Re: XV432 Oscillates with Amp connected - FIXED

[Stu2-3]

Well - probably not fixed, but I was able to find a workaround.

If you haven't been following this thread, the basic issue is that the
PA on my XV432 self oscillates. Thanks to everybody on the list,
including Gary from Elecraft, I had several different things to try.

1) Make sure the paint on the inside of the case was removed in the
appropriate spots (connectors, joints, etc.)
2) Add some vias near U6 and K2
3) Touch up the solder joints
4) Make sure the leads on the resistor specified in the XV432 mod was as
short as possible
5) Change the PA Bias Adj
6) Add a pad in the TX path between the amp and XV432
7) Add a power resistor on a T connector at the XV432 output (or Amp input)

Ultimately, I found that varying the length of the coax between the
XV432 and Henry 2004A was the answer.  I discovered this when I got to
idea #7. First, I used a Bird Watt meter and adjusted the Henry amp to
match my TS2000. (no reflected power) Then I reconnected the XV432. I
added a 50ohm 20W dummy load on a BNC T at the output of the XV432 with
a bunch of different adapters to make the N-BNC connection work. The
XV432 was still connected to the Henry. This essentially put the dummy
load in parallel with the Amp. The self oscillation stopped.

So I removed the dummy load to prove the hypothesis and still no
oscillation. Then I renewed my original track of varying the coax
lengths and stumbled on a good length, which stopped the self oscillations.

I know the length of cable changes the load. So it would seem the next
step might be developing a simple matching network at the output of the
XV432. But since I don't know all the variables (e.g. S parameters) this
would probably need to be trial and error with some good guess work.

Thanks to everybody on the list who helped out. Perhaps I'll see you in
Dayton.

73's

Stu2
W7IY
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[Don Wilhelm-4]

  Stu,

Since the length of the coax changes things, I would suggest putting the
matching circuit at the Amp end of the cable rather than the XV432.  The
cable length should not matter if it is matched at the amplifier end,
but since the length does change things, it is apparent to me that the
amp input does not present a 50 ohm load.

73,
Don W3FPR

On 5/16/2011 5:50 PM, Stu2 wrote:

> Well - probably not fixed, but I was able to find a workaround.
>
> If you haven't been following this thread, the basic issue is that the
> PA on my XV432 self oscillates. Thanks to everybody on the list,
> including Gary from Elecraft, I had several different things to try.
>
> 1) Make sure the paint on the inside of the case was removed in the
> appropriate spots (connectors, joints, etc.)
> 2) Add some vias near U6 and K2
> 3) Touch up the solder joints
> 4) Make sure the leads on the resistor specified in the XV432 mod was as
> short as possible
> 5) Change the PA Bias Adj
> 6) Add a pad in the TX path between the amp and XV432
> 7) Add a power resistor on a T connector at the XV432 output (or Amp input)
>
> Ultimately, I found that varying the length of the coax between the
> XV432 and Henry 2004A was the answer.  I discovered this when I got to
> idea #7. First, I used a Bird Watt meter and adjusted the Henry amp to
> match my TS2000. (no reflected power) Then I reconnected the XV432. I
> added a 50ohm 20W dummy load on a BNC T at the output of the XV432 with
> a bunch of different adapters to make the N-BNC connection work. The
> XV432 was still connected to the Henry. This essentially put the dummy
> load in parallel with the Amp. The self oscillation stopped.
>
> So I removed the dummy load to prove the hypothesis and still no
> oscillation. Then I renewed my original track of varying the coax
> lengths and stumbled on a good length, which stopped the self oscillations.
>
> I know the length of cable changes the load. So it would seem the next
> step might be developing a simple matching network at the output of the
> XV432. But since I don't know all the variables (e.g. S parameters) this
> would probably need to be trial and error with some good guess work.
>
> Thanks to everybody on the list who helped out. Perhaps I'll see you in
> Dayton.
>
> 73's
>
> Stu2
> W7IY
>

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[gm3sek]

Stu2 wrote:
>
>1) Make sure the paint on the inside of the case was removed in the
>appropriate spots (connectors, joints, etc.)

A stainless steel scratch pen is your best friend -

USA:<http://www.cooltools.us/Scratch-Brush-Pen-Style-Steel-p/brn-209.htm>

UK:
<http://www.eileensemporium.com/index.php?option=com_virtuemart&page=shop
.browse&category_id=1&Itemid=9>

Use the centre hole of a large steel washer to isolate the area you want
to clean. Don't stop when the black paint has gone - make sure you also
cut away the primer, which is very tough and not so easy to see.


--

73 from Ian GM3SEK
http://www.ifwtech.co.uk/g3sek
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html