Update: Definitely a toasted zener on the 3.3V PSU board.

Hooked up 3.3 board to bench supply. High current draw, zener gets hot. Removed zener. Current draw now reasonable. Installed 3.3 board into motherboard. Connected JP101 pin 8 to to JP201 pin 8 to trick CPU into thinking 5v rail is healthy. Connected bench supply to JP101 pins 4 and 2. CPU seems happy, I see PWM on the PSU PWM pins. If anything powered by 3.3 failed it didn't fail short. Still waiting on my buddy with the thermal camera to see if I can ID what failed short on the 5V rail.

First fire up on QDX 20-10. Led flashes normally. L4 started smoking. Yikes!
Now terminal shows only 80m, no other bands. Wsjt works into 50 ohms but usb sketchy
Replaced L4. Removed Diodes D1 and D4, they were shorted.
Where to next?

4 watts vs 3 watts is 1.25 db. IMO there are bigger fish to fry.

I intend to run off a car battery inverter so I took some extra steps to ensure thermal performance at 13.8 vs. 12. I put thermal paste between the bss170s and the ground plane, and a thermal pad on top to try and conduct more heat into the washer holding them down. Maybe at some point ill replace the washer with an aluminum bracket that attaches to the inside of the enclosure to conduct the heat out that way.

We'll see how it works when I figure out what I blew up in it...

My QDX was working fine until I was attempting to use it to drive a 100 watt HF amplifier. ?That amp failed, spectacularly, and now the QDX has these symptoms —
1. RX is good on all bands, with decodes appearing properly?
2. ?RX RSSI bar graph previously read (typically) 35-40 dB and was green, now reads 10 dB and is red.?
3. ? RX filter sweep is normal for all 4 filters
4. ?Attempts to transmit instantly shuts down the 9 volt 1 amp power supply after which the QDX restarts with the 5 second sequence and returns to normal receiver operating, albeit low sensitivity ??

I assume something in the common path is damaged such as the LPF switching, T/R switch or perhaps the transmitter final(s). ?

Suggestions for troubleshooting and fault isolation?

N6JH

?

Yup! Looking forward to the 'with' results!
--
73, Dan? NM3A

开云体育

开云体育

Chris, G5CTH

Chris, is this with or w/o Zeners?
--
73, Dan? NM3A

The short is somewhere on the main board, verified by pulling out the PSUs and testing each board individually. I wish the PSU was the problem since it's easier to just order another power module than to troubleshoot a very dense main board with a bunch of very tedious labor put into it that I'd rather not repeat... As it is I'm going to borrow a thermal camera from a friend and see what heats up when I try to power up. I live pretty close to a digikey warehouse so I can get pretty quick shipping on replacements.

I'm willing to bet the failure went something like this -

* Bench supply goes into current limit, drops input voltage
* Duty cycle of buck regulators increase
* Bench supply comes out of current limit when TX ends, input voltage comes back up
* Known issue with CPU failing to respond to transients in input voltage fast enough to prevent rails from spiking
* Protection zeners fail open? I don't see any shorts in PSU boards
* Op amps are probably fine since they're all 30 volt parts, but the Tayloe detector, ADC, DAC, and LCD are all liable to have been killed by this.

I would propose the following modification to the power circuit to help prevent this issue:

https://photos.app.goo.gl/shkveUjM5gU6WNqP8

This would shut down the pwm in an overvoltage event much faster than the cpu can.

The ultimate solution would be to use an off the shelf buck converter IC with a frequency synchronization feature. An IC whose control loop runs as fast as its pwm will naturally have better impulse response than a CPU running a control loop at 1khz.

I think it's just a matter of being cautious.

I use an Elecraft T1 to tune but use a tuning indicator (TI) that
ensures no more that 2:1 swr is presented to the rig. I have forgotten
to put the TI in tune mode more than once. All it takes is 1 bad
mismatch and maybe I kiss my finals goodbye. I'd rather not replace
them if I can avoid it. Since Hans puts it in big bold red letters to
not go over 12.0v I'm gonna listen... maybe necessary, maybe not...

Until I'm fully comfortable that these rigs are indeed robust, I'm
erring on the side of caution.

-mike/w1mt

Why would 5W at exactly 9V, for the 9V version be ok, but 5W at exactly 13V, for the 12V version, be dangerous? FWIW, I’ve ran my QMX off a LiFePO4, transmitting with as high as 13.8V, but mostly in the 13.2-13.3V range, with no problems. But to be fair, I’m using a manual tuner. ?But again, it borderline feels like what I’m reading, on a near daily basis, is running more than 4W at 12V is a sin.?

The right encoder / button problems are likely a bad ground - upper left pin on the control board if I remember right.? ?Recheck that one, that six layer? ground takes a lot of heat.??

Yes - I noticed that too late! My bad, the QMX should connect to XCVR. 72 Sandy KB3EOF

Great setup at low cost. One thing just wonders me: is the QMX connected to the ANT side of the tuner?

Ahhhh, 13.8v. That explains it. I target 4w at 12.0v when building
(matches the graph in the manual). Hans has warned against running
more than 12v on page 4 of the manual -"Do not power a QMX with a
higher voltage than you built it for."
Anything more than that risks the finals, especially when using an
automatic tuner. I personally use a 3s Li-Ion battery and when fresh
it's 12.6v and I'm a little nervous. So far so good.


I use 0603 for the limiting resistor cause that's what I have. No
problems with heating.

Enjoy!

-mike/w1mt

Hi Mike,

Nope, I built for 12V.? Changed L513 back to 10T and squeezed turns.? Played with L512 to optimize 40m (and 60m).? Now have right around 6W on all bands with 13.8V PS.

120 ohms looks good for display brightness.? I only had 0805, the 33 ohm is a 1206, but the 120 ohm 0805 does not get warm or hot so think it is fine.

I'll say it again, what a fabulous rig!

Thanks and 73,

Robert, WA2T

开云体育

I've started playing with a spectrum analyser, this is what I have so far.

There's a bit of scruff round the signal and the second and third harmonics are about -60 dB which seems pretty good, better than my Pixie.

And at this stage L513 had 9 turns instead of 10.

At present I'm using a crude 38 dB attenuator and was running about 1W.? I've a 10W 20dB attenuator coming and that should allow me to run my QMX at full power.

Not used a spectrum analyser before so a bit of a learning experience, there must be a way of getting a screen shot onto a SD card.

?And I have some Zeners, 1N4756A.? Hans, I could send you a couple if it helps.

I'm taking my time over this, not applied more than 10V to my 12V QMX, that gives 2 to 2.5W, astonishing how far that gets.

Chris, G5CTH

开云体育

Sure thing Dan - here is a sketch and photos. It is a simple OCFD and I optimized it for 40-30-20 meters using antenna modeling software (MMANA). The 4 state tuner tunes it easily on those three frequencies (and above). Sandy KB3EOF