|
I've had some issues with my QMX+.? When I received the board the L101 inductor was damaged but appeared functional.? I had the unit fully assembled and working but then IC101, 3.3v regulator fried.? While removing that IC I lifted the pin1 and pin 2 traces.? I was able to replace IC101 and put in jumpers to bypass the pulled traces.? The 3.3v supply is now working.? I decided to replace L101 and managed to lift the pad adjacent to C107.? It didn't fully detach but it did break the through board connection to R113.? I have replaced L101 and installed a jumper to restore the connection from VCC to R113.? I added some extra solder to C107 to bridge over to L101 to help lock it to the board.?
|
First guess is that the voltage divider ratio of R113 and R111 is off by about 10%
--
Brent DeWitt, AB1LF
Milford, MA
|
What would cause this? I didn't change either of those resistors and it was working normally at one point.?
One additional piece of info. If I drop the supply voltage to 11v the diagnostic page supply voltage reads 12.1 and the QMX+ works normally. The 5v on the diagnostic page still reads high though.?
First guess is that the voltage divider ratio of R113 and R111 is off by about 10%
--
Brent DeWitt, AB1LF
Milford, MA
|
Eugene, I do find your report puzzling.
Not sure how the 5v SMPS which is regulated by the processor
can be measured at 5.0v with a DVM when the diagnostics report 5.75v.
The diagnostics are measuring the 5v rail using the same ADC that is used for SMPS regulation.
?
Please repeat your measurement of the 5v rail, also make an accurate measurement of the 3.3v rail.
Do this while watching the 5v measurement on the diagnostics.
?
It is important to measure the 3.3v rail with a DVM since that is the reference for the ADC's
which are used by the processor to measure all voltages.
?
Jerry, KE7ER
|
The following was measured with a Multimeter from JP105 using GND as reference.
With the QMX powered off...
Vin = 12.1v
With the QMX powered on...
Vin = 11.98v
V+ = 11.96v
Vcc = 5.110v
Vdd = 2.928v
While taking these measurements the Diagnostic screen shows 5.76v for 5V, 3.31v for 3V3, and 13.4v for the supply.
Eugene, I do find your report puzzling.
Not sure how the 5v SMPS which is regulated by the processor
can be measured at 5.0v with a DVM when the diagnostics report 5.75v.
The diagnostics are measuring the 5v rail using the same ADC that is used for SMPS regulation.
?
Please repeat your measurement of the 5v rail, also make an accurate measurement of the 3.3v rail.
Do this while watching the 5v measurement on the diagnostics.
?
It is important to measure the 3.3v rail with a DVM since that is the reference for the ADC's
which are used by the processor to measure all voltages.
?
Jerry, KE7ER
|
It appears that your 3.3V supply is still not healthy.? Note on the diagnostics screen that duty cycle = duty max for the 3.3V supply, and your multimeter shows it is really only 2.9V.??
?
It is likely that the low processor voltage is the reason the other voltages read high.? It's voltage is about 11%-12% low, and since as Jerry noted, that is used as the voltage reference, the other voltages will read the same percentage high, which is what you see.
?
So there is something still wrong with your 3.3V supply.? Use the standalone test described in the wiki to test it.? Or order a new one from qrp-labs.com, they are inexpensive.?
|
I was assuming that the measured ~2.9v was a result of the load being put on the 3.3v rail by the processor and other components attached to that rail. The IC101, 78M33, is a new replacement from digi-key.?
It appears that your 3.3V supply is still not healthy.? Note on the diagnostics screen that duty cycle = duty max for the 3.3V supply, and your multimeter shows it is really only 2.9V.??
?
It is likely that the low processor voltage is the reason the other voltages read high.? It's voltage is about 11%-12% low, and since as Jerry noted, that is used as the voltage reference, the other voltages will read the same percentage high, which is what you see.
?
So there is something still wrong with your 3.3V supply.? Use the standalone test described in the wiki to test it.? Or order a new one from , they are inexpensive.?
|
From the QMX/QMX+ startup wiki...
- Produce VDD for the start period
In this QMX state the signal LIN_REG_EN is HIGH (via R101) and "+12V" goes via Q102 to IC101 pin VI. IC101 produce (3.3 + 0.6) V at VO and a approx. 0.8 V lower voltage 3.1 V (or 3.0 V) at VDD, connected via JP103/JP104 to the main board.
You should find around 3 V at VDD.
This aligns with what I am measuring at VDD so that should indicate IC101 is functioning normally.?
I was assuming that the measured ~2.9v was a result of the load being put on the 3.3v rail by the processor and other components attached to that rail. The IC101, 78M33, is a new replacement from digi-key.?
It appears that your 3.3V supply is still not healthy.? Note on the diagnostics screen that duty cycle = duty max for the 3.3V supply, and your multimeter shows it is really only 2.9V.??
?
It is likely that the low processor voltage is the reason the other voltages read high.? It's voltage is about 11%-12% low, and since as Jerry noted, that is used as the voltage reference, the other voltages will read the same percentage high, which is what you see.
?
So there is something still wrong with your 3.3V supply.? Use the standalone test described in the wiki to test it.? Or order a new one from , they are inexpensive.?
|
?
Why would the processor increase the 5v SMPS PWM duty cycle up to what it thinks is 5.76v?
I can't make any sense of how that could happen other than a coding error in the firmware, and that is very unlikely.
Making this even more puzzling, the actual value as measured by the DVM is only 5.110v, which is close to correct.
?
The processor thinks the 3.3v supply is at 3.31v, which is close enough,?
but the DVM shows it to be only 2.928v.
This could be explained by either a faulty DVM or that the band gap voltage reference
inside the STM32F is a bit low (a hardware failure inside the 100 pin processor chip).
?
My suggestion:
1)? Verify that your DVM is reading voltages correctly, perhaps comparing to a different DVM from a friend.
2)? Repeat the readings you made, but this time also measure the power supply current and the voltage of LIN_REG_EN.
?
If you can, reduce your supply voltage as measured into the rig to around 8 Volts.
This will reduce the likelhood of damage to the board if a probe shorts something.
Damage is still possible, so be careful.
?
You could measure power supply current with a second DVM in series with the power supply lead and set to the 1 Amp range,
or by measuring the voltage across a 1 Ohm resistor in series with the 12v lead from your power supply.
A voltage drop of 1 Volt across that resistor means the power supply is giving 1 Amp.
?
The voltage of LIN_REG_EN can be measured at the gate of Q101or Q110, the other probe of the DVM should be on ground.
As shown on the first page of the datasheet, if you hold the board so the single drain pin is above then the gate is on the lower left
?
If LIN_REG_EN is high (3 Volts or more) then 3.3v rail is getting powered from the 78M33 at IC101 and supply current should be several hundred mA.
If LIN_REG_EN is low (close to 0 Volts) then the 3.3v rail is powered from the much more efficient 3.3v SMPS at Q109 and supply current should be around 100 mA.
The 78M33 is only used briefly at power up, once the 3.3v SMPS is working the processor switches over to using that instead.
?
Stan said:? "So there is something still wrong with your 3.3V supply.? Use the standalone test described in the wiki to test it.? Or order a new one from qrp-labs.com, they are inexpensive."
The photo in your first post shows that this is a later version of the QMX+ with the 3.3v supply integrated into the main board.
It cannot be tested in standalone mode or replaced.?
?
You describe a number of hardware fixes you made to the board.
Did you conduct diagnostic tests prior to those fixes, and if so can you tell us what voltages were reported then?
?
In your first post you say "I had the unit fully assembled and working but then IC101, 3.3v regulator fried."
Anything you can tell us about why IC101 fried?? That seldom happens.? Were you probing the board when it happened???
?
Unless you have an error in the STM32F's internal band gap reference (we've never seen that happen),
this should be easily fixable if we can figure out what's going on.
?
Jerry, KE7ER
?
?
On Sat, May 31, 2025 at 10:04 PM, Eugene Pressley wrote:
The following was measured with a Multimeter from JP105 using GND as reference.
With the QMX powered off...
Vin = 12.1v
With the QMX powered on...
Vin = 11.98v
V+ = 11.96v
Vcc = 5.110v
Vdd = 2.928v
While taking these measurements the Diagnostic screen shows 5.76v for 5V, 3.31v for 3V3, and 13.4v for the supply.
|
According to the QMX/QMX+ startup wiki page I should be reading 3v at VDD. This aligns with the 2.928v I'm measuring at VDD. The DMM is accurately measuring voltage, my power supply has a digital voltage and current meter which agrees with DMM. At 12v, when the QMX+ is running, I am seeing about 490ma of current draw at the power supply.?
What I think caused the 3.3v regulator to blow was unplugging the usb-c while CAT control was actively connected to WJST-X. Lesson learned there, don't unplug anything from the QMX while powered on.?
?
Why would the processor increase the 5v SMPS PWM duty cycle up to what it thinks is 5.76v?
I can't make any sense of how that could happen other than a coding error in the firmware, and that is very unlikely.
Making this even more puzzling, the actual value as measured by the DVM is only 5.110v, which is close to correct.
?
The processor thinks the 3.3v supply is at 3.31v, which is close enough,?
but the DVM shows it to be only 2.928v.
This could be explained by either a faulty DVM or that the band gap voltage reference
inside the STM32F is a bit low (a hardware failure inside the 100 pin processor chip).
?
My suggestion:
1)? Verify that your DVM is reading voltages correctly, perhaps comparing to a different DVM from a friend.
2)? Repeat the readings you made, but this time also measure the power supply current and the voltage of LIN_REG_EN.
?
If you can, reduce your supply voltage as measured into the rig to around 8 Volts.
This will reduce the likelhood of damage to the board if a probe shorts something.
Damage is still possible, so be careful.
?
You could measure power supply current with a second DVM in series with the power supply lead and set to the 1 Amp range,
or by measuring the voltage across a 1 Ohm resistor in series with the 12v lead from your power supply.
A voltage drop of 1 Volt across that resistor means the power supply is giving 1 Amp.
?
The voltage of LIN_REG_EN can be measured at the gate of Q101or Q110, the other probe of the DVM should be on ground.
As shown on the first page of the datasheet, if you hold the board so the single drain pin is above then the gate is on the lower left
?
If LIN_REG_EN is high (3 Volts or more) then 3.3v rail is getting powered from the 78M33 at IC101 and supply current should be several hundred mA.
If LIN_REG_EN is low (close to 0 Volts) then the 3.3v rail is powered from the much more efficient 3.3v SMPS at Q109 and supply current should be around 100 mA.
The 78M33 is only used briefly at power up, once the 3.3v SMPS is working the processor switches over to using that instead.
?
Stan said:? "So there is something still wrong with your 3.3V supply.? Use the standalone test described in the wiki to test it.? Or order a new one from , they are inexpensive."
The photo in your first post shows that this is a later version of the QMX+ with the 3.3v supply integrated into the main board.
It cannot be tested in standalone mode or replaced.?
?
You describe a number of hardware fixes you made to the board.
Did you conduct diagnostic tests prior to those fixes, and if so can you tell us what voltages were reported then?
?
In your first post you say "I had the unit fully assembled and working but then IC101, 3.3v regulator fried."
Anything you can tell us about why IC101 fried?? That seldom happens.? Were you probing the board when it happened???
?
Unless you have an error in the STM32F's internal band gap reference (we've never seen that happen),
this should be easily fixable if we can figure out what's going on.
?
Jerry, KE7ER
?
?
On Sat, May 31, 2025 at 10:04 PM, Eugene Pressley wrote:
The following was measured with a Multimeter from JP105 using GND as reference.
With the QMX powered off...
Vin = 12.1v
With the QMX powered on...
Vin = 11.98v
V+ = 11.96v
Vcc = 5.110v
Vdd = 2.928v
While taking these measurements the Diagnostic screen shows 5.76v for 5V, 3.31v for 3V3, and 13.4v for the supply.
|
Eugene,
?
Most of that 490 mA is going through the 78M33 to make 3.3v for the processor, the 78M33 would be dissipating (12v-3.3v)*0.490a = 4.3 Watts.
That's a lot, it's probably getting very hot, surprised they don't blow more often when running continuously without switching to the 3.3v SMPS.
That alone is a good reason to reduce your power supply voltage to 7 or 8 volts when diagnosing this, the 78M33 will be much cooler.
Unplugging the USB cable from the QMX+ may have had something to do with it blowing, but I don't see how.
?
At 490 mA, LIN_REG_EN is probably still high because the processor hasn't managed to bring the SMPS into regulation.
I assume LIN_REG_EN goes low as soon as the 3.3v SMPS is in regulation, though possible it waits till the 5v SMPS is behaving as well.
?
We still have those two weird issues at the top of my previous response that I have no explanation for.
The second one might be explained by a bad band gap reference in the STM32F, but that seems improbable.
?
Jerry, KE7ER
?
?
On Sun, Jun 1, 2025 at 09:15 AM, Eugene Pressley wrote:
According to the QMX/QMX+ startup wiki page I should be reading 3v at VDD. This aligns with the 2.928v I'm measuring at VDD. The DMM is accurately measuring voltage, my power supply has a digital voltage and current meter which agrees with DMM. At 12v, when the QMX+ is running, I am seeing about 490ma of current draw at the power supply.?
?
What I think caused the 3.3v regulator to blow was unplugging the usb-c while CAT control was actively connected to WJST-X. Lesson learned there, don't unplug anything from the QMX while powered on.?
|
On Sun, Jun 1, 2025 at 06:15 PM, Eugene Pressley wrote:
According to the QMX/QMX+ startup wiki page I should be reading 3v at VDD. This aligns with the 2.928v I'm measuring at VDD.
Eugene, sorry for the intrusion. Around 3.0 V only for the first fraction of the first second. Please read
QMX/QMX+ startup wiki...
until the end for the start up process. Within the first second or at least at the end of the first second No. 6 should be reached.
From this Wiki page:
6. Switch internal power supply to the SMPS
If both SMPS are up and VDD and VCC are ok the controller switches LIN_REG_EN to GND (0.x V). This will switch off Q102 and disable the linear regulator and will switch on Q111. Via Q111 VDD = 3.3 V is delivered now from the 3.3 V SMPS.
In RX mode the current is lower than 200 mA. The value is dependent upon the voltage at V_IN (around 95 mA at 12 V and a bit higher ar 9 V). It should be possible to connect a terminal emulation (PuTTY or so) and show at least the Diagnostics screen. There should be no red mark for the SMPS values. VDD should be 3.3 V and VCC 5.0 V.
So in regular operation the linear regulator must be switched off (LIN_REG_EN at 0 V).
Nearly all this was written before in this thread. And like written before the shown voltages are affected by the to low VDD. We have to check using a DMM.
At 12v, when the QMX+ is running, I am seeing about 490ma of current draw at the power supply.?
To much, around 120 or 140 mA would be ok. 490 mA also before IC101 was fried?
What I think caused the 3.3v regulator to blow was unplugging the usb-c while CAT control was actively connected to WJST-X.
But I guess not because you interrupted the USB data transfer. And in regular operation IC101 must be off, how should it be affected?
Maybe the mechanical force from unplug the USB triggered a "latent" failure in the QMX+.
?
For further debugging I recommend a voltage of 7.5 V to lower the danger for more faults.
Could you report please these values (after the QMX+ was running for some seconds):
- I in @ 7.5 V
- screen dump from Diagnostics
- Voltage at LIN_REG_EN
- Voltage at S, G and D of Q111
- Voltage at S, G and D of Q102
- Voltage at IC101 pins in, out and "GND"
- Voltage "+12V", VDD, VCC, ADC_5V and ADC_3V3
I know some points should show the same voltage. But better we check after the history of this QMX+.
?
73 Ludwig
?
?
|
And something more:
?
When you start the terminal session do you have to go through the menue structure to get Diagnostics or is Diagnostics starting direct?
It looks like the first way but I ask to be sure.
?
When the QMX+ is starting via IC101 the current should be not higher than around 200 mA, maybe 250 mA. And this only for the first second or less.
?
73 Ludwig
|
1. I in?@ 7.5V 0.331A measured at the power supply, current limited to 0.5A. 2. Screenshot of Diagnostic page with power supply?@ 7.5V, see below.? I had to select the Diagnostic Page from the menu. 3. Voltage at LIN_REG_EN 0.002V measured at PIN1 (G) of Q101, which is connected to LIN_REG_EN. 4.? Q111 Voltages at S, G, D PIN1 (G) = 0.002V, PIN2 (S) = 2.939V, PIN3 (D)?= 2.939V 5. Q102 Voltages at S, G, D PIN1 (G) = 6.82V, PIN2 (S) = 7.36V, PIN3 (D) = 7.32V 6. IC101 Voltages at IN, OUT, GND IN = 7.32V, OUT = 3.886V, GND = 0.635V 7. Voltages at?+12V, VDD, VCC, ADC_5V, ADC_3V3 Measured at JP105 +12V = 7.44V, VDD = 2.935V, VCC = 5.009V Measured between R111 and R113 ADC_5V = 2.527V Measured between R110 and R112 ADC_3V3 = 1.452V 
On Sun, Jun 1, 2025 at 06:15 PM, Eugene Pressley wrote:
According to the QMX/QMX+ startup wiki page I should be reading 3v at VDD. This aligns with the 2.928v I'm measuring at VDD.
Eugene, sorry for the intrusion. Around 3.0 V only for the first fraction of the first second. Please read
QMX/QMX+ startup wiki...
until the end for the start up process. Within the first second or at least at the end of the first second No. 6 should be reached.
From this Wiki page:
6. Switch internal power supply to the SMPS
If both SMPS are up and VDD and VCC are ok the controller switches LIN_REG_EN to GND (0.x V). This will switch off Q102 and disable the linear regulator and will switch on Q111. Via Q111 VDD = 3.3 V is delivered now from the 3.3 V SMPS.
In RX mode the current is lower than 200 mA. The value is dependent upon the voltage at V_IN (around 95 mA at 12 V and a bit higher ar 9 V). It should be possible to connect a terminal emulation (PuTTY or so) and show at least the Diagnostics screen. There should be no red mark for the SMPS values. VDD should be 3.3 V and VCC 5.0 V.
So in regular operation the linear regulator must be switched off (LIN_REG_EN at 0 V).
Nearly all this was written before in this thread. And like written before the shown voltages are affected by the to low VDD. We have to check using a DMM.
At 12v, when the QMX+ is running, I am seeing about 490ma of current draw at the power supply.?
To much, around 120 or 140 mA would be ok. 490 mA also before IC101 was fried?
What I think caused the 3.3v regulator to blow was unplugging the usb-c while CAT control was actively connected to WJST-X.
But I guess not because you interrupted the USB data transfer. And in regular operation IC101 must be off, how should it be affected?
Maybe the mechanical force from unplug the USB triggered a "latent" failure in the QMX+.
?
For further debugging I recommend a voltage of 7.5 V to lower the danger for more faults.
Could you report please these values (after the QMX+ was running for some seconds):
- I in @ 7.5 V
- screen dump from Diagnostics
- Voltage at LIN_REG_EN
- Voltage at S, G and D of Q111
- Voltage at S, G and D of Q102
- Voltage at IC101 pins in, out and "GND"
- Voltage "+12V", VDD, VCC, ADC_5V and ADC_3V3
I know some points should show the same voltage. But better we check after the history of this QMX+.
?
73 Ludwig
?
?
|
Ok, nice data. lets check each.
?
1. I in @ 7.5V
0.331A measured at the power supply, current limited to 0.5A.
I think 0.331 A is to much. It looks like an additional load.
?
2. Screenshot of Diagnostic page with power supply @ 7.5V, see below. ?I had to select the Diagnostic Page from the menu.
At switch up process No. 6 the MCU is checking both SMPS. Before this, when the MCU is running from IC101, only Diagnostics but no menu is shown via terminal. In case of the SMPS are checked ok the MCU is switching from IC101 to SMPS (LIN_REG_EN to 0 V). As far as I know the 3.3 V SMPS is checked when feeding the internal dummy load R114. This load is necessary to run the SMPS well.
?
Let's assume there is a to high load at the VDD rail. When VDD is produced by IC101 there is no alarm. IC101 can deliver the current. When switching to the SMPS (LIN_REG_EN = 0 V) VDD produced by the SMPS is dropping because the to high load. The SMPS is going to the limit (Duty cycle = Duty max) but couldn't produce enough to hold 3.3 V. Or there is a failure inside the SMPS reducing the amount of current produced from a particular duty cycle.
?
The screen shows both SMPS are running. The 5V SMPS is red only because of the wrong 3.3 V reference. The value 3.3 V is shown only because of some special calculation to build the displayed numbers. The MCU has already checked both SMPS as ok.
?
3. Voltage at LIN_REG_EN
0.002V measured at PIN1 (G) of Q101, which is connected to LIN_REG_EN.
see above
4. ?Q111 Voltages at S, G, D
PIN1 (G) = 0.002V, PIN2 (S) = 2.939V, PIN3 (D) = 2.939V
SMPS out and VDD are connected, ok
5. Q102 Voltages at S, G, D
PIN1 (G) = 6.82V, PIN2 (S) = 7.36V, PIN3 (D) = 7.32V
Voltage at G should be close to voltage at S, so here not ok. It looks like the switch Q102 is ON, not ok because of LIN_REG_EN = 0 V. -0.54 V G to S shouldn't be enough to switch it on. It looks like something is broken, maybe Q102.
6. IC101 Voltages at IN, OUT, GND
IN = 7.32V, OUT = 3.886V, GND = 0.635V
It looks like IC101 is working. Q102 delivers current (see above).
7. Voltages at +12V, VDD, VCC, ADC_5V, ADC_3V3
Measured at JP105 +12V = 7.44V, VDD = 2.935V, VCC = 5.009V
Measured between R111 and R113 ADC_5V = 2.527V
Measured between R110 and R112 ADC_3V3 = 1.452
Maybe only IC101 delivers current to VDD and the 3.3 V SMPS not (enough). VDD is connected via Q111 to the 3.3 V SMPS output and so ADC_3V3 shows 0.5 * VDD = 1.452 V.
?
In the end we may have two or three problems:
- Q102 is on despite LIN_REG_EN = 0 V, maybe a broken Q102
- 3.3 V SMPS delivers not enough current, maybe S of Q111 isn't well connected to the SMPS output. Or more probably there is something wrong around Q113 and D105. A failure there would reduce the current produced from a particular duty cycle.
- Where goes the extensive current?
Nice, failure No 1 would compensate failure No 2. Without No 1 after switch over to 3.3 V SMPS the MCU would stop to work (not enough power from the SMPS).
?
It's late in the evening. I hope I made no failure.
?
73 Ludwig
|
Eugene,
?
?
I looked hard at your photographs in the initial post showing the wires you added.
Didn't notice any errors.
?
Ludwig gave you some great input.
Here's my comments on what Ludwig said:
?
>? At switch up process No. 6 the MCU is checking both SMPS. Before this, when the MCU is running from IC101,
>? only Diagnostics but no menu is shown via terminal. In case of the SMPS are checked ok the MCU is switching
>? from IC101 to SMPS (LIN_REG_EN to 0 V). As far as I know the 3.3 V SMPS is checked when feeding the
>? internal dummy load R114. This load is necessary to run the SMPS well.
?
Some of that might be hard to understand.
Regardless, he is saying that the MCU firmware has made it through step 6 of the QMX/QMX+ wiki,
both SMPS's are up,? and LIN_REG_EN is low to shut down IC101 by turning off Q102.
?
>? Let's assume there is a to high load at the VDD rail. When VDD is produced by IC101 there is no alarm. IC101 can deliver the current.
>? When switching to the SMPS (LIN_REG_EN = 0 V) VDD produced by the SMPS is dropping because the to high load.
>? The SMPS is going to the limit (Duty cycle = Duty max) but couldn't produce enough to hold 3.3 V.
>? Or there is a failure inside the SMPS reducing the amount of current produced from a particular duty cycle.
?
Ludwig is saying that the 3.3v rail has a load somewhere on the board that is taking extra power.
The 78LM33 at IC101 is fairly strong, and was able to get the processor working in spite of that extra power.
But even with the SMPS running properly, it cannot get the 3.3v rail all the way up to where it should be
because the SMPS is limited in firmware to some maximum duty cycle.
?
I'm not so sure about that last part.? Maximum current of an SMPS is not a function of duty cycle,
the duty cycle of a buck mode SMPS is primarily determined by the ratio of output voltage to the input voltage.
?
My guess is that there is something in the 3.3v SMPS that prevents it from working properly.
The processor thinks it is working properly since the 3.3v rail remains high even after it turns off Q102,
but it is fooled because as Ludwig says, Q102 is also broken somehow and continues driving IC101.
?
>? I think 0.331 A is to much. It looks like an additional load.
?
That additional load on 3.3v may have caused the initial failure of IC101.
?
?
>? Voltage at G should be close to voltage at S, so here not ok.
>? It looks like the switch Q102 is ON, not ok because of LIN_REG_EN = 0 V. -0.54 V G
>? to S shouldn't be enough to switch it on. It looks like something is broken, maybe Q102.
?
Agreed.
Processor has de-asserted LIN_REG_EN, so now Q102-G should be very close to Q102-S, turning Q102 off.
It is kind of close,? only 7.36v - 6.82v = 0.54v, but not as close as we would expect.
From the BSS84 datasheet, Q102 should not turn on till that difference exceeds 0.80v,
so should be close enough to turn off Q102.
But apparently Q102 is still on for some reason, since Q102-D is at 7.32v, supplying power to IC101.
Very weird.
?
?
?
?
Why do the diagnostics report the 3.3v rail to be 3.31v when?
a DVM shows that rail to be at 2.93v?? The processor can accurately measure the Vdd rail
by reading the voltage of its internal 1.1v band gap reference with an ADC.
That ADC in turn uses the 3.3v rail as its reference.?
A bit confusing, but from this the processor can compute the true value of the 3.3v rail.
There should be no difference between the DVM reading and what the diagnostics report.
Perhaps the current firmware for the diagnostics does not do this computation,
the diagnostics simply report the reading from the ADC?
?
Similarly, the diagnostics report a 5v SMPS duty cycle of 63%, and a voltage of 5.63v.
Perhaps the diagnostics report just the reading from the ADC, but the firmware?
that sets the SMPS duty cycle is doing a more complete computation that includes?
using the internal 1.1v band gap reference?
?
Not sure I have any advice right now on how to proceed.
You might look for a trouble around Q102 and get a replacement BSS84 lined up,
as it seems that part is somehow failing.
Finding the cause of excessive current on the 3.3v rail may be more difficult.
?
Quite the puzzle.
Hope you enjoy puzzles.
?
Jerry, KE7ER
?
?
?
On Sun, Jun 1, 2025 at 02:47 PM, @Ludwig_DH8WN wrote:
Ok, nice data. lets check each.
?
1. I in @ 7.5V
0.331A measured at the power supply, current limited to 0.5A.
I think 0.331 A is to much. It looks like an additional load.
?
2. Screenshot of Diagnostic page with power supply @ 7.5V, see below. ?I had to select the Diagnostic Page from the menu.
At switch up process No. 6 the MCU is checking both SMPS. Before this, when the MCU is running from IC101, only Diagnostics but no menu is shown via terminal. In case of the SMPS are checked ok the MCU is switching from IC101 to SMPS (LIN_REG_EN to 0 V). As far as I know the 3.3 V SMPS is checked when feeding the internal dummy load R114. This load is necessary to run the SMPS well.
?
Let's assume there is a to high load at the VDD rail. When VDD is produced by IC101 there is no alarm. IC101 can deliver the current. When switching to the SMPS (LIN_REG_EN = 0 V) VDD produced by the SMPS is dropping because the to high load. The SMPS is going to the limit (Duty cycle = Duty max) but couldn't produce enough to hold 3.3 V. Or there is a failure inside the SMPS reducing the amount of current produced from a particular duty cycle.
?
The screen shows both SMPS are running. The 5V SMPS is red only because of the wrong 3.3 V reference. The value 3.3 V is shown only because of some special calculation to build the displayed numbers. The MCU has already checked both SMPS as ok.
?
3. Voltage at LIN_REG_EN
0.002V measured at PIN1 (G) of Q101, which is connected to LIN_REG_EN.
see above
4. ?Q111 Voltages at S, G, D
PIN1 (G) = 0.002V, PIN2 (S) = 2.939V, PIN3 (D) = 2.939V
SMPS out and VDD are connected, ok
5. Q102 Voltages at S, G, D
PIN1 (G) = 6.82V, PIN2 (S) = 7.36V, PIN3 (D) = 7.32V
Voltage at G should be close to voltage at S, so here not ok. It looks like the switch Q102 is ON, not ok because of LIN_REG_EN = 0 V. -0.54 V G to S shouldn't be enough to switch it on. It looks like something is broken, maybe Q102.
6. IC101 Voltages at IN, OUT, GND
IN = 7.32V, OUT = 3.886V, GND = 0.635V
It looks like IC101 is working. Q102 delivers current (see above).
7. Voltages at +12V, VDD, VCC, ADC_5V, ADC_3V3
Measured at JP105 +12V = 7.44V, VDD = 2.935V, VCC = 5.009V
Measured between R111 and R113 ADC_5V = 2.527V
Measured between R110 and R112 ADC_3V3 = 1.452
Maybe only IC101 delivers current to VDD and the 3.3 V SMPS not (enough). VDD is connected via Q111 to the 3.3 V SMPS output and so ADC_3V3 shows 0.5 * VDD = 1.452 V.
?
In the end we may have two or three problems:
- Q102 is on despite LIN_REG_EN = 0 V, maybe a broken Q102
- 3.3 V SMPS delivers not enough current, maybe S of Q111 isn't well connected to the SMPS output. Or more probably there is something wrong around Q113 and D105. A failure there would reduce the current produced from a particular duty cycle.
- Where goes the extensive current?
Nice, failure No 1 would compensate failure No 2. Without No 1 after switch over to 3.3 V SMPS the MCU would stop to work (not enough power from the SMPS).
?
It's late in the evening. I hope I made no failure.
?
|
The 3.3 SMPS must be working well enough that the processor/MCU sees?
that it got up to 3.3v by reading an ADC, and then switches the 78LM33 out
and the 3.3v SMPS in by bringing LIN_REG_EN low.
?
I believe it has to reach 3.3v, 3.29v isn't good enough, before that switch over happens.
Before the switch over happens, the 3.3v SMPS is loaded by the 47 Ohms at R114
which is about as much of a load as the radio takes after the switch over.
After the switch over, the current through R114 is shut down by LIN_REG_EN
taking the gate of Q112 low.
?
However, the failure at Q102 means the 78LM33 didn't get switched out,
so both the 78LM33 and the 3.3v SMPS are operating in parallel after the switch over.
But since the 3.3v rail is only at around 3.0v, our best guess is that with
the 3.3v SMPS switched in by turning on Q111, that SMPS got dragged down
from the 3.3v that the processor saw before to only 3.0v with Q111 on.
This happens because the rest of the rig is apparently taking more power than expected.
?
At first power up when only the 78LM33 is powering the 3.3v rail, it is probably
even lower than 3.0v since it doesn't have the SMPS helping.
The processor may not care much so long as it is working well enough
to get the 3.3v SMPS going and do the switch over.
?
Probably more than you want to know, but that's how I think it works.
Issues are as Ludwig identified them:
? Q102 does not turn off, looks like that transistor is bad.
? Something in the rig is drawing more current from 3.3v.
?
You may find where the extra current is going by feeling around and?
finding something that is getting warmer than it should after a few minutes with power on.
Do this with that a 7.5v supply, do not increase it till the problem is solved..
?
Check the voltage across R114,? should be close to zero volts.
The top and bottom of that resistor should both be at around 3.0 volts.
Seems unlikely, but perhaps LIN_REG_EN cannot turn off Q110, just like it cannot turn off Q102.
If so, the load presented by R114 could be the extra current on the 3.3v rail we need to find.
?
Jerry? KE7ER
?
?
On Sun, Jun 1, 2025 at 05:36 PM, Jerry Gaffke wrote:
My guess is that there is something in the 3.3v SMPS that prevents it from working properly.
The processor thinks it is working properly since the 3.3v rail remains high even after it turns off Q102,
but it is fooled because as Ludwig says, Q102 is also broken somehow and continues driving IC101.
?
|
I decided to replace Q102 and Q101, since I have spares.? Once that was replaced I decided to tidy up the patch from pin 1 to C102.? Somehow, I created a short across C102.? This was a?+12V rail to GND short.? It took me a bit to find and correct that.? I removed Q102 and Q101 again while trying to find that short.? This short was new and not part of the original issue but it may have created new problems.? I replace Q102 and Q101 with fresh components (just in case the short caused an issue with them).? Now when I power on the display will not turn on.? When I go to the console it is in emergency diagnostic mode.  Repeating the previous measurements I [email protected] is 0.240A.? Oddly when I open the diagnostic screen the current jumps to 0.340A. 2. Diagnostic Screenshot, See Above. 3. I measured the LIN_REG_EN previously at the wrong pin of Q101.? Pin2 is the S and what is connected to LIN_REG_EN.? It now reads 4.941V. 4. Q111 Voltages PIN1 (G) = 4.945V, PIN2 (S) = 4.039V, PIN3 (D) = 2.896V 5. Q102 Voltages PIN1 (G) = 0V, PIN2 (S) = 2.43V, PIN3 (D) =? 7.00V 6. IC101 Voltages IN = 7V, OUT = 3.88V, GND = 0.632V 7. Other Voltages +12 = 7.44V, VDD=2.89V, VCC=2.267V. ADC_5V between R111&R113 = 1.121V ADC_3V3 between R110&R112 = 1.98V
Ok, nice data. lets check each.
?
1. I in @ 7.5V
0.331A measured at the power supply, current limited to 0.5A.
I think 0.331 A is to much. It looks like an additional load.
?
2. Screenshot of Diagnostic page with power supply @ 7.5V, see below.? I had to select the Diagnostic Page from the menu.
At switch up process No. 6 the MCU is checking both SMPS. Before this, when the MCU is running from IC101, only Diagnostics but no menu is shown via terminal. In case of the SMPS are checked ok the MCU is switching from IC101 to SMPS (LIN_REG_EN to 0 V). As far as I know the 3.3 V SMPS is checked when feeding the internal dummy load R114. This load is necessary to run the SMPS well.
?
Let's assume there is a to high load at the VDD rail. When VDD is produced by IC101 there is no alarm. IC101 can deliver the current. When switching to the SMPS (LIN_REG_EN = 0 V) VDD produced by the SMPS is dropping because the to high load. The SMPS is going to the limit (Duty cycle = Duty max) but couldn't produce enough to hold 3.3 V. Or there is a failure inside the SMPS reducing the amount of current produced from a particular duty cycle.
?
The screen shows both SMPS are running. The 5V SMPS is red only because of the wrong 3.3 V reference. The value 3.3 V is shown only because of some special calculation to build the displayed numbers. The MCU has already checked both SMPS as ok.
?
3. Voltage at LIN_REG_EN
0.002V measured at PIN1 (G) of Q101, which is connected to LIN_REG_EN.
see above
4.? Q111 Voltages at S, G, D
PIN1 (G) = 0.002V, PIN2 (S) = 2.939V, PIN3 (D) = 2.939V
SMPS out and VDD are connected, ok
5. Q102 Voltages at S, G, D
PIN1 (G) = 6.82V, PIN2 (S) = 7.36V, PIN3 (D) = 7.32V
Voltage at G should be close to voltage at S, so here not ok. It looks like the switch Q102 is ON, not ok because of LIN_REG_EN = 0 V. -0.54 V G to S shouldn't be enough to switch it on. It looks like something is broken, maybe Q102.
6. IC101 Voltages at IN, OUT, GND
IN = 7.32V, OUT = 3.886V, GND = 0.635V
It looks like IC101 is working. Q102 delivers current (see above).
7. Voltages at +12V, VDD, VCC, ADC_5V, ADC_3V3
Measured at JP105 +12V = 7.44V, VDD = 2.935V, VCC = 5.009V
Measured between R111 and R113 ADC_5V = 2.527V
Measured between R110 and R112 ADC_3V3 = 1.452
Maybe only IC101 delivers current to VDD and the 3.3 V SMPS not (enough). VDD is connected via Q111 to the 3.3 V SMPS output and so ADC_3V3 shows 0.5 * VDD = 1.452 V.
?
In the end we may have two or three problems:
- Q102 is on despite LIN_REG_EN = 0 V, maybe a broken Q102
- 3.3 V SMPS delivers not enough current, maybe S of Q111 isn't well connected to the SMPS output. Or more probably there is something wrong around Q113 and D105. A failure there would reduce the current produced from a particular duty cycle.
- Where goes the extensive current?
Nice, failure No 1 would compensate failure No 2. Without No 1 after switch over to 3.3 V SMPS the MCU would stop to work (not enough power from the SMPS).
?
It's late in the evening. I hope I made no failure.
?
73 Ludwig
|
Eugene,
?
This is looking worse.
Looks like something on the 5v rail is now failing.
And that 5v rail is not touching any of the parts you were working on.
?
The 4.62v on the 3.3v SMPS is especially alarming, but your DVM measurements show the 3.3v rail to be a safe 2.89v.?
I'm having a hard time figuring out why the diagnostics report 4.62, even after factoring in?
how the actual 2.89v will affect what the ADC readings since 2.89v is used as the ADC reference.
And I have no idea why the output of IC101 would be high at 3.88v.
?
The 2.22v reported on the 5.0v rail also doesn't make sense, since the low 3.3v rail
would cause the ADC to report a higher value than your 2.267v DVM reading.
?
Whacking at those power supplies is rather dangerous, at some point those supplies may decide to go
too high instead of too low and take out half the parts on the board.
I had thought I mostly understood what was going, but these new readings have me puzzled.
If you are willing to accept the loss of the board as a good educational experience,?
then continue on assuming somebody in the forum can figure out a direction to go on this.
If you care more about the board, there might come a time to consider having Jeff Moore to sort this out for you.
?
Jerry, KE7ER
?
?
?
On Sun, Jun 1, 2025 at 09:19 PM, Eugene Pressley wrote:
I decided to replace Q102 and Q101, since I have spares.? Once that was replaced I decided to tidy up the patch from pin 1 to C102.? Somehow, I created a short across C102.? This was a?+12V rail to GND short.? It took me a bit to find and correct that.? I removed Q102 and Q101 again while trying to find that short.? This short was new and not part of the original issue but it may have created new problems.? I replace Q102 and Q101 with fresh components (just in case the short caused an issue with them).? Now when I power on the display will not turn on.? When I go to the console it is in emergency diagnostic mode.
|
Yeah, I've already been thinking calling this a learning experience and ordering a new kit. I'd then pull the toroids from the current board since they were working well.?
As a last ditch effort I'm thinking about disabling the 5v supply and possibly the 3.3v supply. I have two very small adjustable voltage regulator boards, from DROK, one set to 5v and the other set to 3.3v. I can power these from the switched +12v then supply vcc and vdd on JP105.?
Eugene,
?
This is looking worse.
Looks like something on the 5v rail is now failing.
And that 5v rail is not touching any of the parts you were working on.
?
The 4.62v on the 3.3v SMPS is especially alarming, but your DVM measurements show the 3.3v rail to be a safe 2.89v.?
I'm having a hard time figuring out why the diagnostics report 4.62, even after factoring in?
how the actual 2.89v will affect what the ADC readings since 2.89v is used as the ADC reference.
And I have no idea why the output of IC101 would be high at 3.88v.
?
The 2.22v reported on the 5.0v rail also doesn't make sense, since the low 3.3v rail
would cause the ADC to report a higher value than your 2.267v DVM reading.
?
Whacking at those power supplies is rather dangerous, at some point those supplies may decide to go
too high instead of too low and take out half the parts on the board.
I had thought I mostly understood what was going, but these new readings have me puzzled.
If you are willing to accept the loss of the board as a good educational experience,?
then continue on assuming somebody in the forum can figure out a direction to go on this.
If you care more about the board, there might come a time to consider having Jeff Moore to sort this out for you.
?
Jerry, KE7ER
?
?
?
On Sun, Jun 1, 2025 at 09:19 PM, Eugene Pressley wrote:
I decided to replace Q102 and Q101, since I have spares.? Once that was replaced I decided to tidy up the patch from pin 1 to C102.? Somehow, I created a short across C102.? This was a?+12V rail to GND short.? It took me a bit to find and correct that.? I removed Q102 and Q101 again while trying to find that short.? This short was new and not part of the original issue but it may have created new problems.? I replace Q102 and Q101 with fresh components (just in case the short caused an issue with them).? Now when I power on the display will not turn on.? When I go to the console it is in emergency diagnostic mode.
|
Brent DeWitt