开云体育

I'm running out of ideas and need some help. First, it would be great if there was a schematic with voltage measurements on it that one could use for debugging a non-working ubitx.
I have the following set up:
ubitx to WM-1 wattmeter to dummy load. I'm using CEC's software v1.04. I can receive just fine. I've adjusted the bias of the 510's to 100ma each. I've done calibration and BFO setting. Once again the radio works fine on receive. I have very low output on all bands.
I have made the following measurements and I'm looking for someone to help me figure out where the power is going.
I'm running at 13.8V (was at 12V, but thought extra would help the PA). The 12V numbers are equally bad, just a bit lower.
So, I measured from the wiper of RV1, the power out, and the antenna voltage (on a 50mhz scope) the following:
RV1 at full clockwise:
Band ?? RV1 ? ? Power ?? Ant V
80 ? ? ?? 640mv ?? 4.8W ?? 21V
40 ? ? ?? 520mv ? ? 700mw? 12.2V
30 ? ? ?? 660mv ? ? 700mw ? 15v
20 ? ? ?? 1.2V ? ? ? ? 300mw ? 14v
15 ? ? ?? 640mv ? ?? 7mw ? ? ? 2.8V
10 ? ? ? ? 340mv ? ?? 3mw ? ?? 600mv
Moving RV1 counterclockwise will lower these values as would be expected. Do these values for RV1 look right? If they do then the problem is further down the PA chain. Again, I wish there was a table of receive/transmit voltages that one could reference.

So, then I thought I'd measure the clocks to see if there was anything amiss there:
at 7.150 Clk0 12Mhz 1.44V on receive; 0V on XMT
Clk1? 660mw at 33Mhz; 0V receive
CLK2 1.2V XMT (7.150mhz); 520mv receive 43Mhz

Help would be very much appreciated. Thanks.
73, Mike W1USN

Mike,

If you hook an ammeter into your power lead going to the ubitx what
does it show when you push the PTT?

Mine shows about 600ma with no microphone connected.

tim ab0wr

On Thu, 22 Mar 2018 16:05:25 -0700
"Mike R." <ubitx@...> wrote:

I'm running out of ideas and need some help. First, it would be great
if there was a schematic with voltage measurements on it that one
could use for debugging a non-working ubitx. I have the following set
up: ubitx to WM-1 wattmeter to dummy load. I'm using CEC's software
v1.04. I can receive just fine. I've adjusted the bias of the 510's
to 100ma each. I've done calibration and BFO setting. Once again the
radio works fine on receive. I have very low output on all bands. I
have made the following measurements and I'm looking for someone to
help me figure out where the power is going. I'm running at 13.8V
(was at 12V, but thought extra would help the PA). The 12V numbers
are equally bad, just a bit lower. So, I measured from the wiper of
RV1, the power out, and the antenna voltage (on a 50mhz scope) the
following: RV1 at full clockwise: Band ?? RV1 ? ? Power ?? Ant V 80
? ?? 640mv ?? 4.8W ?? 21V 40 ? ? ?? 520mv ? ? 700mw? 12.2V 30
660mv ? ? 700mw ? 15v 20 ? ? ?? 1.2V ? ? ? ? 300mw ? 14v 15
640mv ? ?? 7mw ? ? ? 2.8V 10 ? ? ? ? 340mv ? ?? 3mw ? ?? 600mv Moving
RV1 counterclockwise will lower these values as would be expected. Do
these values for RV1 look right? If they do then the problem is
further down the PA chain. Again, I wish there was a table of
receive/transmit voltages that one could reference.

So, then I thought I'd measure the clocks to see if there was
anything amiss there: at 7.150 Clk0 12Mhz 1.44V on receive; 0V on XMT
Clk1? 660mw at 33Mhz; 0V receive
CLK2 1.2V XMT (7.150mhz); 520mv receive 43Mhz

Help would be very much appreciated. Thanks.
73, Mike W1USN

I suggest sticking with 12v into the rig till you get it working,
higher voltages just make it easier to blow the finals.?

Best to fuse the supply, perhaps 2A on the PA-PWR to the finals and 0.5A on the main board.
It's real easy to short something out when probing around all day looking for trouble.
Don't ask how I know.


You show 4.8W on 80m, which might be reasonable.
Are you transmitting by sending CW???

Moving RV1 counterclockwise should raise those values, not lower them,
as that pot is backwards from 99.99% of the other pots in this world.

Perhaps the band-switching relays for the transmit low-pass filters are stuck in the 80m position?
Use the scope to check the RF voltage at KT1 pin 12, is it much much higher on those other bands
than you see at the antenna when attempting to transmit?? ?

You measure a bunch of RF voltages, but don't state if they are pk, pk-pk, or rms.

These look a bit strange.
? So, then I thought I'd measure the clocks to see if there was anything amiss there:
? at 7.150 Clk0 12Mhz 1.44V on receive; 0V on XMT
? Clk1? 660mw at 33Mhz; 0V receive
? CLK2 1.2V XMT (7.150mhz); 520mv receive 43Mhz

My best guess is that clk1 is 660mv on receive and 0v on transmit.
And that you are transmitting CW, and that the CEC firmware
is shutting down the CLK0 and CLK1 during transmit, which is fine.
For CW, clk2 operates at the transmit frequency, as you report here,
whereas for receive I'd expect clk2 to be your operating freq above the 45mhz first IF,
so at 45+7.15=52.15mhz, not the 43mhz you report.

The lowest freq clocks are at 1.2 and 1.44v, the higher freq clocks less than half that.
The Si5351 should not have any trouble maintaining the same output voltage across the range,
I suspect this is an issue with your scope or scope probes somehow.

You seem to have more or less reasonable signal at RV1 for transmit.
Check the voltages at the bases and gates of the 3 following stages, see if they
are roughly the same order of magnitude for the upper bands as they are for 80m.

Jerry, KE7ER

Hi Tim,
Thanks for the reply. In receive it draws about 160ma. In CW transmit with RV1 all the way counterclockwise it draws about 1A on 7.150 into a dummy load. With PTT it is 470ma in LSB.
Mike W1USN

Thanks Jerry!
I can't get to everything tonight but will tomorrow afternoon.
To answer a couple of your questions now...
I'm using a 2A fuse inline from a current limited PS that maxes out at 3A (I'm not current limiting at the present, but have it set for the max of 3A). I currently have the red and brown wires to 12V so the PA is not separated.
Ha, I misstated the position of RV1. Of course I meant when I turn it clockwise...it is fully counterclockwise for these tests.
I like the 80m relay theory. I'll check that out first thing tomorrow when I get back to the rig. Are they latching relays?

I'm measuring the voltages on a scope. I assume they are pk-pk. I'm reading the value from the menu and not by guestimating from the waveform.
I'll recheck the clocks and also look at the stages past RV1 tomorrow.
Thanks a lot for the help - it is appreciated!
73, Mike W1USN

Mike,

On CW, frequency 3.95Mhz I get about 35v pk-to-pk on the tab of the
irf510's. That's a little higher than you show.

Ck message 43844. It has a jpg from ve3pe showing some of the voltages
(dc and rf) for the driver stage. This might be of some help to you.

tim ab0wr

On Thu, 22 Mar 2018 16:05:25 -0700
"Mike R." <ubitx@...> wrote:

I'm running out of ideas and need some help. First, it would be great
if there was a schematic with voltage measurements on it that one
could use for debugging a non-working ubitx. I have the following set
up: ubitx to WM-1 wattmeter to dummy load. I'm using CEC's software
v1.04. I can receive just fine. I've adjusted the bias of the 510's
to 100ma each. I've done calibration and BFO setting. Once again the
radio works fine on receive. I have very low output on all bands. I
have made the following measurements and I'm looking for someone to
help me figure out where the power is going. I'm running at 13.8V
(was at 12V, but thought extra would help the PA). The 12V numbers
are equally bad, just a bit lower. So, I measured from the wiper of
RV1, the power out, and the antenna voltage (on a 50mhz scope) the
following: RV1 at full clockwise: Band ?? RV1 ? ? Power ?? Ant V

80 ? ?? 640mv ?? 4.8W ?? 21V
40 ? ? ?? 520mv ? ? 700mw? 12.2V
30 660mv ? ? 700mw ? 15v 20
1.2V ? ? ? ? 300mw ? 14v 15

640mv ? ?? 7mw ? ? ? 2.8V 10 ? ? ? ? 340mv ? ?? 3mw ? ?? 600mv Moving
RV1 counterclockwise will lower these values as would be expected. Do
these values for RV1 look right? If they do then the problem is
further down the PA chain. Again, I wish there was a table of
receive/transmit voltages that one could reference.

So, then I thought I'd measure the clocks to see if there was
anything amiss there: at 7.150 Clk0 12Mhz 1.44V on receive; 0V on XMT
Clk1? 660mw at 33Mhz; 0V receive
CLK2 1.2V XMT (7.150mhz); 520mv receive 43Mhz

Help would be very much appreciated. Thanks.
73, Mike W1USN

Mike,

I dunno. Those figures sound a little low to me. I don't remember what
mine pulls in CW but I know it is more than an amp. And it is 610ma
with the PTT down on LSB at 3.95Mhz.

470ma is just a little bit more than what mine pulled before I
adjusted the two bias pots for the IRF510's.

You might try checking your bias pots again.

tim ab0wr

On Thu, 22 Mar 2018 18:25:12 -0700
"Mike R." <ubitx@...> wrote:

Hi Tim,
Thanks for the reply. In receive it draws about 160ma. In CW transmit
with RV1 all the way counterclockwise it draws about 1A on 7.150 into
a dummy load. With PTT it is 470ma in LSB. Mike W1USN

Nope.??
They are cheap relays.
As they must be on an all band HF transceiver at this price.

They have been known to fail.
Though to be stuck in 80m, seems all three would have to fail.
Seems unlikely.

Jerry

At 23-03-18, you wrote:

The relays used for the output filters are put together in a clever way. They use three relays to control four different filters. When KT1, KT2 and KT3 are not energized, it selects the 21-30MHz filter. With KT1 energized it bypasses the 21-30MHz filter and selects the 14-21MHz filter. With KT1 and KT2 energized it bypasses both of those filters and selects the 7-10MHz filter. With KT1, KT2 and KT3 energized it bypasses those and selects the 3.5-5MHz filter. The relays are only energized when transmitting (and only when needed).?

You have figured it out. I gave up.. Farhan has his mysterious ways of being thrifty. The real reason is as I think is to save one
data out control line from Raduino.

As Jerry said the relays are known to fail so I cut the TX line to filter ralays and linked it to +12V . The relays remain On and
when you tune across bands you can hear them clicking!. Some more power consumption but I can live with that.

LEDs on the board to show the relay status would have been a nice touch! For V2..

Raj

It is a strange configuration but it does save the expense of a 4th relay.?

Keep in mind now that all the relays are in play when the radio selects an output filter. I had a problem where I had no power out on 30m or 40m. Receive was fine. It ended up that one of the legs on relay KT2 was not soldered. A quick soldering of the pin and I was back in business.?

BTW, they are not latching relays.?

Good luck
Mike M.?
KU4QO

On Thu, Mar 22, 2018 at 9:48 PM, Jerry Gaffke via Groups.Io < jgaffke@...> wrote:

Nope.? ?

They are cheap relays.

As they must be on an all band HF transceiver at this price.

They have been known to fail.

Though to be stuck in 80m, seems all three would have to fail.

Seems unlikely.

Jerry

This wiring scheme is often used for bandswitching relays, keeps the traces short as possible for the upper bands.
? ??

With no relays energized as shown in the uBitx schematic,
RF power from the finals goes through KT1-14 to the 30m lowpass filter then through KT1-3 out to the antenna.
That's as short as it can get.

With KT1 energized and KT2 not, RF power goes through KT1-16 and KT2-14 to go through the 20m lowpass filter.
Still pretty short.

With KT1 and KT2 both energized, we get the 40m lowpass filter through KT3-14.

With KT1,KT2,KT3 all energized, we get the 80m lowpass filter through KT3-16.
That's getting to be a pretty long path for that rf, going through all three relays.
But at 4mhz it doesn't matter.


The most obvious way to implement this would be to have four relays wired to the IRF510's in parallel,
only one relay on at a time.? But then you have a trace routed off to each relay all the time, even when on 30m.
On both the IRF510 side and on the antenna side.? That's a significant bit of extra loading.

If we had lots and lots of bands, a binary tree on the input and another on the output would reduce the relay count
but clearly adds to trace lengths on those critical high bands.

What we got is as good as it gets for minimal traces on 30m.

Jerry

Finally got around to working on the ubitx again.
I traced the logic for each of the three lines (TXA,TXB, and TXC) to ensure the logic was operating as it should. They were. I then traced through each of the relays and filters to see if and where the signal might be getting attenuated. KT1 looked somewhat strange on the scope when going into transmit - the signal started out strong but soon went very low. Hmmm.
I then went and re-heated the pins of each of the three relays (KT1, 2, and3). I also re-heated some of the leads on some of the components in the 4 filters that looked dull to me. I then went back and had another look at the bias to the 510's. Both were set to 100ma each. Here are the results:
Band ?? Current ?? Wattage
80 ? ? ? ?? 1.84A ? ?? 9W
60 ? ? ? ?? 1.46A ? ?? 6W
40 ? ? ? ?? 1.26A ? ?? 7.8W
30 ? ? ? ?? 1.26A ? ?? 4.2W
20 ? ? ? ?? 1.61A ? ?? 5.5W
17 ? ? ? ?? 1.25A ? ?? 3W
15 ? ? ? ?? 1.07A ? ?? 2W
12 ? ? ? ?? 1.3A ? ? ?? 2W
10 ? ? ? ?? 1.0A ? ? ?? 1.5W
Judging by others results, it doesn't look like I can get much more out of the rig as it stands without replacing finals, etc.
It looks like my rig is now "normal" for the ubitx.
That seem like a fair conclusion?
Thanks to all for the help, now to start playing with larger displays, new finals, new software, etc. It never ends!
73, Mike W1USN

Very cool, good job!

If it was a bad solder joint in one of the lowpass filters, then only one of the 4 bands
would have failed to pass a signal.? So that probably was not the problem.
If all relays were permanently on we would only get the 80m lowpass filter and thus
none of the upper bands would get out.? But that would require 3 separate failures,
I don't see a single point of failure for this to happen.

If any relays are stuck off (perhaps the coil is burnt out) then that relay is stuck in the position it is drawn
in the uBitx schematic.? We would get a lowpass filter for one of the bands above the operating frequency.
That lowpass filter will? let our lower band through just fine and things will seem to operate normally.? ?
But not quite normally enough, as the rig will be transmitting?harmonics as well.?
Most of us would never notice till perhaps that letter from the FCC arrives.
Speaks for the need of a good cheap HF spectrum analyzer.

Jerry, KE7ER

Mike,

I spent all afternoon with my spectrum analyzer, a two-tone generator,
and my ubitx and came up with 10watts on 80m at 1.4amp total current
draw being the cleanest signal I could get.

I didn't chk 40m but on 20m I could only get 2 watts out.

I tried every combination of PA bias and drive to mimimize 3rd order
IMD products. I could crank the power output up further but 10watt out
on 80m seemed to be the cleanest.

I wonder if running higher voltage on the IRF510's would help with the
3rd order IMD so you could be more clean power out? Or will you wind up
saturating the driver stages trying to drive the PA to get more power
out and thus make the 3rd order IMD worse?

tim ab0wr







On Sat, 24 Mar 2018 13:53:09 -0700
"Mike R." <ubitx@...> wrote:

Finally got around to working on the ubitx again.
I traced the logic for each of the three lines (TXA,TXB, and TXC) to
ensure the logic was operating as it should. They were. I then traced
through each of the relays and filters to see if and where the signal
might be getting attenuated. KT1 looked somewhat strange on the scope
when going into transmit - the signal started out strong but soon
went very low. Hmmm. I then went and re-heated the pins of each of
the three relays (KT1, 2, and3). I also re-heated some of the leads
on some of the components in the 4 filters that looked dull to me. I
then went back and had another look at the bias to the 510's. Both
were set to 100ma each. Here are the results: Band ?? Current
Wattage 80 ? ? ? ?? 1.84A ? ?? 9W 60 ? ? ? ?? 1.46A ? ?? 6W 40
?? 1.26A ? ?? 7.8W 30 ? ? ? ?? 1.26A ? ?? 4.2W 20 ? ? ? ?? 1.61A
5.5W 17 ? ? ? ?? 1.25A ? ?? 3W 15 ? ? ? ?? 1.07A ? ?? 2W 12
1.3A ? ? ?? 2W 10 ? ? ? ?? 1.0A ? ? ?? 1.5W Judging by others
results, it doesn't look like I can get much more out of the rig as
it stands without replacing finals, etc. It looks like my rig is now
"normal" for the ubitx. That seem like a fair conclusion? Thanks to
all for the help, now to start playing with larger displays, new
finals, new software, etc. It never ends! 73, Mike W1USN

Satish,
That is a good idea. At some point I'll look for a voltage booster to run the 510's at a higher voltage.
Thanks,
Mike

Hi Tim,
I did look at the outputs on my scope and they looked like nice even sine waves. Not an absolute measurement of course.
Using CEC's software I could go to 160M and look at that waveform. It was awful! No output filtering at all as the ubitx lowest filter is 3.5 - 5 Mhz. A lot of garbage in that waveform. Lots of power though. I'll have to go back and measure the power out on 160, it had to be in excess of 20W.
Mike W1USN

I just finished assembling my uBitx and get what I think is low power output.? I tried two different power meters to verify my results.? Measurements were taken in CW with a dummy load and 13.8VDC.

80M - 11.0W
40M - 7.3W
20M - 3.4W
10M - 0.7W

Is this normal power output?? I hear relays clicking when I change bands..

Also, in SSB mode, I get very weak power.? Unless I blow into the mic, I get no needle deflection on my peak-reading power meter.? I see a few threads about changing some resistors.? I'm using the stock mic element however and don't expect to need this.? Any comments please?

Rich - KE1EV