|
Re: My NanoVNA is overheating
The results are in, it is the si5351.
On the first attached sweep the chip was not cooled, the other sweeps were made while holding a frozen cotton swab against the chip for increasing amounts of time. I started out using a frozen blueberry, but that only worked for so long, eventually the blueberry juice ran down onto the traces and NanaoVNA-Saver crashed. So I ate the rest of the frozen blueberries. I wonder what my options are, besides buying a new H4. Gary AG7TH |
|
Re: VBAT resistor voltage divider, polynomial fitting, VBAT sampling
On 2020-07-16 16:19:-0700, you wrote:
This may be why the VBAT is now enabled all the time...that would mean no diode startup time. That would mean a steady state current of maybe 30 to 36 uA over the battery discharge cycle. I wonder if this is correct. It seems to be, because the VBAT voltage divider is a constant 100K ohms. The question is what the voltage drop across the 1N4148WS (if that is what is used) is at 40 uA. The smaller the voltage drop across the diode, the larger the drop across the 100K network. Somewhere there is a balance. I have not seen the design specification for this part of the board, so I don't know what the designers calculated for the circuit.Resistors won't have that instability.In this wiring the diode does not represent a source of "instability", imho, except its -2mV/K TC. I still opine that I am not interested in more than a rough estimate of the battery charge state. I am more interested in the number of CPU cycles and the amount of memory used by a routine that takes 25 averages of the voltage. I actually don't know if that is correct, either. I was told that months ago, but have not seen a requirements doc or design doc that explains it, and I have not taken the time to troll the code trying to find out. I understand that the code is under more or less constant revision in each of the branches, with little in the way of specification, review, or V&V. -- 72/73 de Rich NE1EE On the banks of the Piscataqua |
|
Re: My NanoVNA is overheating
The ?T is 18¡ãC; 13¡ãC; and 15¡ãC. Some of the cold spots are actually on the circuit board. The two closest to the IC packages may be holes for mounting. I would not consider the holes significant. The large square cold spot on the right of all three is probably not significant either. The high temperature ranging from the Human Body Temperature +9¡ãC or +13¡ãF is a different issue.
toggle quoted message
Show quoted text
My imager is due back tomorrow from it¡¯s annual Calibration Trip to the Factory. I will take a couple pictures of mine for control. See what a more normal temperature may be. Are those images from the front or back of the unit. If they are through the LCD display, I would not trust the accuracy. John Nicholas Certified Thermographer Level II On Jul 16, 2020, at 6:19 PM, Gary Hale <gary@...> wrote: |
|
Re: My NanoVNA is overheating
Hot pictures! Looks like all three chips are candidates for a heat sink.
toggle quoted message
Show quoted text
I installed the old 0.8.4.5 firmware, then I put the NanoVNA in the freezer for a few minutes to see if I could make it through a calibration, looks okay so far after calibrating. Need to wait until the device warms up again to see if it stays stable. After a few minutes I am now seeing the same thing, instability and noise all over. So it may be something that the firmware cannot fix. Gary AG7TH On 7/16/20 15:52, DiSlord wrote:
Made thermal photo NanoVNA |
|
Re: VBAT resistor voltage divider, polynomial fitting, VBAT sampling
On 2020-07-16 08:05:-0700, you wrote:
FYI - In the current source code the internal on-chip vbat divider is always enabled. <SNIP>Well, if it is always on, then the circuit draw 4V / 100K ohms = 40 e-6A = 40 uA. If the battery is a 400 mAh battery, then it will take 10 000 hours to drain the battery. Is that a concern? With the resistor voltage divider in place of the diode, the numbers are roughly the same, so battery drain there is not a problem either. Feel free to correct my estimates...I have been known to miss a decimal point or two when I do this in my head. -- 72/73 de Rich NE1EE On the banks of the Piscataqua |
|
Re: VBAT resistor voltage divider, polynomial fitting, VBAT sampling
On 2020-07-16 06:56:-0700, you wrote:
The Vbat, at least what I see in the sources for H4, is being sampled such the Vbat voltage divided by 2 (internal on-chip divider, most probably disconnected when not in use) is converted via the ADC channel 17. There are no external dividers needed.Skipping over "most probably" as an excellent example of what was not shared in the early days... We come to "There are no external dividers needed." But...again, according to the early design...the batteries peaked at 4.2+V, and the VBAT signal is spec'd at a max of 3.6V. It can go slightly higher, but not by design spec. So the original designer included a diode in the path to VBAT. The idea was that there would be a roughly 0.6V drop across the diode, and that would drop the 4.2 to 3.6. So the diode was taking the role of a voltage divider. My concerns were a) the diode was operating in an unstable region, due to the low current through it. Why the low current? Because the VBAT sampling circuit used a 100K resistor to sample across. (Actually 2 50K, but the diode sees 100K). b) if the diode is not allowed to settle, then its voltage drop is not predictable. In higher current circuits, using a diode in this fashion is fine. c) the clue to the design flaws is in the fact (as I understood it) that VBAT was sample 20+ times, and the results averaged. d) the whole reporting of the voltage to 4 significant figures is pointless. All we really care about it is estimated battery charge to the nearest 20%, right? The rather unusually high noise in the vbat data readings I presented in my other post could be caused by something missing in the H4 code, ie. a small delay should be inserted after activating the divider - before reading the vbat, or something like that..Well...exactly the original point I made...and I am not saying that I am "correct", only that I surfaced a reasonable question, and was snubbed by people who said that they had not read the manual. It's pretty standard to take into account all the timing delays in an ADC circuit. None of them are instantaneous. And this problem is worse if my suspicion about the unstable diode response is valid. Resistors won't have that instability. These are just discussion points...not everyone will see the situation in the same light...or need...after all, we have some estimate of battery voltage now, right? -- 72/73 de Rich NE1EE On the banks of the Piscataqua |
|
Re: Protective Case for H4
That is correct.
toggle quoted message
Show quoted text
Darrell Sent from my over-rated IPhone 7 Plus. Any Mis-spellings or grammar errors are due to my IPhone auto correct feature. On Jul 16, 2020, at 15:52, David Webb <kb1pvh@...> wrote: |
|
My NanoVNA is overheating
NanoVNA-H V3.4
DiSlord 0.9.3.4 firmware I performed and open calibration only with no cables. The temperature in the room is 80 degrees Fahrenheit. First picture was taken when device was turned on. Second picture was taken 6 minutes after turning on. Third picture was taken 15 minutes after turning on. If I put a bag of frozen vegetables on top of the device it works great, it is perfectly stable. If I issue a threshold 200000000 command it overheats more slowly. I bought the device in February and it worked well in the cooler months. Thinking about adding a heat sink somewhere, but not sure where. Any ideas? -- Gary AG7TH |
|
Re: Protective Case for H4
No 3-D printer here. So I modified the very nice box my H-4 was shipped in. I provided some large open cell rigid packing material to hold items in place. The small bottle holds the calibration pieces. This is secure and quick, cosmetic touches lacking.
toggle quoted message
Show quoted text
Took about an hour to complete. KEoZUW John On Jul 16, 2020, at 3:52 PM, David Webb <kb1pvh@...> wrote: |
Rich NE1EE
John Nicholas
Darrell Carothers
