After I build my QMX and thoroughly tested and inspected al the soldering points, I applied 12V to the QMX. As instructed starting at 7V and current limited to 250mA I pressed the left rotary switch. The supply voltage dropped to abt. 3,5V and the current was 250mA, falling in the limit.

And that was it.

Checking the mainboard afterwards, with both power supply boards removed, I found a short on the 3,3V line and 0 Ohms on PMV 5V (pin 26, PA3) at the MCU.

A couple of days later a HAM-friend of mine (a pro on electronics), Peter PD5PD, tested also the mainboard, with removed power supply boards. After some testing he applied abt. 2V over C433 and a Flir IR camera revealed an overheating MCU.

1)????? Is this a DOA, or did I do something wrong?

2)????? With the help of my friend a replacement of the MCU is an option. But I do read contradictory messages in this group.io on replacing that chip. It is clear to me that the MCU is programmed with the bootloader. Are in the foreseeable future programmed MCU’s available? Some messages say No, some other messages say Yes.

3)????? Do I need in the end to buy a new QMX?

In addition: another HAM-friend, Frans PE0F, build a QMX and experienced exact the same problem.

Did more people had exact this, and what was your solution.

?

73,

Martin, PA0MJM

QRZ.com/db/PA0MJM

开云体育

Unfortunately, this seems to come up frequently. At this time, the MCU is not user replaceable. I don't know that anyone has isolated the root cause of the MCU failures-- Certainly a bad SMPS board could do it (over-voltage on the MCU), but aside from a build or board error, it's hard to imagine a credible scenario.

What would be interesting is, on a board with a shorted MPU, to remove the MPU and then look at what's supplied to the MPU to determine what caused it to fail... (Presumably, what ever caused the MPU to die is still present!)

Comes up often enough that I think it would be naive to believe that all of these failures are end-builder problems...

73, Paul -- AI7JR

After I build my QMX and thoroughly tested and inspected al the soldering points, I applied 12V to the QMX. As instructed starting at 7V and current limited to 250mA I pressed the left rotary switch. The supply voltage dropped to abt. 3,5V and the current was 250mA, falling in the limit.

And that was it.

Checking the mainboard afterwards, with both power supply boards removed, I found a short on the 3,3V line and 0 Ohms on PMV 5V (pin 26, PA3) at the MCU.

A couple of days later a HAM-friend of mine (a pro on electronics), Peter PD5PD, tested also the mainboard, with removed power supply boards. After some testing he applied abt. 2V over C433 and a Flir IR camera revealed an overheating MCU.

1)????? Is this a DOA, or did I do something wrong?

2)????? With the help of my friend a replacement of the MCU is an option. But I do read contradictory messages in this group.io on replacing that chip. It is clear to me that the MCU is programmed with the bootloader. Are in the foreseeable future programmed MCU’s available? Some messages say No, some other messages say Yes.

3)????? Do I need in the end to buy a new QMX?

In addition: another HAM-friend, Frans PE0F, build a QMX and experienced exact the same problem.

Did more people had exact this, and what was your solution.

?

73,

Martin, PA0MJM

QRZ.com/db/PA0MJM

-- 
Paul -- AI7JR

Hello Martin,?

I have the same problem,? see /g/QRPLabs/message/116603

It seems that there is no ESD protection of the CPU at all, so that you have a good chance to kill it.?
There was a discussion about providing CPUs programmed with bootloader by qrp-labs, but I don't see them in the shop.?
Buying just a new board for $85 will not make much sense vs. a complete QMX kit for $95.
The CPU is approx. $10 at Mouser or Digikey, but without the bootloader it will be useless.?
--

73, DB2OO -Joerg

You might check D109 (the 3V6zener on the #2 plug-in board. Mine was "soft" and although it began conducting a bit below 3.6 volts, it allowed the voltage to climb above 5 volts and caused a hard short to ground within my MCU pretty much like yours.
This is for my curiosity only. I haven't seen any offers of a loaded CPU yet and the replacement would be difficult even if it were.
At this juncture it might be best to order a new kit and start over. If you do then you might check D109 before applying power just to make sure. Couldn't hurt.
Please keep us posted if you find out more.
-Don-

开云体育

D109 is really just a short term fail safe (if the regulator loses control of the 3V supply. It shouldn't get activated unless there's a problem somewhere else (like a short to the +12 supply...)

If it's toast, there was undoubtably a problem but where? :-(

Paul -- AI7JR

You might check D109 (the 3V6zener on the #2 plug-in board. Mine was "soft" and although it began conducting a bit below 3.6 volts, it allowed the voltage to climb above 5 volts and caused a hard short to ground within my MCU pretty much like yours.
This is for my curiosity only. I haven't seen any offers of a loaded CPU yet and the replacement would be difficult even if it were.
At this juncture it might be best to order a new kit and start over. If you do then you might check D109 before applying power just to make sure. Couldn't hurt.
Please keep us posted if you find out more.
-Don-

-- 
Paul -- AI7JR

"an overheating MCU."

How was that determined?
--
Compton
VK2HRX
Sydney, Australia

I've come to a similar conclusion about the availability of a replacement boot-loaded MCU, Instead of starting my hi-band QMX kit I built a hi-band (tr)uSdx kit to go with my lo-band (tr)uSdx. I'm not convinced I made any mistakes that caused my MCU failure, and, frankly I'm enjoying the clunky little orange boxes.

Paul (and others),
Yes, that zener is only the last line of over-voltage defence. I had used the QMX for quite a while before it went dead. It failed before Hans corrected the interrupt priority of the voltage control and I am assuming (barring another explanation) that it was a prolonged hiccup in that voltage control that pressed the bad zener into service. I believe that that diode was faulty all of the time since I have never had one fail in "high impedance" mode. With the older firmware mine locked up frequently. I would have to de-power and start again. When most MCU s lock up like that usually more than one function is not performing and perhaps voltage control happened to be one of those halted processes. If that diode was the cause then I suppose that there might be more of its ilk out and about. Since it isn't a transient fault then a simple test would suffice. Just limit your current applied to the diode to 20 milliamps and if the voltage across it exceeds 4.2 volts then you may have the same problem.
How did I find the short? I used a decent bench DMM, one that measures 10 milliohms (with very low test voltage applied so semiconductors are mostly ignored), and traced it to where the lowest resistance from VDD to common return ground occurred. I methodically unsoldered and lifted each VDD pin on that MCU until the short cleared, measured the MCU VDD pins to ground and found that the short was indeed still there inside the chip.
No heat noticed within the chip by the time I got to it and no sign of heating visible. Failure was quick and enduring. I had to be certain before I mentioned it so that I wouldn't make a fool of myself. I was sad to see it go but I will wait a few more months and order a new one to build. No hurry since I have many more things to do.
72, Don

With my build is was the same. It now a paperweight and what rests is the building pleasure but very pity that the short on the supply board destroyed the mcu and although this was is a factory fault there is non supprt of qrp labs

On 27/01/2024 13:14, Len via groups.io wrote:

although this was is a factory fault there is non supprt of qrp labs

Create a ticket.

Hans will do his best to help.

73 Alan G4ZFQ

Hello all,

I’ll throw my 2 cents in here.

?For the last several months I’ve been repairing many of the qrplabs rigs. ?I’ve done over 80 of the QMX version 1 & 2 rigs. ?Of those, I’ve received 11 that had blown MPUs. ?I use a milliohm meter to isolate the shorts and methodically lift pins. ?In 9 of the 11 cases, it was clear that a gross over voltage of the 3.3v line occurred at some point. ?That will take out the mpu and a few other chips as well. ?In 2 of the cases, the failure in the mpu was due to one pin being bad “LIN_REG_EN”. ?I believe that line went bad because voltage greater than 12v was used to power the rig (like 13.8 or 14.4). ?The mpu pin for that signal is not even rated for 12v! ?So the higher the voltage the more likely that pin will fail.

I think that The gross over voltage on the 9 others was due to intermittent power connections to the rig. ?This is from the stories that the various hams told me. ?This makes sense because the loop in software that controls the PWMs cannot always keep up with a very rapidly changing input voltage. ?In all 9 cases the big tantalum cap and Zener were blown as well. ?I don’t replace the mpu chips on this class of failure. ?Often many other chips and even capacitors were probably destroyed as well. ?While most anything is fixable, there are only so many hours in the day. ?And it’s easier to start over sometimes.

i replaced mpu chips on the 2 that only had that one pin destroyed. ?I got my mcu from a donor board with a known good working mcu. ?Sometimes I get these from hams. ?I also have a small stock of new boards that I bought from Hans for $40 a piece. ?Both transplants were successful.

As an aside, this over voltage can also happen on the 5v line. ?I’ve successfully replaced the entire RX chain of ICs in those cases. ?But I see this less frequently.

73
Jeff
W1NC

"there is non supprt of qrp labs"

That is not my or the experience of others I know.
--
Compton
VK2HRX
Sydney, Australia

开云体育

Agreed!

"there is non supprt of qrp labs"

That is not my or the experience of others I know.
--
Compton
VK2HRX
Sydney, Australia

-- 
Paul -- AI7JR

At the moment of this IR photo we measured 0,9A @2V.
We fed the MCU directly. See the wires in the picture

We determined the heat @ the MCU by a Flir IR addition to a android smartphone.

About no support: I did create 2 times a ticket and my buddy with the same issue created also a ticket.
Until today: no reaction.

in other words a repair is beyond the scope or ability of the average kit builder, so all we can do is bin it? or donate it as a donor board.
Paul DJ0CU G4ADF?

just a note looking at the thermal picture the DIE of the micro is at the center of
the package not offset to one side.

Either something under the chip or even on the back side of the board should
be checked before jumping off the cliff.

--
Allison
------------------
Post online only,?
direct email will go to a bit bucket.