QDX low power output - where to start troubleshooting? #qdx #troubleshooting

Luke / AD0KI

Just finished up my QDX Rev 3 yesterday. No apparent problems from the initial build. I went slow and checked each component as much as possible before installation. Everything seems to be functioning normally - Connects to PC just fine, status light functioning as expected, receives FT8 signals just fine, transmits just fine (made an FT8 contact). My only problems seems to be lower than expected power output. Looking for tips on what I should start checking first?

- 80m = 2.5 watts 40,30 & 20m = 2 watts
- It is built for 9V (3:3 winding).
- Powerwerx DC power supply (14V) stepped down to 9.3V using a buck convertor. Voltage drops 0.3V under TX load to 9V even.
- Beta 1.04 firmware 
- Windows 10 with latest ver 2.4 WSJT-X 

Luke / AD0KI

Bruce K1FFX

Suggest you start by adjusting spacing on the LPF toroids.  On my (9 volt, 3:3) QDX build, compressing
the turns cranked output power from < 4 watts to a bit over 5 watts.

- Bruce K1FFX

Luke / AD0KI

Compressing the turns didn't do much of anything that I could tell. RX current is about 144ma, and on TX it's about 330ma which is definitely lower than it should be. Any other suggestions? 

The reciever is definitely sensitive! Had a couple FT8 decodes while I was testing it and connected to the dummy load!

Luke / AD0KI

I just reflowed the majority of the solder joints with no improvement, and also used a 9v battery just to eliminate the power supply. 9v battery worked, but with even lower output power (as would be expected).

Luke / AD0KI

Also checked the LPF diodes and they're working correctly (0V on current band, and about 6-7V on the "off" band filters).

Evan Hand

I would also check T1 for stripped wire.  The pictures are not high enough resolution to tell, but it is either lighting or bare copper showing on the vertical picture.  I have found that the ferrites for the transformers have sharp edges that took aggressive deburring to remove.   On my first QDX build, I found red flakes of enamel coating where I was winding the transformer.  

Verify that you are getting 9volts on the drains of all 4 BS170s.

The other area to check would best be done with an oscilloscope.  Verify that the gates of the 4 BS170 MOSFETs are getting close to a square wave, and if you have dual-trace, the gates of Q8 and Q10 are 180 degrees out of phase of Q9 and Q11.  If the signals are not close to a 5volt square wave, I suspect IC5 has failed.  If you do not have a scope, measuring 2.5volts DC on all 4 gates when transmitting indicates that IC5 is OK.

The fact that the current on transmit is so low for even 2 watts suggests that maybe one of your instruments is not reading correctly.  For 2watts, I would expect closer to 450ma.

Let us know what you find.

Luke / AD0KI

Just checked the BS170 drains before running out the door for work and verified they're all at about 9.2V.

I'll check that the gates are at 2.5V during TX this evening.

I'm measuring the power output using a home built dummy load roughly copied from the Elecraft DL1 (4x 10w 220ohm non-inductive resistors in parallel with a 1N5711 diode and 0.01uF capacitor.). I checked it against both my IC-705 and TX-500 at 5 and 10 watts and I was achieving accurate power numbers by measuring the output voltage, squaring it, and then dividing by 100.


For the DL1, elecraft’s stated formula to calculate watts is (V+.25)squared/100. The .25 is to compensate for the diode.

Luke / AD0KI

Thanks for the clarification. I wish my findings were just a simple math error, but even adding back in the 0.25V only changes the calculated output by about 100mw. I just need to find where the other 2 or 3 watts are going! :)

Al Holt

It may be possible one or more of the BS-170's is/are bad. With QDX's push-pull output I would think you'd see a very distorted RF waveform on an oscilloscope or maybe a very distorted tone in a nearby SSB receiver.

Were you able to test the FET's before installing them?

Being it's a new build I wouldn't think they would so quickly fail. Just a thought. Hope it's an easy repair!




I had a low power fault (less than a watt) at the first start up on several bands.
After deeply checked what I could have wrongly done, I have found my mistake :

Where : the solder side of L1 and C14

What : I have put a big solder blob on two pads connecting the two components which must not be. 
The result is a pull down to ground of the VCC line (thru a 1k res) dedicated to the switching (pin diode) of the BP filters.

Why : the distance between the two pads is so close.

Attached three screenshots, two taken with a microscope, as an help in locating, IMOA, one risky area of the PCB.
I hope this will help.

73, F4IPL, Nicolas

Luke / AD0KI

Still scratching my head on this one...

This evening I verified 2.5 volts on all four of the BS170 gates during transit into dummy load.

Checked for solder bridges on L1 & C14 - no problems there.

CW "tune" signal from WSJT-X sounds clean and undistorted from neighboring transceiver. 

I played around a little more with the toroids and compressing the turns resulted in slightly lower output power, not higher.

After messing around with everything I double checked the power output at 10v input. Here's the results (+/- 100ma). 80m=2.6w, 40m=2.3w, 30m 3.3w, 20m=2w.

Tomorrow I will take the QDX in to work and check the BS170's and RF output on the oscilloscope. 

Luke / AD0KI

Can someone post a picture of what both the left and right sides of the T1 should look like for a 3:3 winding and a 3:2 winding? 

Luke / AD0KI

Can someone verify this is correct? 4 loops visible from left side of T1 and 3 from right?

Al Holt

My QDX is set to operate on 9V thus a 3:3 wound T1. Here are my photos:
It's not easy to count turns. Photo on left looks like 6. Photo on right (with center-tap) has 6, too.
Hope this helps!


Luke / AD0KI

Thanks again, Al! Considering that everything is working as expected, minus the power output, it has me wondering if I somehow botched up T1. The power output example in the build manual would be almost spot on if I had accidentally wound T1 3:2 (for 12v operation) and was only running it off of 9v.

Al Holt

You could simulate winding a 3:2 transformer using a couple of short tubes taped together making a binocular core and some random wire or cord and see what each side shows. Rather than remove what you have installed. 

It's better for me to have some physical model to see rather than trying to work it out in my head. Usually it more reliable :-) 

Just a thought!


Al Holt

BTW, when I wound mine I used one of those low cost (read offshore) LC meter and measured the inductances of each side. 
To me, the turns count got confusing when you get to the secondary.

https://toroids.info/ can be used to get the approximate inductances which you can check against what you read with your LC meter. So, 3 turns in a BN-43-202 core is 19.80uH. 2 turns is 8.8uH. 

And Bob's your uncle :-)


Luke / AD0KI

Update: I took off T1 and it was definitely missing a turn on the secondary so I think I accidentally had it wired 3:2. I re-wound it and now I'm getting:

80m= 4.7 watts
40m= 3.6 watts
30m= 4.7 watts
20m= 3.9 watts

Much better than it was, but 40m and 20m are still a wee bit low. If I could get them up into the 4.5 watt range I'd be a happy camper. I'll try tinkering with the LPF's and see if I can get a slightly better result.

Thank to all for your help and suggestions!