|
Hello all,
I'm not at my first kit building, I built several ones from Small Wonder Labs, a KX1 from Elecraft, a Bitx, and other ones involving arduinos or esp32. So I ordered a? QMX a while back and finally had the time to put it on the bench a few days ago.
It went not so bad, Well ok this kit doesn't allow any shortcuts in building, even with experience you have to stick to the manual or you gonna break something. Fine. I fixed what I broke and it went ok to the end. I plugged the psu I'm using with the kx1, uploaded the firmware, all was fine. Then I bumped into the psu and uplugged it from the QMX... An this one died. D106 and D108 died on the 5V board. I ordered two new boards as they where upgraded since I ordered my kit, received them yesterday and installed them. Everything went fine. I plugged back the QMX on the same psu this morning, put it on the side of the bench to plug it on my tinysa and while moving it, it went dead. AGAIN! god damn it, I've controlled the psu, it outputs 12.3V, everything else is working fine on it so what the hell is going on with my QMX... It's really frustrating to have a rig designed for portable that is so incredibly sensitive to the power connector you plug in. I don't know what to do now, I'll check what burned this time later when I'll cool down but I'd like some advices as how to fix this issue as it's really frustrating.
?
73 de Tom, F8COD
|
Did you check that your barrel connector is 2.1mm and not 2.5mm? Many issues like yours have been caused by the short disconnect / connect of a 2.5mm barrel plug that cannot ensure a good connection.
?
73, Peter
|
Tom,?
Yes, that is frustrating, and shouldn't have happened since the new 5 volt power board has a larger diode on it and the latest firmware is designed to ramp the power up and down, and Hans also tested for that issue specifically with good results. See this article, maybe it will help:
In the mean time, check around the power barrel jack on the PC board for loose or cold solder joints. Check the power supply boards themselves to make sure the tops and bottoms on the connectors are soldered properly. I had the same thing happen to me once, but not since I replaced the boards. I also put new diodes on the old boards so I would have backups.
?
--
73, Dan - W2DLC
|
Yes, that's a good point too. The time I blew the diode on my power supply boards was as a result of having the wrong size barrel plug coming from my power supply.
--
73, Dan - W2DLC
|
Are these blown power supplies still happening with the latest version of firmware for the QMX?
We haven't been seeing many reports of this lately in the forum, I was hoping it was mostly resolved.
?
I think a very brief interruption of power could be causing the problem, perhaps a disconnect of 100 microseconds or less.
The 5v rail falls, and if the processor inspects the ADC_5V input from the SMPS at just the right time when there is still
enough power to keep the processor going, it will bump up the PWM_5V duty cycle and the current through the inductor
quickly ramps up to far more than the nominal 100 ma that is needed before the next time the processor inspects
the ADC input some 1000 microseconds later.? (Likewise for the 3.3v rail with ADC_3V3 and PWM_3V3).
Since this depends on exactly when the processor inspects the ADC input, it won't happen very often.
?
All conjecture on my part, would be worth looking into further if this is happening with current firmware.
Hard to model without knowing exactly what the software is doing to control the PWM signals.
For example, if the software sees that the 12v input has suddenly gone low, it could immediately inhibit
any increase in the PWM duty cycle.? I believe current firmware does something like that, but there may?
be a digital filter between the inspection of the 12v rail and where the PWM duty cycle is calculated.
?
Using a 2.5mm plug in the 2.1mm jack in back of the QMX could definitely make for an intermittent connection there.
And if both truly are 2.1mm, there might still be an issue if these are budget parts with wide dimensional tolerances
and nothing springy in the metal to maintain connection.
?
Moving up to a similar barrel connector from Switchcraft that has a screw lock to keep the plug
securely attached may fix this, but that might be $10 instead of $0.10.
?
Moving to an SMPS IC would allow the supply rail to be inspected at a rate 1000 times faster
and should solve the issue while using reasonably small caps and coils and no increase in cost.
But it is not yet clear to me that this is still a problem worth solving.
?
Jerry, KE7ER
?
?
On Wed, Jan 22, 2025 at 07:59 AM, Daniel Conklin wrote:
Yes, that's a good point too. The time I blew the diode on my power supply boards was as a result of having the wrong size barrel plug coming from my power supply.
|
Jerry,
My issue happened with an older version of the firmware installed and the older power supply boards.? With the installation of new boards and the new firmware, and a proper barrel connector, I haven't had a problem since.
--
73, Dan - W2DLC
|
I can vouch for locking screw-ons of 2.5 mm jacks and plugs.
Switchcraft jacks are not TOO pricey but lordy their plugs cost a bunch.
Found some decent 4 pack screw together 2.5mm jack/plug combos on Amazon for about $10 per set. Enter this in the Amazon Search box:
Ruiwaer 4Sets 5.5mm x 2.5mm Male DC Power Plug Connector with Screw Lock Female Panel Socket Mount Adapter Soldering Power Jack Connector and Screw Lock Female Panel Socket Mount AdapterRuiwaer 4Sets 5.5mm x 2.5mm Male DC Power Plug Connector with Screw Lock Female Panel Sockets?
The jacks are panel mounts so had to fiddle those into place and run insulated, short jumpers to the existing barrel connector pads (after removing the OEMs). Also had to enlarge the through holes in the end panels.?
And have to remember to remove the holding nut before removing the applicable end panels on my QMX HB and QDX LB units. I'll be retro fitting my QMX+ soon too.?
I soldered the plugs onto shortish, flexible wire runs with 12 and 9 volt USB C trigger adapters hard wired on the other ends. Pic attached.?
I put BIG '9' or '12'? labels on the cases of the Q's to remind me which adapter to use, but generally just leave the correct power-in cables screwed onto? the Q's.?
Reasonable bumping or jiggling has not caused any? glitches (or worse) so far.?
GREG KI4NVX?
Yahoo Mail: Search, Organize, Conquer
toggle quoted message
Show quoted text
??
??
On Wed, Jan 22, 2025 at 12:11 PM, Jerry Gaffke via groups.io
<jgaffke@...> wrote:
??
Are these blown power supplies still happening with the latest version of firmware for the QMX?
We haven't been seeing many reports of this lately in the forum, I was hoping it was mostly resolved.
?
I think a very brief interruption of power could be causing the problem, perhaps a disconnect of 100 microseconds or less.
The 5v rail falls, and if the processor inspects the ADC_5V input from the SMPS at just the right time when there is still
enough power to keep the processor going, it will bump up the PWM_5V duty cycle and the current through the inductor
quickly ramps up to far more than the nominal 100 ma that is needed before the next time the processor inspects
the ADC input some 1000 microseconds later.? (Likewise for the 3.3v rail with ADC_3V3 and PWM_3V3).
Since this depends on exactly when the processor inspects the ADC input, it won't happen very often.
?
All conjecture on my part, would be worth looking into further if this is happening with current firmware.
Hard to model without knowing exactly what the software is doing to control the PWM signals.
For example, if the software sees that the 12v input has suddenly gone low, it could immediately inhibit
any increase in the PWM duty cycle.? I believe current firmware does something like that, but there may?
be a digital filter between the inspection of the 12v rail and where the PWM duty cycle is calculated.
?
Using a 2.5mm plug in the 2.1mm jack in back of the QMX could definitely make for an intermittent connection there.
And if both truly are 2.1mm, there might still be an issue if these are budget parts with wide dimensional tolerances
and nothing springy in the metal to maintain connection.
?
Moving up to a similar barrel connector from Switchcraft that has a screw lock to keep the plug
securely attached may fix this, but that might be $10 instead of $0.10.
?
Moving to an SMPS IC would allow the supply rail to be inspected at a rate 1000 times faster
and should solve the issue while using reasonably small caps and coils and no increase in cost.
But it is not yet clear to me that this is still a problem worth solving.
?
Jerry, KE7ER
?
?
On Wed, Jan 22, 2025 at 07:59 AM, Daniel Conklin wrote:
??
Yes, that's a good point too. The time I blew the diode on my power supply boards was as a result of having the wrong size barrel plug coming from my power supply.
|
Yup, I'd seen the Ruiwaer parts previously but haven't ordered any.
Good to know they work well.
?
Jerry,? ?KE7ER
?
On Wed, Jan 22, 2025 at 11:50 AM, Greg McCain wrote:
I can vouch for locking screw-ons of 2.5 mm jacks and plugs.
Switchcraft jacks are not TOO pricey but lordy their plugs cost a bunch.
Found some decent 4 pack screw together 2.5mm jack/plug combos on Amazon for about $10 per set. Enter this in the Amazon Search box:
Ruiwaer 4Sets 5.5mm x 2.5mm Male DC Power Plug Connector with Screw Lock Female Panel Socket Mount Adapter Soldering Power Jack Connector and Screw Lock Female Panel Socket Mount AdapterRuiwaer 4Sets 5.5mm x 2.5mm Male DC Power Plug Connector with Screw Lock Female Panel Sockets?
The jacks are panel mounts so had to fiddle those into place and run insulated, short jumpers to the existing barrel connector pads (after removing the OEMs). Also had to enlarge the through holes in the end panels.?
And have to remember to remove the holding nut before removing the applicable end panels on my QMX HB and QDX LB units. I'll be retro fitting my QMX+ soon too.?
I soldered the plugs onto shortish, flexible wire runs with 12 and 9 volt USB C trigger adapters hard wired on the other ends. Pic attached.?
I put BIG '9' or '12'? labels on the cases of the Q's to remind me which adapter to use, but generally just leave the correct power-in cables screwed onto? the Q's.?
Reasonable bumping or jiggling has not caused any? glitches (or worse) so far.?
GREG KI4NVX?
|
This may be a stupid question but how can you tell by looking at the plug if it is a 2.5 or 2.1mm plug? Having read this discussion I thought it prudent to make sure the plug I am using is the right one. Mine has no markings regarding size.
|
Actually a very good question, it's hard to tell by eye.
You can measure the outside diameter of the plug barrel to verify it is 5.5mm with a micrometer, but all other measurements are difficult.
?
I have 2.5mm jack (such as would be mounted on the board), if the plug does not fit into that jack it is probably 2.1mm.
I also have a 2.1mm plug, if it fits into an unknown jack, then that jack is probably 2.1mm.
?
Most plugs and jacks of approximately this appearance are either 2.1 or 2.5mm.
Some plugs have longer barrels than others (and some jacks accept longer barrels), but that usually doesn't matter too much.
Poorly manufactured jacks and plugs may not be exactly 2.1 or 2.5mm.
?
The majority of jacks and plugs for consumer electronics are 2.1mm.
The QMX+ uses a 2.1mm jack, and most 12v supplies I have laying about the house can plug into it.
The down side is that it is possible to plug a 2.5mm plug into it and get a dicy connection.
?
None of this matters much on most equipement, a flakey power connector just causes a reboot.
?
Jerry, KE7ER
?
On Wed, Jan 22, 2025 at 03:25 PM, Grant wrote:
This may be a stupid question but how can you tell by looking at the plug if it is a 2.5 or 2.1mm plug? Having read this discussion I thought it prudent to make sure the plug I am using is the right one. Mine has no markings regarding size.
|
You can use the shank of a 2.5mm drill bit (or 3/32" in US).?
If it fits in the plug, it is a 2.5mm plug.
If it doesn't fit, it is the correct 2.1mm plug.
|
To determine the jack size, search messages for drill bit sizes. Someone in group gave a simple way to determine using drill bit.?
Actually a very good question, it's hard to tell by eye.
You can measure the outside diameter of the plug barrel to verify it is 5.5mm with a micrometer, but all other measurements are difficult.
?
I have 2.5mm jack (such as would be mounted on the board), if the plug does not fit into that jack it is probably 2.1mm.
I also have a 2.1mm plug, if it fits into an unknown jack, then that jack is probably 2.1mm.
?
Most plugs and jacks of approximately this appearance are either 2.1 or 2.5mm.
Some plugs have longer barrels than others (and some jacks accept longer barrels), but that usually doesn't matter too much.
Poorly manufactured jacks and plugs may not be exactly 2.1 or 2.5mm.
?
The majority of jacks and plugs for consumer electronics are 2.1mm.
The QMX+ uses a 2.1mm jack, and most 12v supplies I have laying about the house can plug into it.
The down side is that it is possible to plug a 2.5mm plug into it and get a dicy connection.
?
None of this matters much on most equipement, a flakey power connector just causes a reboot.
?
Jerry, KE7ER
?
On Wed, Jan 22, 2025 at 03:25 PM, Grant wrote:
This may be a stupid question but how can you tell by looking at the plug if it is a 2.5 or 2.1mm plug? Having read this discussion I thought it prudent to make sure the plug I am using is the right one. Mine has no markings regarding size.
|
It is easy to just look at them.? The one with the yellow insert and the two indents which form a spring is the 2.1mm and the one which is straight metal is the 2.5mm.
?
|
Use the shank of a 2.5mm drill as Stan said. There are many different suppliers. I have 2.1mm plugs that look identical to the 2.5mm ones except for diameter.
?
73, Peter
|
I tend to measure the insides of my barrel plugs by shaving down a flat piece of wood (tongue depressor size) to approximately the right width, poking it into the plug, and shaving it till it fits without too much force. THen measure it with calipers. (The internal-measurement jaws of a caliper don't QUITE work for the task by themselves.
?
Steve
|
Yup--that's how I do it, but I grow weary of walking outside to my shop during our Idaho winters to fetch a drill bit whenever I wish to double check a plug. So...I think I'll find an extra 3/32" bit laying around, or even buy a cheap $9 set at Harbor Freight when on sale, drill the bit's fluted end into the end of a larger dowel or block of wood, drill again for a cord, and hang it on the wall in my shack!
?
73,
Cal AD8Q
?
On Wed, Jan 22, 2025 at 05:15 PM, Stan Dye wrote:
You can use the shank of a 2.5mm drill bit (or 3/32" in US).?
If it fits in the plug, it is a 2.5mm plug.
If it doesn't fit, it is the correct 2.1mm plug.
|
I think I'll get one of these and wear it around my neck:
-- rec -- ad5d --
Yup--that's how I do it, but I grow weary of walking outside to my shop during our Idaho winters to fetch a drill bit whenever I wish to double check a plug. So...I think I'll find an extra 3/32" bit laying around, or even buy a cheap $9 set at Harbor Freight when on sale, drill the bit's fluted end into the end of a larger dowel or block of wood, drill again for a cord, and hang it on the wall in my shack!
?
73,
Cal AD8Q
?
On Wed, Jan 22, 2025 at 05:15 PM, Stan Dye wrote:
You can use the shank of a 2.5mm drill bit (or 3/32" in US).?
If it fits in the plug, it is a 2.5mm plug.
If it doesn't fit, it is the correct 2.1mm plug.
|
2 totally unnecessary frustrations of modern ¡®ham¡¯ life are :
2.5mm ¡®sloppy fit¡¯ DC connector plugs?
USB cables that don¡¯t carry data?
?
Owning a 2.5mm drill bit or a 2.5mm chassis socket quickly finds the first sort. ?I rarely trust a 2.5 plug on a 2.1 socket ..?
?
Either way - Both get a narrow wrap of red heatsink on arrival/detection at my lab and my life gets a little easier?
|
I have to check the connector, I will change the one from the power supply with a 2.1mm, thanks Peter.
|
Thanks for the info Dan, I'll check the two pc boards. I was using the last FW, I thought I was safe with it but well, I'll change the components that failed.
|
Peter HB9EBE
Daniel Conklin
Grant
Steve H
Bruce Akhurst