Page 1 of 1

PA100-D Problem with F-Sense not detecting Band

PostPosted: 20 Apr 2017 21:38
by HB9VQQ
Hi all,

ELAD FDM-Duo connected to newly purchased (second hand) Juma PA100D FW 3.00c. Connected 3.5mm PTT out Jack to T/R Jack of the PA. Auto Band Detect set to F-Sense. Auto Band selected so I expect to detect TX Band by F-Sense. Lower line of LCD reads:

Image


When I hit the tune button (feeding 3 Watts to the PA) on the Duo the lower Line of LCD reads:

Image

When I use the manual Band select Mode all is working es expected. So there seems to be a problem with F-Sense I guess. Any ideas ?

73's
Roland

Re: Problem with F-Sense not detecting Band

PostPosted: 20 Apr 2017 21:39
by HB9VQQ
After further examination I found C1 in a rather bad condition on the F-Sense board

Image

According to the parts list this seems to be a 1nF ceramic capacitor size 1206 X7R or G0G. Hopefully this is the only reason why F-Sense is not working. Wonder how that could happen.

73's
Roland

Re: Problem with F-Sense not detecting Band

PostPosted: 20 Apr 2017 21:39
by HB9VQQ
50 VDC ok for the capacitor ?
Removed the remaining pieces of C1 and cleaned the area around carefully, this is what it looks like..
Apparently this part of the board became very hot, the black area is completely burnt.

Image

Re: Problem with F-Sense not detecting Band

PostPosted: 21 Apr 2017 10:25
by HB9VQQ
After soldering a new 1000pF 100VDC capacitor (C1) I still get an "Err" when Auto Band select mode is enabled...

Any ideas ?

Re: Problem with F-Sense not detecting Band

PostPosted: 22 Apr 2017 11:47
by 5B4AIY
PA-100D - 2.jpg
PA-100D driven from Elecraft KX3
PA-100D - 1.jpg
PA-100D, F-Sense, showing frequency and I/P power (0.4W)
Hi,

Attached are two images showing the PA-100D in the F-Sense mode displaying both the input frequency and power, and the AUTO Band Select Mode correctly selecting 14MHz, this is with version 3.00c Build 3 of my firmware, showing that the software is correct.

The '?' indication in the AUTO Band Select mode shows that although the AUTO mode is selected, there is no valid band data. The 'Err' message shows that you have attempted to go from STDBY to XMIT with no valid band select in the AUTO mode.

A 50v 1nF capacitor is fine. The blackened board rather suggests someone injected WAY too much input power, in which case, let's hope that this did not propagate through and cause too much damage.

Unfortunately, you have a hardware fault. Equally unfortunate, the schematics are wrong when it comes to identifying the various plugs and sockets. The correct connections are:

The F-Sense board pin 1, marked with a dot, is connected to J3 pin 1 on the main board, and that is directly connected to J5 pin 2. This is incorrectly identified as J2 on the schematic.

From here the signal is taken via a ribbon cable to the front panel control board J2 pin 2, to C7, and there to the biasing resistors R6/R7 and to pin 2 of IC1, a NC7S14M5 single inverting Schmitt Trigger. The output, from pin 4, is connected to R8 and then directly to pin 60 of IC9, the main microprocessor.

Unfortunately you will need an oscilloscope to positively determine the failure. It will need a bandwdith of at least 5MHz, and preferably 10MHz. You do not need to have the amplifier in the transmit mode, the F-Sense logic works in both STDBY and XMIT.

Remove the top and bottom covers. Connect your driving source to the input, and a dummy load to the output. Inject a suitable RF drive signal, at 1.9MHz, 1W is more than enough, leave the amplifier in STDBY. Check for the presence of a clipped sinewave at pin 1 on the F-Sense board, indicating that the input transformer and diode limiters are working.

Turn the amplifier upside down, and check that this clipped sinewave appears at pin 2 of J5. (You could also verify this connection from J3 pin 1 to J5 pin 2 with a multi-meter.)

With the amplifier standing vertically on its front panel, (The mounting screws will space the panel off the worktop ensuring no buttons are pressed.) place the scope probe on R8 and verify the presence of a 5V square wave.

If you have a square wave at this point then the fault is in the input circuit of the microprocessor, and that's fatal, there's nothing you can do other than replace IC9. You could check with a multimeter the physical printed circuit board connection from R8 to pin 60 to make quite sure that there is a connection.

In all probability you will find the fault on the F-Sense board, or either there is an open circuit trace or ribbon connection, or a fault with IC1.

73, Adrian, 5B4AIY

Re: Problem with F-Sense not detecting Band

PostPosted: 22 Apr 2017 15:30
by HB9VQQ
Thanks Adrian, I will need to send the PA in for repair since I neither have the required equipment nor the required knowledge.
Since the PA is working fine with the manual Band select I guess there is not too much damage. As you already pointed out it seems there was way to much input RF to the F-Sense circuit at some time.

Thanks for clarification.
73s
Roland