Topics

uBitx V3 low power continues

Ted
 

This is a follow-up to another thread that is months old. I'm now recovering from cervical spine surgery & attendant complications of infection, so I've had time to repeat tests though not necessarily any clearer mind. Having only 15-20 degrees of rotation of my neck isn't helping, either, with bench work.

To review: a perfectly-functioning radio with typical RF out levels (12w out in 80, 10w in 40, 8w on 20, etc) has devolved to less than a watt on CW on all bands and even less on SSB. The exception is 20 meters, where a couple of watts in CW show up.

Let us assume the finals are fine and that all drivers and pre-drivers are OK. They all get warm, in fact. Suggestions were made to trace the RF thru the LPF array to ensure power wasn't being eaten by a stuck relay or bad filter. I think that this possibility has been eliminated, though I can't explain why the finals get warm and yet the output generally stinks AND worse on one mode than on another.

Again, the SSB signal is far below the CW/dead carrier level, perhaps it is 25-30% of carrier level (according to my oscillscope).  Output at the antenna connection shows approximately the same level as measured at the final transistors.

What dies, prior to Q90, that kills SSB and CW as described?   My sledge hammer works just fine and requires little skill to operate - and at this point makes a better communication tool than my uBitx. I'm not allowed to pick up 15 lbs, though, so that option is on standby. I just don't get it. 


Thanks,

Ted   K3RTA

Curt
 

Ted

If this were my rig, I would suspect my own wiring before anything on the board. Yes been there sometimes. Lets check as much of the path from board to antenna connector as possible. Something might look soldered and may not be connected, yes I was there troubleshooting for weeks. Use your dvm to examine continuity,  both through and insuring no short. Next check the keying circuit. Power tends not to vanish like this on the board. Be patient,  the problem is there someplace. Limit your sessions to say 20 minutes max. Fresh eyes better at finding things, along with a dvm.

Curt wb8yyy

MVS Sarma
 

Ted,
as and when time permits  take a photo of the Post PA  LPF sections. you would locate the LPF covering 40m band.
 check the turns of coils L17, L18 and l19.
Cross check with ft37-43 core on site calculator   at kits and parts. you will find that it  might be having additional turns.
   we can remove coils from top itself   no need to pull board out. try to tweak the coils by reducing and measuring, if possible.
 The power on 40m would come up.
secondly,
 perhaps one could try increasing the  local VFO , intermediate and BFO osc levels. by manipulating the code in si5351 firmware.
 they are set at 2mA by defining 1, 1, 1  for clocks 0, 1 and 2
we can perhaps try with 2, 2, 2 instead.

 notice that on v5 the designer had made it 3, 3, 3

 we can always restore them to old levels by a click of button on arduino ide.

all the best for a speedy recovery from spinal issues.

 sarma vu3zmv



On Thu, Oct 31, 2019 at 12:13 AM Ted via Groups.Io <k3rta=yahoo.com@groups.io> wrote:
This is a follow-up to another thread that is months old. I'm now recovering from cervical spine surgery & attendant complications of infection, so I've had time to repeat tests though not necessarily any clearer mind. Having only 15-20 degrees of rotation of my neck isn't helping, either, with bench work.

To review: a perfectly-functioning radio with typical RF out levels (12w out in 80, 10w in 40, 8w on 20, etc) has devolved to less than a watt on CW on all bands and even less on SSB. The exception is 20 meters, where a couple of watts in CW show up.

Let us assume the finals are fine and that all drivers and pre-drivers are OK. They all get warm, in fact. Suggestions were made to trace the RF thru the LPF array to ensure power wasn't being eaten by a stuck relay or bad filter. I think that this possibility has been eliminated, though I can't explain why the finals get warm and yet the output generally stinks AND worse on one mode than on another.

Again, the SSB signal is far below the CW/dead carrier level, perhaps it is 25-30% of carrier level (according to my oscillscope).  Output at the antenna connection shows approximately the same level as measured at the final transistors.

What dies, prior to Q90, that kills SSB and CW as described?   My sledge hammer works just fine and requires little skill to operate - and at this point makes a better communication tool than my uBitx. I'm not allowed to pick up 15 lbs, though, so that option is on standby. I just don't get it. 


Thanks,

Ted   K3RTA

 

Hi Ted,

I hope you are recovering well from you surgery and infection.  That's tough breaks for sure.

I think from your former post, you said you have an oscilloscope?  If this was my board, I would start going through it using the 'scope.  For example, rather than just assuming the PA chain works OK because it gets warm, put the rig on 80m and follow the CW signal through.  Check the gain on each stage and make sure it all looks good that way.

For SSB, put a tone out of your cell phone (lots of aps for that) onto the mic and follow that signal.  You should be able to see the audio signal right at the mic and then see gain in each stage up to the mixer.  Then you should see the signal at RF and keep following it.

While not exactly the same rig, this site http://golddredgervideo.com/kc0wox/bitx20a.htm will show you how signals look as you follow them through the rig.  Signal tracing on the TX side is not too bad since the signals are all reletively large.

73,


Mark

ajparent1/KB1GMX <kb1gmx@...>
 


A Scope is nice if your understand what your looking at.

The easy way is the amp has an output line that goes to the
relays and filters.  break that line and measure power there.

There is nothing that dies prior to Q90 that is in the Power amp.

However there are two IF amps that can be all sorts of wrong
and the alignment of the oscillators as well.

However CW is generated differently than SSB and for CW TX
none  of the IF stages are used.

So the only things that can be broke is the 33mhz LPF alone
the edge or the power amp.

Allison

ajparent1/KB1GMX <kb1gmx@...>
 

If the problem is 20M only its either a filter issue or a relay issue for 20M only.

Not a lot of parts but I've seen the same complaint for 40M several times in V3 boards.
Problem wrong number of turns on the filter coils.

Allison

Ashhar Farhan
 

Does the receiver work well? Are stations clear?


On Sat 2 Nov, 2019, 7:51 AM ajparent1/KB1GMX, <kb1gmx@...> wrote:
If the problem is 20M only its either a filter issue or a relay issue for 20M only.

Not a lot of parts but I've seen the same complaint for 40M several times in V3 boards.
Problem wrong number of turns on the filter coils.

Allison

Ted
 

Ashhar,

Yes, the receive is spectacular, with an added RFpreamp installed.  I can pick up whatever my Kenwood TS130S can hear. 

Please note the very low SSB modulation in comparison with a tune/CW note. I'd expect a fully-modulated "FIIIIIVVVE" to be aroud the same 'scope width as a CW note, give or take; my SSB is no more than about 25-30% of that with a compressor in line that had worked perfectly until I let the magic out somewhere along the way.

I was thinking "balanced mixer out of whack" or some such, though my theory isn't up to date.  Now, I asked about circuits prior to Q90 because this wide disparity between SSB peak envelope and CW occurs before that, hence my thought that the
problem could preceed this stage.

Thank you, sir, for any further thoughts on the subject.

Ted
K3RTA

Ashhar Farhan
 

Ted,
The CW is generated independent of the IF and audio system. On the other hand the receiver is working, this rules out any problem in the RF mixer or the low pass filter. 
You must concentrate on the transmitter drivers and PA. The trouble is with those stages. Do you have access to an oscilloscope?

- f

On Sat 2 Nov, 2019, 6:44 PM Ted via Groups.Io, <k3rta=yahoo.com@groups.io> wrote:
Ashhar,

Yes, the receive is spectacular, with an added RFpreamp installed.  I can pick up whatever my Kenwood TS130S can hear. 

Please note the very low SSB modulation in comparison with a tune/CW note. I'd expect a fully-modulated "FIIIIIVVVE" to be aroud the same 'scope width as a CW note, give or take; my SSB is no more than about 25-30% of that with a compressor in line that had worked perfectly until I let the magic out somewhere along the way.

I was thinking "balanced mixer out of whack" or some such, though my theory isn't up to date.  Now, I asked about circuits prior to Q90 because this wide disparity between SSB peak envelope and CW occurs before that, hence my thought that the
problem could preceed this stage.

Thank you, sir, for any further thoughts on the subject.

Ted
K3RTA

Ted
 

I do have an oscilloscope handy. What would you like me to look at?  


- Ted
  K3RTA

Ted
 

On Fri, Nov 1, 2019 at 10:18 PM, ajparent1/KB1GMX wrote:
However there are two IF amps that can be all sorts of wrong
and the alignment of the oscillators as well.
I'd be happy to take a look, if in had an idea what one should be measuring at which points.   

Ted

Ted
 

Mark,

Thanks for the well wishes and such. Doing as Curt suggests, 20 minute & otherwise short intervals would probably help.  Will regroup, apply tone, and work the stages.   This stuff really requires a clear mind! 

Thanks again, all. Will report back with data.

Ted
k3rta

 

You're welcome Ted,

Good luck and I think short stints is a good idea too.  Take written or audio notes as you go, so you can note measurements and remember what you did and where you left off.

You might check the compressor too.  Perhaps it has a power problem or a level control or compression control got changed?

Regardless of the SSB and CW difference, it seems you still need to sort the overall lower power output.  So investigation of the PA power chain seems a good thing to do.  If it would help, I could dig out my rig (just moved) and do some 'scope and RF probe measurements of it.

Someone here may have already done that and made a list of measurements or taken 'scope pictures?

73,


Mark.

Ted
 

Mark [et al],

Using a basic oscilloscope for measurements and a tone generator app for audio, each stage of the amp section seems to raise the RF level an appropriate amount. The exception to this is Q90, whose Collector output is lower than the Emitter input.  Maybe that's how that stage is meant to operate; the voltage measrements are about what the V4 troubleshooting guide says they should be.  Actually, all transistor voltage levels in the driver and PA stages are where they belong.

Regarding mike compression, that part works great as verified by an audio tracer and on the 'scope.  Even with tons of gain cranked up for the sake of following patterns through the circuit board, the peak-to-peak level only comes to about 2/3 of CW levels.  This dispairity is evident as early on as L1-L4.  My checks at Q20-Q22 are hard to read (by me at least).

That's where my hour of patience got me today.  I can try tomorrow to repeat some stuff, maybe/possibly get some pics if it's of any help.

73,

Ted
K3RTA

Ted
 

On Fri, Nov 1, 2019 at 10:18 PM, ajparent1/KB1GMX wrote:
There is nothing that dies prior to Q90 that is in the Power amp.

However there are two IF amps that can be all sorts of wrong
and the alignment of the oscillators as well.

However CW is generated differently than SSB and for CW TX
none  of the IF stages are used.

Allison,

Thank you for those insights. Now might I test the IF sections for fitness ?  

Just as an aside, the radio was working for a good while up until it didn't.  I was doing one thing or another on the bench the specifics of which I don't recall, when this power diminishment occured.   The amp section seems fine, but can't amplify what it has not been given from earlier stages;  I'll try anything at this point.


Tnx Agn ,

Ted
K3RTA

Evan Hand
 

Here are the oscilloscope pictures of test points 3 to 7.  First two are at different time scale than the last one.  The last picture is the power reading using my NISSEI SWR/Power meter.

Hope this helps.
Evan
AC9TU

 

Evan Hand
 

I should have included that the frequency was 7.077 mhz into a 50 ohm dummy load.

Also, I have two pictures of TP 5.

Evan Hand
 

Mode was CW.

Ted
 

Cool!  My scope is a bit more primitive than that one :)   I'll give it a workout and compare.

Thanks very much.



-ted
 K3RTA


On Tue, Nov 5, 2019 at 19:01, Evan Hand
<elhandjr@...> wrote:
Mode was CW.

Ted
 

Well, I went and tested the points specified, and could not get the resolution seen in the photos.  

Here are TP 3 -




and TP4 -




Considering some of the Chinese hand-held scopes on the market, newer technology be helpful.  What sort of buffering are you using between the test point and the end of the test leads?

-Ted
 K3RTA


On Tue, Nov 5, 2019 at 18:56, Evan Hand
<elhandjr@...> wrote:
Here are the oscilloscope pictures of test points 3 to 7.  First two are at different time scale than the last one.  The last picture is the power reading using my NISSEI SWR/Power meter.

Hope this helps.
Evan
AC9TU