Problem connecting to external PA

[DO1OKE]

Hey guys,

I analyzed the underlying issue and also have an easy solution for you!

The problem is, that the PA is keyed using an internal 12V relay. The tip of the truSDX PTT socket is directly connected to the MCU which is running at 5V. That means that if PTT is off truSDX will output 5V, while 12V are coming from the PA on the PTT line (just measure the PTT line from the PA with a multimeter to see the 12 V). That means that the truSDX can't 'block' the 12V with its 5V output (the 12V could also potentially destroy that MCU Port permanently!). This could be fixed by adding an MOSFET inside the truSDX that is switched by the MCU, so the MCU is isolated and only enables or disables the MOSFET (this is acutally done be the QDX REV 3 https://qrp-labs.com/images/qdx/schem3.png).

The keying circuit by KB9RLW is insufficient in this case because it does not protect the truSDX from the 12V by the PA.

But I found another solution. Actually only a single N-channel mosfet would be required with the gate connected to the truSDX and the drain and source to the PA. However, the output from the MCU is inverted so I added an inverter in front using another MOSFET. The final circuit is shown in the attached picture. I also added the LTSpice simulation as a ZIP file.

The circuit can be built using almost any N-channel MOSFETs. The pullup resistor R2 should be at least 1M ohm. Otherwise the current through the pullup will already key the PA (took me a while to notice that problem ). The resistor R1 only limits the current through the MOSFET.


The MX-P50M is working great. It gives me about 10dB more SNR on CW at the RBN and also on pskreporter when using FT8 with my QDX. So it really makes a difference for DX.

Hey Chris,

Thanks for the circuit! It is working. But for some reason it keeps killing my mosfets. It already sent 3 BS170 and 2 2N3906 to heaven. Any idea why? It's always the PA facing mosfet that gets destroyed. I wonder if its got to do with the PA relays and some peak voltage when turning it on...


73! Simon

[dl6sez]

Hello OM Simon,

probably you are right...
I would add 1N4148 or 1N4007 as Drain to Kathode and Source to Anode. If it works than the internal protection diode of the FET was to week.
But maybe other possibility: the current is to high for one BS170, take two in parallel or a NMOS-FET with higher current capabilities.

73 de Chris

[DO1OKE]

Hello OM Simon,

probably you are right...
I would add 1N4148 or 1N4007 as Drain to Kathode and Source to Anode. If it works than the internal protection diode of the FET was to week.
But maybe other possibility: the current is to high for one BS170, take two in parallel or a NMOS-FET with higher current capabilities.

73 de Chris

Yes indeed, putting a diode across prevented the next mosfets from the same faith These inductive voltage peaks are really not to be messed with...

[Creidiki]

Apologies If I'm resurrecting a zombie thread but here's my cirquit to key MX-P50M. KiCad files available if anyone interested.
BR
OH2GVD

Screenshot from 2023-04-28 21-08-14.png (6.93 KiB) Viewed 581 times