开云体育

Stojan, but missing some important information. Could you show please screen shots?

?

73 Ludwig

An internal short in the transformer would do bad things and?

would likely be sensitive to temperature .

?

It is also a place where its easy to cause said short during

construction.? ?using a small drill bit or abrasive tool to take

off rough edges from the holes in the core (in the manual!)

and even some nail polish painted on to further smooth

those edges works too ( I do that).??

?

Should power reduce?? ?No, its not normal.

?

All measurements say some reduction (milliwatts)

as in .1W or less was seen when the BS170s and driving

74ACT were hot.? A loss of .1W (or less) at 4-5 Watts is

insignificant.

?

--

Allison
------------------
Post online only,?
direct email will go to a bit bucket.

On Sun, Apr 13, 2025 at 12:33 AM, Stan Dye wrote:

I also found that in Windows, the audio output level created by Winlink's VARA-HF modem does not seem to be full scale, so the output power is somewhat lower.?? It would be nice if Hans could add a level adjustment to the USB sound interface, so we could compensate.

I agree with this, it may only be a small thing, but having a little less than 3W output on Vara is great for the finals, but it would be nice if we could adjust (boost/cut) the level of the audio coming in on the USB somewhat. To be clear, it's not a massive deal, but it would be a nice to have.

QMX+, Rev1 board, 1-00-027 Program

I checked the advanced section and all 3 items are set to Enable.

I still have a horrible Clunk when keying the QMX+.?

No clunk when not in transmit mode. So I think its when the PA gets switched on??

Please advise!

Evan,

?

The? saturated resistance of a mosfet switch goes up with temp...true

The yabut is that the gate threshold goes down with heating.

The latter means for a fixed drive the mosfet switches on harder

(lower resistance with higher drive).

?

Its why IRF510 (and others) in power amps show increased idle

current when hot unless there is circuit to lower gate bias

based on heating.

?

Likely cause is somewhere something is not delivering enough

voltage to the fets or coupling power from them.

?

If its BS170 heating, a finger on the heatsink would make that

obvious.

?

--

Allison
------------------
Post online only,?
direct email will go to a bit bucket.

Just to close this thread out, I had to go on the road for a couple of weeks and was only able to get back work on this problem this week. I removed T1, ?rewound it with fresh wire, and reinstalled it. The QMX is now putting out 4 to 4.5 watts into a dummy load on all the bands with no dropping of power, so that fixed the problem. I did not try to count the turns as I unwound it to see if I had miss-wound it, but I did see where some of the enamel installation had been scrapped off where I had a couple of false starts winding it the first time. I suspect that was the issue.?

?

Thanks for pointing me in the right direction.?

?

73 Doug WB5WRC

Sorry for all the questions, I thought I had everything working, but now when I open the measurements screen and press the button, nothing happens. It will do an SWR scan with the Serial interface but not just show when you press tune.

On Thu, Apr 17, 2025 at 03:44 PM, OE1IRA wrote:

ah, I almost forgot: I had a long fight with the 20/17/15m-filter for a 4xTN0110 modification.
You could try it with your 6xBS170 mod?
See /g/QRPLabs/files/QMXplus_Hardware_Mods_V1.0.pdf

Hello Gottfried,
?
Thank you for your help and I did not see your paper until now.
?
As tinkering on this PCB is not so easy, I opted for simulating the 20 / 17 / 15M LPF with RFsim99.
?
A source impedance of 50 ohm produces a graph that is virtually flat, confirming a sound design.
?
With a source impedance of 125 ohm and changing C532 from 330 to 430 pF, the filter graph rises gradually from 20M to 15M, which is not what we need.
Changing the source impedance to 20 ohm and using also 430 PF for C532, insertion loss on 20 and 15M is very similar and a bit more on 17M.
?
In practice, adding a 100 pF SMD 1206 capacitor at the bottom of the board parallel to C532, produced 5.3 W on 20, 4.3 W on 17 and 4.6W on 15M.
Apparently, 100 pF is a bit too much as the passband shifts down a little too far.
?
82 pF extra did the trick and the internal LPF sweeps looks now much flatter.
This implies that the output impedance of the amplifier with 2 x 3 BS170's is well below 50 ohm.
?
With 2 x 3 BS170's, and a power supply of 10V, P-out is now 5.4 W on 20, 4.3 W on 17 and also 5.4 W on 15.
Harmonic suppression is very good. Nearly 60 dB on 20M and over 50 dB on 17 and 15M.
?
My main use will be CW contesting on 160, 80, 40, 20, 15 and 10M and on these bands the QMX+ now delivers 5W or more.
I gather this is good enough for the time being.
?
Maybe I will test some real power fets in the future, but for now adding just 2 BS170's is by far the cheapest and simplest option.
An other idea might be to try a different transformation ratio for T501, providing a better match.
?
I don't intend to use it off the grid, so power consumption is not a big deal.

?

73,

Roelof, pa0rdt

?

?

Ludwig, thank you for the hint.?Yes, exactly there is something wrong:

Values by decreasing voltage:

Supply (Multimeter)? ? Supply (Diagnostics)? ?Current RX/TX? ? ? 5V SMPS(Diagn.)

12.1 V? ? ? ? ? ? ? ? ? ? ? ? ? ? 12.3 V? ? ? ? ? ? ? ? ? ? ? ? ? ? 90 / 940 mA? ? ? ? ? ?5.00 V

11.0 V? ? ? ? ? ? ? ? ? ? ? ? ? ? 11.2 V? ? ? ? ? ? ? ? ? ? ? ? ? ? 90 / 900 mA? ? ? ? ? ?4.99 V

10.62 V? ? ? ? ? ? ? ? ? ? ? ? ? 10.4 V? ? ? ? ? ? ? ? ? ? ? ? ? ?100 mA / -? ? ? ? ? ? ? 5.01 V

10.60 V? ? ? ? ? ? ? ? ? ? ? ? ? 13.2 V? ? ? ? ? ? ? ? ? ? ? ? ? ?180 mA / -? ? ? ? ? ? ? 6.36 V

10.59 --> switching off

by increasing voltage again from this point - different behavior:

10.59V? ? ? ? ? ? ? ? ? ? ? ? ? ?13.7 V? ? ? ? ? ? ? ? ? ? ? ? ? ?140 mA / -? ? ? ? ? ? ? 6.56 V

11.25 V? ? ? ? ? ? ? ? ? ? ? ? ? ?8.8V? ? ? ? ? ? ? ? ? ? ? ? ? ? ?120 mA / -? ? ? ? ? ? ? 3.97 V

11.26 V? ? ? ? ? ? ? ? ? ? ? ? ? 11.0 V? ? ? ? ? ? ? ? ? ? ? ? ? ? 90mA / -? ? ? ? ? ? ? ? 5.0 V??

(I hope, after sending, my "table" is staying aligned in columns).

Stan:

I tried to calibrate three QMXs (low, medium and high band) and a QMX+.??

None calibrated correctly.? About half the bands failed early with crazy plots and stopped at close to 1000 on the x axis.

On testing, into a dummy load and receiving on my base station, audio was bad until I turned off predistortion.

I know Hans is working on making the calibration process more robust.? I assume it is not easy to trace the software issue.

On Fri, Apr 18, 2025 at 09:49 AM, Stojan Ratkovic wrote:

The minimal voltage in protection settings is 7V, why isn't it working with 9V?

Stojan, this setting doesn't switch off the QMX, see operational manual.

There is an other problem.

Could you show please screen shots from Terminal - Hardware tests - Diagnostics for 12V, 11V, 10V+x and 10V-x? And please add the supply current value for each.

?

73 Ludwig

Kai, please read QMX QMX_plus Audio PCM1804 from QRP Labs Wiki.

Could you send please screen shots from Terminal - Hardware tests - Rf filter sweep and Test ADC I/Q for a higher band and a lower band?

?

73 Ludwig

UPDATE:

i again did a factory reset and then programmed the QMX+ by display and knobs, not via putty.

i saw that is a special "character" for the space sign in the display. Now it seems to work correctly...

Will have a look for and inform you again.

Thanks for your help and tips, may now all is fine?

?

vy 73 de Udo DF4BJ

Stan,

?

- If you did not test the filter as installed in the QMX+, you should.

?

?

Hi Stan, I've investigated this further. The keyer actually cuts out irregularly even when I'm using it normally with my paddle. For example, I simply held the DAH paddle (on the dummy), and the keyer cuts out irregularly, sometimes after 3 or 10 strokes, sometimes only after 20 or so. Nothing else was connected or switched on that could have caused interference. with my two QMX before (80-20 and 20-10) all was fine , with newest ssb fw too...

no idea ...

?

vy 73 de Udo DF4BJ

My QMX is working perfect with 12V, but I would like to try 9V to reduce power on FT8.

By lowering voltage with a regulated power supply, qmx is going off at <10V (qmx seems not to be completely off, but there is no audio and no display).

By increasing voltage again, QMX is starting again at >10V (without pressing "On/Off knob).

At exactly 10V QMX is alternately going on and off.

The minimal voltage in protection settings is 7V, why isn't it working with 9V?

Vy 73 and thanks in advance, Stojan, DK4SR

I recently finished?building?my QMX. However, when I plug in my headphones to the audio port, all I get is static that stays the same regardless of whether or not I have an antenna?attached or not.

No, there is no message.? And I don't think any of us know yet definitively what constitutes successful 'completion'.

?

I will give my definition of 'complete' next, but first let me state that you don't need to do calibration for good SSB performance.? Just don't do it at all, or if you do it and are not happy with it, turn 'Phase predistortion' off, and you will still have good audio.? The calibration is supposed to give even better audio.? And if you have unsatisfactory audio, turn off all, then turn on, one at a time, each of the options in the SSB menu, to see which one is contributing to the problem and which are making it sound better.

?

Based on my own experimentation, I have been defining 'complete' "Phase Error" calibration on a band this way:

- The calibration plot should have * all the way to the highest pertinent DAC (x-axis) value

- This should happen on all of the bands; a band where it doesn't happen has 'incomplete' phase error calibration.

- The highest x-axis value is defined by your supply voltage and your 'drive level' (default 900) setting.? With the default drive level, that corresponds to about 2000 for a 9V supply, and 2600 for 12V.

?

Sometimes the phase error plot stops well short of 2000 or 2600 on the x-axis.? In my experiments, when that happens, the audio output on that band seems to be corrupted.? I think the corruption happens whenever the incoming audio needs a DAC value that is not initialized on the curve.? [But one other poster said he didn't think this was happening, so it is just my perhaps-inexact theory.? And some posters have said they think the corruption is due to an erratically shaped curve, not how complete the curve is.? But on my units an irregular shape doesn't seem to corrupt the audio.? We don't really know for sure.]

?

Sometimes just repeating the phase-error calibration on an 'incomplete' band will get the plot to complete.?? Often that doesn't happen, and it seems that if we 'tweak' the center frequency and/or band edges in the Band configuration tool, we can get it to complete calibration, sometimes after several trials.

?

And in my experience, this is mainly an issue with the 'Phase error' calibration.? The IMD Sync optimizations seem to always complete.

Stan KC7XE

?

Most of my bands are below 2 watts,? also I have a horrific "Click when I use my paddle and