Keyboard Shortcuts
ctrl + shift + ? :
Show all keyboard shortcuts
ctrl + g :
Navigate to a group
ctrl + shift + f :
Find
ctrl + / :
Quick actions
esc to dismiss
Likes
Search
Firmware and Wiki question
#firmware
|
I am experiencing something that makes me wonder if the information in the Wiki reflects the current firmware.
Background: When I start up, measuring on a good meter, I get 7.55 V input 5.0? V Vcc 3.01 V Vdd The Vdd is too low, obviously, and I have been looking into this.? I have bought a new set of SMPSs, but they apparently have not yet been shipped, so I am still looking at my existing supplies to see if I can find the problem.?? The other way to look at this is the measurements from the microcontroller. The 3.3 volts is still running linear, certainly because the power supply voltages are off.? If the Vdd volts is being used as a reference voltage (likely), then the measured voltages would be off by? a factor of 3.3/3.1? The supply voltage of 8.1 V is close to (7.55 * 3.3 / 3.1), which is 8.04V.? Similarly, (5V * 3.3/3.1), or 5.32 is close enough to 5.47 to be believable.? I am therefore poking around the 3.3V linear supply.? The output of the 78M33 is 3.9V, but the voltage at the zener is 3.03V, so the diode on the back side of the board is dropping? 0.87 V, vs about .6 volts expected.? I? My question: I keep going back to the Wiki, which says: What I'm saying ladies and gentlemen is that if you are able to log in to your QMX using PuTTY then the program has already gone through its entire initialization successfully. The buck converters have both been powered up and are in use producing the 3.3V and 5V rails. It's already gone past initialization of the ADC/DAC chips which also require the 5V rail
I get to the PuTTY page, but it seems to be limited to the hardware diagnostic page. I keep coming back to this statement in the wiki and wondering if a newer version of the firmware has changed this, or if there is something that I am not understanding.? Can someone who has direct knowledge of the firmware comment?? I imagine that is Hans, but I would appreciate input from anyone who has direct knowledge. Bonus questions: - What is the criteria for turning on the SMPS and turning off the linear power supply?? ?Specific voltages please. - How is the Voltage in the 3V3 SMPS section of the diagnostic page calculated.? My guess is that it is some factor multiplied by ADC_3V3.? What is the factor? - How the heck am I getting a solid 5V on Vcc if the power supplies? are in linear mode?? There is no linear 5V regulator, so I must be on switching mode, and the PWM must be set to a good value to deliver 5.0V, but the diagnostic page is showing that the measured voltage is 5.47 volts.? The controller must know something that it is not displaying on this page.? Maybe?? Thanks, Mark W8GU |
|
I think that Vdd being low when running on the linear is a feature of the linear circuit.
The ground of IC1 is connected to ground through the diode D102 and this lifts the output by the diode drop across D102.? If isn't taking much current so maybe 0.6V giving an output from IC101 of 3.9V.? This is reduced by the diode drop across D103, however D103 has a much higher current through it, maybe 100mA so will have a higher voltage drop, maybe 0.8V which gives a VDD of 3.1V. Chris, G5CTH |
|
On Tue, Jul 2, 2024 at 11:23 AM, Chris wrote:
3.9V.? This is reduced by the diode drop across D103, however D103 has a much higher current through it, maybe 100mA so will have a higher voltage drop, maybe 0.8V which gives a VDD of 3.1V.Mark, I think Chris is right. So VDD = 3.01V is consistent and enough to run the controller. then the measured voltages would be off by ?a factor of 3.3/3.1I think, you are also right calculating this factor. But the calculated Duty max don't fit to 8.1 V shown supply voltage. Testing my QMX (FW 018) I get the shown Duty max from 7.4 V. (Afaik Duty max is calculated from the supply voltage.) 7.4 V is close to your real input. So it looks like the controller know something more. (Later I will check something.) The 3V3 voltage shown by Diagnostics should be the one from the output of the SMPS. I guess it's calculated from the signal ADC_3V3. In state of "Linear" (LIN_REG_EN = HIGH) the SMPS output is separated from VDD by Q111. The SMPS is working toward the artificial load R114 (47 ohm). Following Diagnostics the 3V3 Duty cycle 49% is equal to Duty max. This means you have 3V3 SMPS overload with the result of only VDD = 2.54V. The question is the place of this overload. If Q111 is working well the overload should exist on the SMPS board. And I guess Q111 is working well because the VDD rail has 3.01V. I would test the SMPS board using the off line method (see the link in the QRP Labs Wiki document QMX and QMX+ Internal power supply. 73. Ludwig |
|
So, based on responses from Ludwig and Chris, I am hearing that the wiki "QMX_QMX_plus_Startup" is incorrect.? It appears that this is a true wiki, and I can edit it.??
Would anyone be upset if I made the following change? What I'm saying ladies and gentlemen is that if you are able to log in to your QMX using PuTTY (and you can get to pages other than the diagnostic page) then the program has already gone through its entire initialization successfully. The buck converters have both been powered up and are in use producing the 3.3V and 5V rails. It's already gone past initialization of the ADC/DAC chips which also require the 5V rail.? |
|
In addition, the "QMX PSU test process.txt" document that is referred to in the "QMX QMX_plus internal power supply" wiki should be changed.? It says:
PCB #2 1 ?Connect a 5.6V, 100mA max power supply to the GND and +12V pins. ? ? 1.1 ?VDD should be 3.3V, supplied by the linear regulator IC101. Current should be around 50mA (45 to 55) and supplying current to VDD should be possible. The SMPS section should be inactive. (The current is taken by R114; Q111 is off because LIN_REG_EN is high but it will still conduct through its parasitic diode.) I suggest PCB #2 1 ?Connect a 5.6V, 100mA max power supply to the GND and +12V pins. ? ? 1.1 ?VDD should be 3.3V, but might be as low as 3.0V, supplied by the linear regulator IC101. The linear regulator is only used during initial startup and firmware update, so a full 3.3 volts is not necessary.? ?Current should be around 50mA (45 to 55) and supplying current to VDD should be possible. The SMPS section should be inactive. (The current is taken by R114; Q111 is off because LIN_REG_EN is high but it will still conduct through its parasitic diode.) This edit, and the one I just sent about the startup wiki would really have kept me from chasing a non-issue. |
|
On Tue, Jul 2, 2024 at 06:33 AM, @Ludwig_DH8WN wrote:
The 3V3 voltage shown by Diagnostics should be the one from the output of the SMPS. I guess it's calculated from the signal ADC_3V3. In state of "Linear" (LIN_REG_EN = HIGH) the SMPS output is separated from VDD by Q111. The SMPS is working toward the artificial load R114 (47 ohm).OK.? This is something that I did not understand.? I assumed that the /LIN_REG_EN (I always? use / to indicate an active-low signal) was turned on during the startup test and then turned back off when it failed.? My understanding is probably wrong. One other thought.? The processor is apparently calculating the max duty cycle based on the input voltage.? It thinks the input voltage is 8.1V, but it is wrong because the linear supply is only putting out 3.01V.? The input voltage is only 7.55 volts.? I wonder if 49% is high enough for an input voltage of 7.55V. Looking in the Operation manual page 84, the diagnostic page of a known good device shows Operating at 12 volts, the max duty cycle is only 2% above the current duty cycle, so the calculated max is rather restrictive.? If the processor believes that the 3.01V reference voltage is really 3.3V, it may be setting the Duty max too low. There is no way to tell the processor that the linear supply voltage is too low, though. Which brings me back to wondering if 3.01V out of the linear supply might be just too low. If I could change the firmware, I would relax the max duty cycle while /LIN_REG_EN was active to account for the fact that the reference voltage might be too low.? The main board is protected from this voltage by Q111 if it gets too high. If only the processor knew the 3.3 reference voltage was too low. Which brings me back to bonus question 3 from my original post: - How the heck am I getting a solid 5V on Vcc if the power supplies? are in linear mode?? There is no linear 5V regulator, so I must be on switching mode, and the PWM must be set to a good value to deliver 5.0V, but the diagnostic page is showing that the measured voltage is 5.47 volts.? The controller must know something that it is not displaying on this page.? Maybe?? |
|
On Tue, Jul 2, 2024 at 02:42 PM, <w8gu@...> wrote:
Which brings me back to wondering if 3.01V out of the linear supply might be just too low.If the controller talks with the terminal emulation and shows values that means the controller is working and the voltage should be enough. For my working QMX when the linear regulator is active (Update firmware) I see VDD = 3.07V. Using my working QMX I get these values: voltage left?????????????????????????? 3V3 Duty max top at Diagnostics ???? 8.1V??????????????????????????????????? 45% ???? 7.4V??????????????????????????????????? 49% In your first post 8.1V was shown together with 49%. So the controller calculated the value fitting to 7.4V. It looks like the controller was knowing the voltage of 7.4V but showing 8.1V. This is related to your bonus question 3. I assumed that the /LIN_REG_EN (I always ?use / to indicate an active-low signal) was turned on during the startup test and then turned back off when it failed.I don't see a "/" before LIN_REG_EN but maybe I was looking at the wrong place. Regarding schematics page 1 LIN_REG_EN is active when HIGH (Q101, Q102 and Q110 are ON and Q111 is OFF). It would be interesting to know the supply current for off line test of PCB #2 Step 3.1. You may find the 3V3 overload. 73 Ludwig |
|
On Tue, Jul 2, 2024 at 01:37 PM, <w8gu@...> wrote:
Would anyone be upset if I made the following change?The Text is a citation from a post by Hans. So your insertation should be marked as a text from you and not by Hans. 73, Ludwig |
|
On Tue, Jul 2, 2024 at 05:42 AM, <w8gu@...> wrote:
OK.? This is something that I did not understand.? I assumed that the /LIN_REG_EN (I always? use / to indicate an active-low signal) was turned on during the startup test and then turned back off when it failed.? My understanding is probably wrong.LIN_REG_EN is active high, not active low. At start up it is pulled high by R101, the processor pulls it low to disable the linear regulator and switch to the SMPS. One other thought.? The processor is apparently calculating the max duty cycle based on the input voltage.? It thinks the input voltage is 8.1V, but it is wrong because the linear supply is only putting out 3.01V.? The input voltage is only 7.55 volts.? I wonder if 49% is high enough for an input voltage of 7.55V. This will be correct while the linear regulator is enabled. Once it is disabled and the SMPS enabled VDD will be whatever the SMPS supplies.? However if the Vref for the ADC is VDD then the value of VDD measured by the processor will always be 3.3V regardless of what it actually is. What will happen is that ADC_3V3, which is used to measure VDD, will always be 50% VDD because it is defined by the voltage divider R110 and R112, both 22k. The processor will read ADC_3V3 using VDD as the reference, then double it to get VDD. |
|
Thanks for asking, Jeffrey,
The power supply problem has been fixed.? See /g/QRPLabs/topic/107032579#msg125295.? It was a broken wire on the inductor.? Most likely it was taken out by those nylon nuts.? I am considering other ways to hold that board down.? May try putting the screw in upside down (nut on bottom).? For now, I have a flying AWG26 wire going up to make contact with the inductor wire.? It was difficult because the broken wire was the one that led from the inside of the core, so I could not unwind it. Now I have this problem. ?/g/QRPLabs/topic/106620596#msg125481 "Wall to wall noise".? Replacing the ADC. Actually sitting in front of the board with a hot soldering iron and some ChipQuik right now.? Wish me luck! Mark W8GU |
to navigate to use esc to dismiss