开云体育

ctrl + shift + ? for shortcuts
© 2025 开云体育
  1. QRPLabs
  2. Topics

#QMX Q103/Q104 short #qmx

 

Hi all

So I counted that we have made 44 assembled QMX here so far, of which 3 had the Q103/Q104 short. So 7%. This was consistent?with my feeling that of the boards I had personally tested (assembled QMX) the short-rate was single figures.?

We also started testing all SMPS boards?here for the short, so that it can be resolved before shipment. So nobody who ever orders a QMX from now on should ever see the short, because if it was there, we would have already noticed it and fixed it. My colleague ran through 73 QMX boards of which 22 had the fault. So 30%. Quite a significant difference which seems to me non-representative statically of the small sample size; so my guess would be that there was perhaps some systematic?tendency perhaps in the way the SMD assembly machines were loaded, so that boards near each other in the boxes we have would perhaps be more likely than boards from another area of the batch or assembly? Who knows.?

Anyway -?
  1. All boards shipped henceforth have been tested by us and remediated if required?
  2. Checking for this fault is in the assembly manual at as something to do before powering up
  3. The fault is described in the troubleshooting guide?
73 Hans G0UPL
GIUSEPPE
 

Good afternoon Hans, thank you very much for informing us.? I would like to ask you for further information, why the shipping time of the QMX power supply boards is so long!? But before shipping, even those in the laboratory will be tried without finding any surprises!? Thanks and good afternoon.

Giuseppe iu8eun

Il Lun 21 Ago 2023, 15:51 Hans Summers <hans.summers@...> ha scritto:
Hi all

So I counted that we have made 44 assembled QMX here so far, of which 3 had the Q103/Q104 short. So 7%. This was consistent?with my feeling that of the boards I had personally tested (assembled QMX) the short-rate was single figures.?

We also started testing all SMPS boards?here for the short, so that it can be resolved before shipment. So nobody who ever orders a QMX from now on should ever see the short, because if it was there, we would have already noticed it and fixed it. My colleague ran through 73 QMX boards of which 22 had the fault. So 30%. Quite a significant difference which seems to me non-representative statically of the small sample size; so my guess would be that there was perhaps some systematic?tendency perhaps in the way the SMD assembly machines were loaded, so that boards near each other in the boxes we have would perhaps be more likely than boards from another area of the batch or assembly? Who knows.?

Anyway -?
  1. All boards shipped henceforth have been tested by us and remediated if required?
  2. Checking for this fault is in the assembly manual at as something to do before powering up
  3. The fault is described in the troubleshooting guide?
73 Hans G0UPL
 

Hello Giuseppe
?
Good afternoon Hans, thank you very much for informing us.? I would like to ask you for further information, why the shipping time of the QMX power supply boards is so long!? But before shipping, even those in the laboratory will be tried without finding any surprises!? Thanks and good afternoon.

Because I have to manufacture them. I designed a smaller board having just the two SMPS boards on it. But the PCBs have to be manufactured, all the components sourced, and the PCB assembly done. Then shipped to me, import customs and taxes. It all takes time.?

Meanwhile I did already say that for some urgent cases I don't mind breaking SMPS boards off existing QMX boards, we can then use them here in Assembled QMX once the new SMPS-only boards arrive. AND, I already emailed you privately off-list offering this to you... but I have not had any reply. So I suppose you haven't seen that email.?

73 Hans G0UPL
GIUSEPPE
 

Hans, I'm sorry, I still haven't read your email after I bought the QMX traction power supply boards.? I'm working and this year I also have a bit to do at home with some jobs.? I'll reply as soon as I read your email, thank you very much, you are always so kind and quick to reply to all the thousands of emails that arrive every day.? 73

Giuseppe iu8eun

Il Lun 21 Ago 2023, 16:19 Hans Summers <hans.summers@...> ha scritto:
Hello Giuseppe
?
Good afternoon Hans, thank you very much for informing us.? I would like to ask you for further information, why the shipping time of the QMX power supply boards is so long!? But before shipping, even those in the laboratory will be tried without finding any surprises!? Thanks and good afternoon.

Because I have to manufacture them. I designed a smaller board having just the two SMPS boards on it. But the PCBs have to be manufactured, all the components sourced, and the PCB assembly done. Then shipped to me, import customs and taxes. It all takes time.?

Meanwhile I did already say that for some urgent cases I don't mind breaking SMPS boards off existing QMX boards, we can then use them here in Assembled QMX once the new SMPS-only boards arrive. AND, I already emailed you privately off-list offering this to you... but I have not had any reply. So I suppose you haven't seen that email.?

73 Hans G0UPL
 

开云体育

A slightly radical 'fix' may be replacing Q103 by a bit of wire between its source and drain. As far as I can see the purpose of Q103 is to prevent reversed polarity problems.? If it's replaced by a short then the reverse polarity protection is lost but Q105 will still control the power.? It might not help if the board has been smoked but may keep people going while waiting for a replacement board, because moving Q103 failed.

BTW Hans I noticed that you have ordered 500 PSU boards, that's another estimate - 25%.

Chris, G5CTH

On 21/08/2023 15:19, Hans Summers wrote:

Hello Giuseppe
?
Good afternoon Hans, thank you very much for informing us.? I would like to ask you for further information, why the shipping time of the QMX power supply boards is so long!? But before shipping, even those in the laboratory will be tried without finding any surprises!? Thanks and good afternoon.

Because I have to manufacture them. I designed a smaller board having just the two SMPS boards on it. But the PCBs have to be manufactured, all the components sourced, and the PCB assembly done. Then shipped to me, import customs and taxes. It all takes time.?

Meanwhile I did already say that for some urgent cases I don't mind breaking SMPS boards off existing QMX boards, we can then use them here in Assembled QMX once the new SMPS-only boards arrive. AND, I already emailed you privately off-list offering this to you... but I have not had any reply. So I suppose you haven't seen that email.?

73 Hans G0UPL
 

Hello Giuseppe, all

My colleagues continued checking?and found more boards with the fault; so in the end I had a pile of about 40 here (similar fault proportion to before). I managed to easily repair them all in less than half an hour.?

In almost?all?cases I observed the AOD403 was angled a few degrees anticlockwise of true, putting its metal drain tab clearly too close to Q104 and causing the solder bridge. About 10 seconds blast with the hot air until I saw the solder look liquidy, then a quick nudge with a small screwdriver to angle it back inline (or slightly angled away clockwise in the other direction), was enough to clear the short. Then re-test with the DVM to make sure. Very quick.?

It is indeed disappointing and upsetting that this happened at all, of course it would be nice if the manufacturers didn't have this degree of slop?in the angle of the AOD403. OR if I had known and moved Q103 a bit further away so it would never have mattered. But these things happen. Over the years the suppliers and manufacturers become trusted partners rather than combatants so it's better to work with them to look for solutions rather than cast blame and destroy the working relationship.

73 Hans G0UPL



On Mon, Aug 21, 2023 at 5:41?PM GIUSEPPE <iw8rsb@...> wrote:
Hans, I'm sorry, I still haven't read your email after I bought the QMX traction power supply boards.? I'm working and this year I also have a bit to do at home with some jobs.? I'll reply as soon as I read your email, thank you very much, you are always so kind and quick to reply to all the thousands of emails that arrive every day.? 73

Giuseppe iu8eun

Il Lun 21 Ago 2023, 16:19 Hans Summers <hans.summers@...> ha scritto:
Hello Giuseppe
?
Good afternoon Hans, thank you very much for informing us.? I would like to ask you for further information, why the shipping time of the QMX power supply boards is so long!? But before shipping, even those in the laboratory will be tried without finding any surprises!? Thanks and good afternoon.

Because I have to manufacture them. I designed a smaller board having just the two SMPS boards on it. But the PCBs have to be manufactured, all the components sourced, and the PCB assembly done. Then shipped to me, import customs and taxes. It all takes time.?

Meanwhile I did already say that for some urgent cases I don't mind breaking SMPS boards off existing QMX boards, we can then use them here in Assembled QMX once the new SMPS-only boards arrive. AND, I already emailed you privately off-list offering this to you... but I have not had any reply. So I suppose you haven't seen that email.?

73 Hans G0UPL
 

On Mon, Aug 21, 2023 at 12:57 PM, Hans Summers wrote:
it would be nice if the manufacturers didn't have this degree of slop?in the angle of the AOD403.
Perhaps the PCB footprint of the part(s) needs to be changed, to shrink the ground pad and its solder mask opening, which will constrain the movement of the part.??
Remember that during reflow soldering the part is literally floating on liquid solder and the only thing constraining its movement is surface tension.??
I had a product with some large Mini-Circuits surface mount amplifiers; the first pass had ~25% shorts using the MCL-recommended pattern.? When we shrank the ground pad (by about 25%) and widened the solder mask clearances we had 0 shorts.? Making the pad and solder mask openings symmetrical with respect to the centerline of the part will also help avoid skewing.
73, Don N2VGU
GIUSEPPE
 

Of course Hans, the producers are trusted people, better keep him quiet, then in the QMX power supply PCB design phase it was necessary to calculate Q104 1 millimeter more spaced, and what I think, I wouldn't want to be wrong.? The card printers do what was previously loaded by the designers so it is obvious that there are errors.? Thanks Hans for all the information you give us.

Giuseppe iu8eun

Il Lun 21 Ago 2023, 20:11 Donald S Brant Jr <dsbrantjr@...> ha scritto:
On Mon, Aug 21, 2023 at 12:57 PM, Hans Summers wrote:
it would be nice if the manufacturers didn't have this degree of slop?in the angle of the AOD403.
Perhaps the PCB footprint of the part(s) needs to be changed, to shrink the ground pad and its solder mask opening, which will constrain the movement of the part.??
Remember that during reflow soldering the part is literally floating on liquid solder and the only thing constraining its movement is surface tension.??
I had a product with some large Mini-Circuits surface mount amplifiers; the first pass had ~25% shorts using the MCL-recommended pattern.? When we shrank the ground pad (by about 25%) and widened the solder mask clearances we had 0 shorts.? Making the pad and solder mask openings symmetrical with respect to the centerline of the part will also help avoid skewing.
73, Don N2VGU

 

Yes, I totally agree on reducing the tab’s footprint. It’s well known that SMD components move during IR reflow and will float freely within the copper region on the board. Where they finally settle is a random variable with component mechanical tolerances also being a factor.

On Mon, Aug 21, 2023 at 1:11 PM Donald S Brant Jr <dsbrantjr@...> wrote:
On Mon, Aug 21, 2023 at 12:57 PM, Hans Summers wrote:
it would be nice if the manufacturers didn't have this degree of slop?in the angle of the AOD403.
Perhaps the PCB footprint of the part(s) needs to be changed, to shrink the ground pad and its solder mask opening, which will constrain the movement of the part.??
Remember that during reflow soldering the part is literally floating on liquid solder and the only thing constraining its movement is surface tension.??
I had a product with some large Mini-Circuits surface mount amplifiers; the first pass had ~25% shorts using the MCL-recommended pattern.? When we shrank the ground pad (by about 25%) and widened the solder mask clearances we had 0 shorts.? Making the pad and solder mask openings symmetrical with respect to the centerline of the part will also help avoid skewing.
73, Don N2VGU

 

Totally agree with the above admonitions to reduce the tab size for
the component.
Tony S. nailed it when he referred to the float and tendency to be
randomly surface tension driven.

JZ

On Mon, Aug 21, 2023 at 2:27?PM Tony Scaminaci <tonyscam@...> wrote:

Yes, I totally agree on reducing the tab’s footprint. It’s well known that SMD components move during IR reflow and will float freely within the copper region on the board. Where they finally settle is a random variable with component mechanical tolerances also being a factor.

On Mon, Aug 21, 2023 at 1:11 PM Donald S Brant Jr <dsbrantjr@...> wrote:

On Mon, Aug 21, 2023 at 12:57 PM, Hans Summers wrote:

it would be nice if the manufacturers didn't have this degree of slop in the angle of the AOD403.

Perhaps the PCB footprint of the part(s) needs to be changed, to shrink the ground pad and its solder mask opening, which will constrain the movement of the part.
Remember that during reflow soldering the part is literally floating on liquid solder and the only thing constraining its movement is surface tension.
I had a product with some large Mini-Circuits surface mount amplifiers; the first pass had ~25% shorts using the MCL-recommended pattern. When we shrank the ground pad (by about 25%) and widened the solder mask clearances we had 0 shorts. Making the pad and solder mask openings symmetrical with respect to the centerline of the part will also help avoid skewing.
73, Don N2VGU
 

Hi all?

The SMD pad size is already smaller than the physical tab (someone in a former?post pointed this out as an error and the pad size should be made larger). I'm not sure how much smaller it should be made...?

Anyway there is going to be no problem here going forward; it's in the manual, in the troubleshooting, quite easy to fix generally, and we are testing all boards for it now before shipping; and for future PCB revisions I moved Q104 a little further away, for which there is plenty of space on the PCB at that location, so there is no further trouble.?

73 Hans G0UPL



On Mon, Aug 21, 2023 at 9:35?PM John Z <jdzbrozek@...> wrote:
Totally agree with the above admonitions to reduce the tab size for
the component.
Tony S. nailed it when he referred to the float and tendency to be
randomly surface tension driven.

JZ

On Mon, Aug 21, 2023 at 2:27?PM Tony Scaminaci <tonyscam@...> wrote:
>
> Yes, I totally agree on reducing the tab’s footprint. It’s well known that SMD components move during IR reflow and will float freely within the copper region on the board. Where they finally settle is a random variable with component mechanical tolerances also being a factor.
>
> On Mon, Aug 21, 2023 at 1:11 PM Donald S Brant Jr <dsbrantjr@...> wrote:
>>
>> On Mon, Aug 21, 2023 at 12:57 PM, Hans Summers wrote:
>>
>> it would be nice if the manufacturers didn't have this degree of slop in the angle of the AOD403.
>>
>> Perhaps the PCB footprint of the part(s) needs to be changed, to shrink the ground pad and its solder mask opening, which will constrain the movement of the part.
>> Remember that during reflow soldering the part is literally floating on liquid solder and the only thing constraining its movement is surface tension.
>> I had a product with some large Mini-Circuits surface mount amplifiers; the first pass had ~25% shorts using the MCL-recommended pattern.? When we shrank the ground pad (by about 25%) and widened the solder mask clearances we had 0 shorts.? Making the pad and solder mask openings symmetrical with respect to the centerline of the part will also help avoid skewing.
>> 73, Don N2VGU
>
>





1 - 11 of 11