Re: Help with partially fried QMX
Your diagnostics page looks fine for an 11V input.
?
On Sun, Apr 13, 2025 at 06:52 AM, Austin K4HYD wrote:
At 11v I see the device auto power on even without the control board connected. As soon as the barrel connector is plugged in the device boots up.
The cause of this is one of the following:
a) The PWR_ON signal (JP101/JP102 8-pin connector) is getting pulled low.? It should be at the Vin voltage (via pull-up resistor R104).
b) The PWR_HOLD signal (JP101/JP102 8-pin connector) is getting pulled high.? It should be at 0V (via pull-down resistor R105).
c) [less likely] - multiple component failures on the 5v smps.
You should be able to measure the PWR_ON and PWR_HOLD voltages as you turn the QMX on.? PWR_HOLD is easy to measure with a DMM.? If it is significantly below the Vin voltage, that is your problem and you need to find and fix it.? PWR_HOLD is harder to measure - you may need an oscilloscope: when the QMX processor boots, within about 1/2 second it will assert the PWR_HOLD voltage high, and you want to see the voltage on it before this.
?
Since it is worse when you have the control board attached, it suggests that the added load on these two control lines exacerbates the problem.
?
Good luck in finding it - Stan KC7XE
|
Re: QMX/QMX+ internal Transmit Voltage measurement question
On Mon, Apr 14, 2025 at 08:34 PM, Stan Dye wrote:
So I measured the voltage at ADC_PA with my DMM (at top of C529), in case it is helpful:
?- ADC_PA during diagnostic receive mode (off): 0.08V
?- ADC_PA during diagnostic transmit: 1.96V
Using just the 47k/10k voltage divider ratio, this would equate to 9.2V (rather than the full 11.22 measured)
?
Stan,
calculating from 1.96V back to the PA voltage (Q507 D) I came to a value of 11.17V, close to 11.22V you measured.
1.96V * ((10k + 47k) / 10k) = 11.17V?
?
73 Ludwig
|
QMX repair and diagnosis by W1NC
Shout out to Jeffrey Moore W1NC for the excellent service and quick turnaround on my problematic QMX build. ? He has it up and running and fixed all my build problems. ?Lessons learned. ? I can't say enough about how delighted I am. ? I didn't want to have to attempt to build a second kit to correct my mistakes. ? He had found my soldering issues plus replaced the blown CPU. ? I am looking forward to my first CW contact with 5W on this machine. ? ?
?
Bill Buck, KJ5IVB
|
28 AWG enameled wire. Available on Amazon.
What type of wire is used on the torroids? I want to rewind some but I don't know where I put the extra.
-- Gary W0CKI
|
What type of wire is used on the torroids? I want to rewind some but I don't know where I put the extra.
|
Re: QMX/QMX+ internal Transmit Voltage measurement question
Yes, I see similar behavior with my QMX+. See /g/QRPLabs/message/143062Probably a software bug and maybe the cause for SSB calibration failures. Calling out @Hans! :-) Gottfried
|
Re: First attempt at WSJT-X failed
Regarding failure to make contacts in WSTJ-X:
From your notes, it seems you have things set correctly for WSJT-X to transmit (full bar under the 'A') and it is decoding fine.
?
So a few things to check:
1) Is QMX in practice mode?? (P in the display after the frequency) CW practice mode kills TX in all modes.
2) Are you getting SWR protect? (S in the display after the frequency) - that would kill TX
3) VFO mode: make sure you are set to ONLY VFO-A or ONLY VFO-B.? Don't try split mode; it confuses WSJT-X, and you may not be transmitting on the frequency you expect.? Make sure WSJT-X Radio 'mode' is set to 'none' and 'split operation' is set to 'none'.? And manually set DIGI mode and one VFO (not split) on the QMX.
?
Stan KC7XE?
|
Re: First attempt at WSJT-X failed
On Mon, Apr 14, 2025 at 10:38 AM, Jim Brown wrote:
Should the 3.3v status also say "OK?"
No, it is fine.? When in linear mode it says 'LINEAR'.
|
QMX/QMX+ internal Transmit Voltage measurement question
Hans (and others who may know):
?
On my QMX and QMX+, I have often wondered why the measured TX voltage, as shown in the Diagnostics page, always has a significant drop (about 1.5V) from the supply voltage.? Looking at the schematic, it should have only the very small drop due to Q507.
?
So (while answering someone else's question today) I measured the voltages at Q507 and ADC_PA with my DMM and oscilloscope, on my 12V QMX+ build, all measurements done at 7.074MHz.
?
Diagnostic screen shows:
?- supply voltage 11.8V receive, 11.5V during TX
?- transmit voltage 0.4V receive, 9.8V during TX? (the unexpected drop of 1.7V in TX voltage)
?
DMM measurement shows:
?- supply voltage 11.84V receive, 11.48V during TX (measured at Q507 source terminal)
- agrees well with QMX+ supply voltage measurement
?- transmit voltage 0.43V receive, 11.22V during TX (measured at Q507 drain terminal, a drop of 0.26V)
- does NOT agree with QMX+ transmit voltage measurement
- but is as expected from Q507 drop
?
So it appears that measurement of TX voltage on QMX/QMX+ is not well calibrated.
Could this be causing some of the problems seen in SSB calibration, etc?
?
So I measured the voltage at ADC_PA with my DMM (at top of C529), in case it is helpful:
?- ADC_PA during diagnostic receive mode (off): 0.08V
?- ADC_PA during diagnostic transmit: 1.96V
Using just the 47k/10k voltage divider ratio, this would equate to 9.2V (rather than the full 11.22 measured); hmmm, is this part of the problem? Or is there an expected additional resistance in the STM32 input that completes the voltage divider?
?
Stan KC7XE
|
Right after posting this I was informed that the TX voltage is NOT the actual voltage supplied to the PA. In fact, full supply voltage is supplied to the PA so, while I am correctly showing the TX voltage (in parentheses), 10.5 volts is too much for my 9 volt
build.
Tony
AD0VC
toggle quoted message
Show quoted text
Noting that 9V supply only gives 7.5 volts TX voltage, I upped the supply to 10.5 which gives more like 8.7 volts for the TX? voltage. I then added another pair of TN0110.
Tony
AD0VC
Here is mine, measured with Boonton 92BD and 10 watt 20db attenuator:
Tony
AD0VC
This is the sort of power I get:
I'm measuring the power in 3 ways, PWR is what my QMX+ reports, PWR(pk) uses the peak voltage from my dummy load and PWR(pp) is the peak to peak voltage reported by my DSO2D10 oscilloscope. They are about what I would expect, although it's a puzzle why
the 12m and 17m powers are so much lower, it doesn't seem to be anything obvious such as being off the edge of the LPF.
This is all using the standard BS170s, no messing around (so far).? It's a 12V QMX+, running on 12V.
?
|
Thanks Stan. I just confirmed that with my 9V build. I assumed... oops.
Anyway, I will just dial it back to 9 volts.
Tony
toggle quoted message
Show quoted text
Interesting question, Tony.? I happened to have my open QMX+ on the bench along with a DMM and oscilloscope, so I measured the voltages.? My build is for 12V.
?
The diagnostic screen shows supply voltage at 11.8V.
My DMM measures 11.84V at the source of Q507 (about the same as supply voltage, as expected).?
When I do 'T' to transmit in the diagnostic screen, the supply voltage displayed goes down to 11.5V, and the measurement at the source of Q507 is 11.48V (very close, as expected)
?
Measuring at the drain of Q507, which is the TX voltage measured by 'ADC_PA' and displayed by QMX+:
? diagnostic rx mode: 0.4V displayed, 0.43 measured
? diagnostic tx mode: 9.8 displayed, 11.22V measured (just 0.26V drop from V+)
?
So the Tx voltage displayed by my QMX+ is a full 1.4V lower than what is measured by a DMM.?
A measurement with my oscilloscope verifies the measurement, and shows a clean DC signal.
?
Apparently, the QMX Tx voltage measurement via (ADC_PA) is not well calibrated, and gets off by 1.4V on the high end.
?
So in answer to your question, can you run your QMX+ on a higher supply voltage, the answer is: NO.
The real TX voltage is close to the supply voltage.
?
Why?? Seems like a question for Hans, I'll post separately.
?
Stan KC7XE
?
|
Interesting question, Tony.? I happened to have my open QMX+ on the bench along with a DMM and oscilloscope, so I measured the voltages.? My build is for 12V.
?
The diagnostic screen shows supply voltage at 11.8V.
My DMM measures 11.84V at the source of Q507 (about the same as supply voltage, as expected).?
When I do 'T' to transmit in the diagnostic screen, the supply voltage displayed goes down to 11.5V, and the measurement at the source of Q507 is 11.48V (very close, as expected)
?
Measuring at the drain of Q507, which is the TX voltage measured by 'ADC_PA' and displayed by QMX+:
? diagnostic rx mode: 0.4V displayed, 0.43 measured
? diagnostic tx mode: 9.8 displayed, 11.22V measured (just 0.26V drop from V+)
?
So the Tx voltage displayed by my QMX+ is a full 1.4V lower than what is measured by a DMM.?
A measurement with my oscilloscope verifies the measurement, and shows a clean DC signal.
?
Apparently, the QMX Tx voltage measurement via (ADC_PA) is not well calibrated, and gets off by 1.4V on the high end.
?
So in answer to your question, can you run your QMX+ on a higher supply voltage, the answer is: NO.
The real TX voltage is close to the supply voltage.
?
Why?? Seems like a question for Hans, I'll post separately.
?
Stan KC7XE
?
|
Noting that 9V supply only gives 7.5 volts TX voltage, I upped the supply to 10.5 which gives more like 8.7 volts for the TX? voltage. I then added another pair of TN0110.
Tony
AD0VC
toggle quoted message
Show quoted text
Here is mine, measured with Boonton 92BD and 10 watt 20db attenuator:
Tony
AD0VC
This is the sort of power I get:
I'm measuring the power in 3 ways, PWR is what my QMX+ reports, PWR(pk) uses the peak voltage from my dummy load and PWR(pp) is the peak to peak voltage reported by my DSO2D10 oscilloscope. They are about what I would expect, although it's a puzzle why
the 12m and 17m powers are so much lower, it doesn't seem to be anything obvious such as being off the edge of the LPF.
This is all using the standard BS170s, no messing around (so far).? It's a 12V QMX+, running on 12V.
?
|
First attempt at WSJT-X failed
QMX
1_00_027 firmware
?
WSJT-X would not work with the choice of the QMX as the transceiver. Kenwood 480 did work.
?
CAT button works.
?
I did select the QMX sound input and output in Ubuntu on my Dell.
?
WSJT-X was set to full power as was the Ubuntu sound output.
?
On TX there was a full bar under the A in the QMX display.
?
The QMX was set to digital mode.
?
I called CQ on 20 m, which was active at the time and also tried to answer CQs to no avail. I was using my EFHW in my shack. I have worked 131 countries with it and a Icom 7300.
?
I thought maybe I might have blown my finals in some way.
?
I connected to the QMX with Putty and went to the Hardware Test section.
?
With supply voltage of 11.3V I tested every band by pressing the T key on my keyboard for TX and have the following results. A dummy load was attached.
?
15m? ? ?5.2w
17m? ? ?4.2w
20m? ? ? 4.7w
30m? ? ? 4.1w
40m? ? ? 4.4w
60m? ? ? ?5.1w
?
Since I have output on all bands with this test, I assume I didn't blow the finals. The QMX seems okay and I should just keep trying with WSJT-X.
?
?
I do have a question about the display above. The status of the 3.3v SMPS says "SMPS" while the status of the 5v SMPS says "OK."
Should the 3.3v status also say "OK?"
|
Yeah, there's a button to update your hamlib dll from inside WSJT-X; but that's /also/ broken (it's a theme!) unless you have a very specific openssl dll installed that's no longer easy to find.? Whenever I'm setting up a WSJT install now installing WSJT-X (usually the most recent version for me) and then replacing the hamlib dll with its latest version has been my recipe for success.
73 KD7SKY
On Monday, April 14th, 2025 at 7:31 AM, james jones <j2@...> wrote:
Didn’t know about the hamlib update; thot it was always supplied as part of WSJT-X. ?
Fairly sure I downloaded v2.6 from Princeton.edu not that long ago. ?
FWIW: QMX+ receiver works extremely well on FT-8 on 80, 30 and 20 even though another receiver can barely hear the signal.
Thanks for the suggestions!
72/73
Jim w0nkn
On Apr 14, 2025, at 2:40?AM, Muhsin TA1MHS via groups.io <muhsin.dogrular@...> wrote:
?
Hi?
I had similar problems
cure:
1- update WSJT to 2.7 (no r)
2- install Hamlib
4.6.2
3- copy libhamlib-4.dll from?
..\hamlib-w64-4.6.2/bin (somewhere in your c:\program Files)
to
..\wsjtx/bin (it must be in somewhere on your c:\ drive)
overwrite it...
4- select Rig: QRPLabs QMX
5- send a prayer to Tesla
73 DE muhsin TA1MHS
?
|
Hi Stan,
Thanks for taking these measurements. Your 180 MHz scope is obviously rolling off the 50 MHz harmonics above the third but the waveform at the XOR gate output is as clean as can be expected.
The 50 MHz waveform distortion at the FET gates is interesting. I wonder if the sinusoidal response at 0V is due to a combination of the BS170 gate capacitance and the 74ACT08 frequency limitation or if one or the other device is causing this. The only way to know would be to disconnect one leg of the ACT08 and re-measure. It’s possible that this device is better at sourcing current than sinking current.
When JZ and I were simulating the PA circuit way back when, I saw a large current flow exceeding 1A from the ACT08 due to the Miller effect during FET turn-on. I wondered if the two paralleled AND gates could really supply this much current so I added a second pair of AND gates to the simulation to model what Tony is doing and the waveforms did improve a bit. I concluded that the ACT08 driver was one limiting factor at 50+ MHz. Short of using a true gate driver such as OnSemi’s NCP81074, I didn’t find a better way to mitigate the Miller effect.
Tony AC9QY
On Sun, Apr 13, 2025 at 05:54 PM, Tony Scaminaci wrote:
have either of you looked at the output waveforms of the 74ACT08 at 6m?
I just happened to have an open QMX+ and an oscilloscope in front of me when I read this, so I did some quick measurements.
The oscilloscope is inexpensive (FNIRSI DPOX180H), so keep that in mind - but it should be reasonably good at 50MHz.
Using a 10x probe, I see the following at the driver (IC503) outputs.
Measurements are at 3MHz, 14MHz, and 50MHz for comparison.
Then there is a measurement of the 3.3V IC502 output (XOR gate pre-driver) at 50MHz, so you can see what the scope does at that frequency on a signal with (presumably) less loading.
?
So yes, according to these measurements, at 50MHz the driver does give a full-rail 5V signal to the BS170s, but it is no longer a clean square wave, starting to look more sinusoidal on the bottom half, due to the undershoot rebound time.?? I'm not sure how the BS170s will internally see the low-going gate signal when it is below ground potential - maybe it is still treated as 0V, cleaning up the square wave.? And these measurements should be repeated on a higher-quality scope for verification that this isn't just scope/probe measurement degradation.
|
My apologies for sending an unfinished email.
?
On Sun, Apr 13, 2025 at 10:42 AM, dl8lrz Reiner wrote:
The best results were achieved with 2 additional BS170.?(A BS170 can switch about 380-550mA, depending on the manufacturer; with more, the residual drain voltage (and heat) increases significantly. Disadvantage: The higher capacitances can detune the LPF, and this non-linearly. The 20-17-15m filter (and higher) in particular can cause a lot of trouble afterward.
Reiner,
?
Yes, the easiast way I found so far is to use 2 x 3 BS170's instead of 2 x 2.
?
I carefully checked the LPF sweeps on all bands before and after mounting the extra 2 BS170's.
No change is obvious from 160 through 40M and on the higher bands the changes are small.
Some bands look even a bit better and no change is needed at all.
?
My build is a rev1 board set up for 9V, using a WTST on a BN-43-302 core.
The beauty of a 9V build is that you can test is more safely with a higher supply voltage.
?
Going for maximum smoke, 12V deliveres this:
?
160M - 8.8W, 80M -7,7W, 60M - 6.5W, 40M - 8.0W, 30M - 7.0W, 20M - 6.5W, 17M - 4.4W, 15M - 6.5W, 12M - 4.8W, 10M - 6.1W, 6M- 4.2W
?
What puzzles me is the following: using 10V, 2 x 2 BS170's deliver 2.8W on 15M.? This jumps to 5.8W when using 2 x 3 fets.
This seems counter-intuitive as more transistors have a higher input capacitance and hence needs more drive.
?
It is also interesting that 20M, 15M and 10M produce more output than 17M and 12M.
I gather this is caused by ripple in the passband of the LPF's.
?
73,
Roelof, pa0rdt
?
?
|
Re: QMX and power bank PD
Yes, all good now.
And that sounds like a fine POTA kit.
Have fun!
Stan KC7XE?
|
Re: QMX+ with 100 amp/hour lithium iron battery
Let's not get too carried away with special $150 fuses and such, that only discourages the use of these batteries.
Large high current batteries should be treated with respect, but are very useful for the amateur.
Used properly and with some understanding of what's going on, they are not dangerous.
Far less dangerous than the 2000 kV plate supplies of an old school transmitter.
?
My advice for powering a QMX is to just use something like16 AWG wire, not 0000 welding cable.
16 AWG is 0.004 ohms/ft, so 6 ft of wire round trip is 0.024 Ohms, at 1A that's only 0.024 of voltage drop.
A 2A automotive fuse in the power supply wires is fine, and the wire will go if the fuse does not for some reason.
Try not to drop a screwdriver across the battery terminals.
?
As an experiment, I donned my safety goggles and shorted out our 30kwh 48v off-grid battery bank
using two feet of 6 AWG copper wire and a 7.5A ATO automotive fuse, curious if the fuse would arc over.
It did not, just a nice pop and flash of light inside the fuse.
I am not confident that this would always be the case for a 48v supply, for example this may depend
on how large a gap gets established by the initial current surge, or who made the fuse.
I am confident that such an automotive fuse is adequate for a 12v supply.
After all, a lead-acid starting battery can also deliver several thousand amps when shorted.
?
Here's a quick video that drives home why we care about high voltage DC arcs.
https://www.youtube.com/watch?v=Zez2r1RPpWY
The AC arc self quenches whenever the sine wave crosses zero volts, at 60Hz*2 = 120 times per second.
The DC arc has no such quenching action, the ionized air in the arc remains conducting till the arc is quite long.
In this case it's 240 Volts DC, not 12 Volts.
?
LiFePO4 is a very good choice.
Safer than most other lithium chemistries, and has reasonably good power density.
A 100AH battery is good for half a week with key down at 5W out on a QMX.
Excellent for emergency backup communications.
?
Jerry, KE7ER
?
?
On Mon, Apr 14, 2025 at 07:52 AM, Bruce Akhurst wrote:
?
On Thu, Apr 10, 2025 at 02:52 PM, Bruce Akhurst wrote:
Hope you’ve got your fault protection sorted.?
Prospective short circuit will be 10kA and an auto fuse ain’t going to stop that arcing ..
It can be as simple as a fusible link.? Go down 4 wire size numbers, so for example #16AWG link in #12AWG conductor, a foot long, connected in series with the main 12AWG line with compression crimp splices NOT solder.?
The link including splices and at least an inch beyond (I cover about 2-3 inches), is covered with fire sleeve.?
I use an 18 inch long sleeve for a 12 inch link; the idea is to contain and cool the hot gases if the link fuses so leave room
Exactly.? ?I see a lot of DIY/ RV stuff where 100's of Ah of lithium batteries are merrily connected to quite low consumption projects in inflammable locations without properly thinking through the prospective fault/short circuit currents and consequences.? ?
Large current DC arcs are hot and self sustaining and something you only want to see once and not in your workshop /office /Van/ home.?
In theory the better batteries may have BMS or other internal protection but it's hard to test that without damaging the battery and it may not be there or not reliable anyway.? ?
T grade fuses are $150 for a high interrupting current fuse+holder so cheaper alternatives are useful.? ? ? ?
Also don't ignore marine grade Switch/Breakers which are resettable, re-usable? and decent ones by Buss etc (no not ebay clone/copy ones) can typically carry and switch 15 - 100A or more as circuit breakers and interrupt 6000 amps in fault conditions for less money than a T fuse . .??
Note a 'battery switch' or 'breaker' or 'resettable fuse' while cheaper,? are not necessarily able to interrupt large DC fault currents and this figure usually missing from their specs?
Bruce
|
Re: QMX and power bank PD
Thanks for the feedback Stan. I double checked this and it's my fault. I misread the display. I was looking at the output in Watts, not the Voltage reading!
?
I checked again today. Same setup with the dummy load and tapped away on my paddle. The Voltage was flitting around 12.04v to 12.16v as I tapped away at the key. When idle it just sat at 12.04v.
?
I removed the QMX and left the power bank, voltage meter and the USB-C cable connected and it still sat at 12.04v with no QMX connected to it. I'm assuming that's the PD board in the USB-C cable telling the power bank to pull 12v as that's the cable I specced when I bought it instead of the 20v or 15v one?
?
With all of this in mind, I'm assuming my QMX will be okay and safe to use out and about? What a game-changer this will be compared to lugging my FT-891 around for POTA and SOTA! My kit list will be QMX, Sony Earbuds or KOSS PortaPro, Chameleon LEFS 8010 - plus optional tree! (or my favourite, the JPC PAC-12 - no need for the optional tree hihi) . Notepad and Pencil, 50m Coax, Throw line and weight in a Brabantia washing peg bag.
?
73
?
?
|
