms. Your plots show the 3.3V Linear regulator coming on at t=0 but something happens to the 3.3V at t=80ms to t=200ms. I don't know what. Then it continues on to about t=250ms when the 3.3V SMPS is turned on?(notice a slightly higher level .....no problem, but you can't tell if the 3.3V from the Linear regulator is now turned OFF .....except for the now back biased D103? ? ?and it has more noise .....no problem).??
The 5V VCC also starts coming up at t=0 along with VDD but does not reach an operational "high" level until well into the 250ms sequence cycle. Can't tell exactly where from those plots.
I wonder if others have seen the same plots ?? Thank you again.
I would guess at t=80ms the processor is starting to initialize and run and is doing something else major at about t=200ms. Maybe you need a larger C103 to help stabilize the 3.3V Linear Regulator output until the SMPS takes over.?
73 Kees K5BCQ
|
DL2ARL