Please Help - K2 Loss of SSB Transmit

[VK1OO, Mike Walkington]

Hi,

My K2 seems to have lost power out in SSB mode. As far as I can tell,
the RX is still working OK in all modes. The inbuilt ATU seems to tune
OK and there is power out in CW. When I hold the MH2 PTT in, there is no
"click", but the RX goes silent, the RF/ALC display stays blank in both
RF and ALC modes ie there are no bars when I speak or whistle into the mike.

Any suggestions would be appreciated.

Mike
VK1OO
K2: 2599
______________________________________________________________
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

[Jeff Cochrane - VK4XA]

CONTENTS DELETED

The author has deleted this message.

[VK1OO, Mike Walkington]

Hi Jeff,

Thanks for this suggestion. I don't have another mike to try, but I've
checked, and the PTT is working OK.

Mike
VK1OO



On 27/06/2010 6:42 PM, Jeff Cochrane - VK4BOF wrote:

> Mike,
> Have you tried another microphone?
> If the rig checks out ok with another microphone, check the PTT switch
> in the MH2, mines a bit 'ordinary' and occasionally does not make
> contact when the PTT button is pressed.
> Jeff Cochrane
> VK4BOF
> Innisfail QLD
> Australia
> Elecraft K3 #4257
>
>     ----- Original Message -----
>     *From:* VK1OO <mailto:[hidden email]>
>     *To:* Reflector Elecraft <mailto:[hidden email]>
>     *Sent:* Sunday, June 27, 2010 6:28 PM
>     *Subject:* [Elecraft] Please Help - K2 Loss of SSB Transmit
>
>     Hi,
>
>     My K2 seems to have lost power out in SSB mode. As far as I can tell,
>     the RX is still working OK in all modes. The inbuilt ATU seems to
>     tune
>     OK and there is power out in CW. When I hold the MH2 PTT in, there
>     is no
>     "click", but the RX goes silent, the RF/ALC display stays blank in
>     both
>     RF and ALC modes ie there are no bars when I speak or whistle into
>     the mike.
>
>     Any suggestions would be appreciated.
>
>     Mike
>     VK1OO
>     K2: 2599
>     ______________________________________________________________
>     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
>

______________________________________________________________
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

[VK1OO, Mike Walkington]

Hi,

I believe I've now solved this problem with the excellent help provided by Don W3FPR, Gary AB7MY, and Andy VK2ZAM.  In the end, the solution was relatively simple - wick up the solder on T1 and T2 and resolder them. Go figure why this was needed after 8 years.

For the record (and in response to a recent email on the reflector), I have attached the lengthy off-reflector email trail as there is lots of useful information in them. In summary, the following four paths needed testing:
(1) microphone output through speech compressor (U3) to balanced modulator (U5),
(2) BFO output (on main K2 board) buffered through Q3 to balanced modulator (U5),
(3) balanced modulator output through crystal filter through Q2 to amplifier (on main K2 board)
(4) ALC circuit.

The problem was in (3) but all the initial symptons pointed to (4) or (1).

Mike
VK1OO (ex VK1KCK, AC7MZ)
K2: 2599


-------------------------------------------------------------------
Email thread with Don, W3FPR

28 Jun 10
Hi Don,
Just wondering if your expertise can help me solve this one..

My K2 seems to have lost power out in SSB mode. As far as I can tell, the RX is still working OK in all modes. The inbuilt ATU seems to tune OK and there is power out in CW. When I hold the MH2 PTT in, there is no "click", but the RX goes silent, the RF/ALC display stays blank in both RF and ALC modes ie there are no bars when I speak or whistle into the mike.

Any suggestions would be appreciated.

Mike
VK1OO
K2: 2599
   

29 Jun 10
Mike,
Well there are several possibilities.  Without audio input, the SSB power out will be almost zero.
1) Are you certain the microphone is working?
2) Is the mike bias resistor properly in place?  You can check for 5 volts on pin 1 of the mike jack (pin 1 is the first pin to the left of the nub at the top of the jack.
3) There is a break in the microphone input path to the KSB2 board.
4) The KSB2 board has a problem.

If you have a 'scope, you should be able to see the audio entering the KSB2 board at P1 pin 5, and you should also see some response to mike audio at U3 pin 4.

Again if you have a 'scope, you should see some RF being generated as you speak into the mike at the balanced modulator output U5 pin 4 or check at D4 cathode.

If you don't have a 'scope, analysis will be difficult - can you borrow one?

73,
Don W3FPR


30 Jun 10
Hi Don,
Well, as expected a CRO is most illuminating.  I'm definitely getting audio through the mike and varying but very low power output on all bands.  40m seems to have the highest level of output, about half of the expected 8W I had the power output set for.  The majority of the other bands were minimal.  Is there likely to be a problem with the VRFDET circuit?  How would you recommend I check it out? I'm now assuming from googling a few archived emails that I've somehow fried MCU Pin 2 (RA0).

Mike
VK1OO
K2: 2599


30 Jun 10
Mike,
The source for the VRFDET circuit is working if you have proper power control on CW (VRFDET is generated by RF Board D9 or by the wattmeter in the KAT2).  However, in SSB, that signal changes the conductivity of KSB2 Q1 to set the level at ALC THR.  The adaptive ALC operation is actually rather complex and subject to a number of variables that are difficult to control for a test.
I would check KSB2 Q1, the soldering at D1, D2, R4, R7, R8 and R9 as well as the control to the BFO injection level (Q2).  It may be easiest to simply put new 2N2222s in Q1 and Q2, they are inexpensive.
If that does not reveal anything, then it is time to try a new SSB firmware chip.

73,
Don W3FPR


1 Jul 10
Don,
Thanks for the leads Don. I think I also need to go back and have a closer look at the CW output.

Mike


1 Jul 10
Mike,
Using your 'scope across the dummy load,  compare the peaks on SSB with the RF level that is present in CW.  If the peaks come up to the CW RF level, that is about all that can be expected, the average may be increased by increasing the compression setting.  If the peaks are not as high as the CW RF, then try increasing the AF Gain (menu SSBA parameter).

73,
Don W3FPR


3 Jul 10
Hi Don,
I'm getting warmer (well not physically as its winter here in Canberra and the temperature is in the single digit degree Celsius range).

The weekend has allowed me to spend some time troubleshooting.
I removed the KAT2, KSB2 and KIO2 and reverted the K2 back to a basic CW rig. I added appropriate jumpers and cap in the KSB2 holes.

I confirmed that the basic rig functions OK, with power control working as expected. So, I believe this means the VRFDET is OK as is MCU Pin 2.
I installed KAT2 and worked through the installation checks. It seems to be OK, including measurement of FWD and REF power.

I installed KSB2 (without KAT2) and confirmed with CRO that previous observations ie SSB peaks are well down from CW level at the same power.  SSB peaks vary significantly by band.  So, as discussed, this is the area of concern.

I haven't replaced any parts yet as you suggested. I'll look at this tomorrow.  In the meantime, I have a previously removed SSB chip (version 1.04). Should I try installing this in place of the one on the board (version 1.08b)? Am I likely to damage this one if there are unserviceable parts? Is the existing one likely to have been damaged?
The KSB2 had been working OK up until it wasn't.  What might have caused the issues you suggest I look for with" KSB2 Q1, the soldering at D1, D2, R4, R7, R8 and R9 as well as the control to the BFO injection level (Q2)? "

I am looking forward to getting to the bottom of this and I'm pretty sure I will be confirming the issue is exactly where you think it is.

Cheers and thanks for the help.

Mike
VK1OO
K2: 2599


3 Jul 10
Mike,
Yes, if you have an old firmware chip, try it to confirm or deny that the firmware chip ALC output is or is not the problem - that is a qucik check,  Just use anti-static measures when changing the chip - dry air and static are problems in the wintertime.

73,
Don W3FPR


4 Jul 10
Hi Don,
The troubleshooting continues.

Changing the firmware chip did not change the erratic power behaviour. I then replaced D1 and D2, no change. Then Q1 and Q2, no change.

The SSB peaks get close to the CW peak on 40m and 30m but are significantly lower on the other bands.  Also, on the other bands, the SSB peaks appear to stay constant regardless of the output level.

Any further avenues for exploration would be appreciated.


Mike


4 Jul 10
Mike,
About all I can think of right now is that you might have a microphone which has developed lower output.
What is the possibility that you can try another one?  The element in that microphone is an electret type (requiring the bias resistor).  If you have no other microphone available and cannot borrow another MH2, do you have an 8 pin Foster plug and one of the inexpensive computer microphones around?  Cut the plug off the computer mic (or build up a 3.5mm stereo jack to Foster plug adapter) and connect the mic AF to pin 1, and the return to pin 7.  Connect a key or paddles to the K2 and close the dot contact for PTT.  That will allow you to tell if the MH2 mic has developed a problem - if it works OK with the computer mic, then the MH2 element has gone bad on you.

You might also check the amount of compression that you are using.  I normally find SSBC = 3-1 is about right for most normal use.  For data modes, it must be 1-1, but the RTTY filter set allows independent setting of the compression.

The TX gain through the RF chain is higher on 40 meters than for the other bands, so that explains why the RF from SSB can reach the CW peaks on those bands - that part is normal.

73,
Don W3FPR



Email thread with Gary, AB7MY

8 Jul 10
Hi Mike,
You need to first be sure the CW and Tune RF output is good on all bands. 14-15w on 80-20m, and 10w or more on 15-10m is expected with a 13-14VDC power supply and a good dummy load.

Test only the basic K2 without a KAT2 or other option like the KPA100, KAT100 until you get this solved.

See if the rig will produce about 1w of carrier or more on 40m when the R1 carrier null pot is purposely set fully CW or CCW. This tests that the carrier injection is good. Re-null R1 when done with the test.

Then check to see if T1 and T2 might be wound wrong, swapped with each other, or turned around wrong on the KSB2 board. This is a common error and the manual shows the correct procedure. Also see the attached pics.

The MH2 should drive the rig well on all bands with SSBA=2 and SSBC=2-1. The 5.6K mic bias resistor must be on pins 1-6 of the front panel mic connector (on the rear of the pins on the PCB).

Look over the KSB2 very, very, closely to be sure there are no assembly or soldering errors. A bad part is very rare.

If the KSB2 MCU has been damaged, or Q1, from an error with the cable to the KAT2 or KPA100 options (if installed), those parts may need to be replaced. See how the ALC works below.

The KSB2 ALC works this way:

Normal ALC action is about 1-5 bars right to left on the bargraph, and it is highest on 80-40m and lowest on 12-10m.

Check the ALC Threshold at U2 pin 5 on the KSB2. At power on or when the mode was changed to CW and back to LSB or USB, U2-5 should be about 6VDC until you press the PTT button. It then should change from 0.5 to about 4.0VDC when the Power knob is moved from min to max while the PTT button is held down.

Damage to U1 pin 2 (U1-2) or a problem with the ALC level from Q1's collector to U1 pin 2 (ALC input) on the KSB2 MCU can cause a lack of ALC. The voltage at U1-2 goes down to about 4.8 to 5.0VDC on loud voice peaks, and then rises back close to 6.0VDC when there is no mic input and thus no SSB RF output. U1-2 monitors this DC level and displays it on the LED bargraph.

The SSB ALC is really pretty simple, and just depends on the 0.5 to 4.0VDC ALC Threshold applied to Q1's emitter from U1 and U2-5, which is then compared to VRFDET on Q1's base. VRFDET comes from the simple D9 RF detector in a basic K2, or from the KAT2 or KPA100 if they are installed.

Q1 then conducts when VRFDET is 0.6VDC higher than the ALC Threshold on Q1's emitter, and that allows current to flow through D2 to shunt away a portion of the SSB signal applied to the base of Q2. Q2 buffers the SSB output back to the U10 TX Mixer on the RF Board. Q3 is used only when in CW or TUNE mode, to buffer and pass the BFO signal from P1-12 to P1-7, effectively bypassing the BFO carrier signal around all of the other parts on the KSB2. The D1 circuit is a separate time delay ALC path directly controlled by U1, and used only when needed to help long-term ALC action or when D2 is unable to maintain control of the RF output (referred to as tracking ALC, and is sometimes used between speech syllables).

The BFO settings for CAL FIL are also important for good SSB operation. Use the KSB2 manual numbers for the crystals that they came in, or better yet - use Spectrogram or some audio analysis software and a PC's sound input to find the best settings.

Here are some sample BFO settings from a K2 I once repaired and aligned. Yours should  be close to these:

LSB

FL1    4913.60
FL2    4913.69
FL3    4913.20
FL4    4913.22

USB  
FL1    4916.32
FL2    4916.26
FL3    4915.37
FL4    4915.19

CW normal

FL1    4913.35
FL2    4913.22
FL3    4913.15
FL4    4913.10

CW reverse

FL1    4915.00
FL2    4914.46
FL3    4914.37
FL4    4914.31

--
73, Gary AB7MY


8 Jul 10
Hi Gary,
Thanks for the help.
Basic K2 in CW checks out OK.
Not sure how to measure the 1W of carrier. But full CW or CCW lights up 7 bars when SSBA is set to BAL, and reduces to zero bars in the about the middle of the R1 range.
I read your SSB Carrier Null notes from the mod webpage, but again, not sure how you set up the K2 to use an external QRP wattmeter to measure the carrier null.
I couldn't see any problems with T1 and T2 and assume they are OK, and the KSB2 has worked OK for the last few years.
MH2 isn't driving on all bands. Almost none.  Most output is on 40m and 30m, but minimal.
ALC Threshold varies as you described.
U1-2 seems to be stuck on 6V at least on 40m despite loud whistles into the mike.

I had previously swapped in an old U1 chip which made no difference to transmit level and then replaced Q1.  Could a fault with Q1 have fried U1-2? ie I may have fried both U1s as I replaced Q1 after trying both.
 How do I confirm if I still have a Q1 problem or have a U1 problem?  Am I on the right track in thinking this?

Mike


9 Jul 10
Have you tried another mic to see if that might be the problem?

A Kenwood mic with an 8 pin connector will work with no changes.

40m has the higher transmitter gain, so if the mic output is low, the 40m band would be the band that worked the best. 10m has the lowest transmitter gain. If there is not enough mic output, the effect of that would be seen most on 40m (good) and 10m (bad).

I use either an oscilliscope or a OHR WM-2 QRP wattmeter to measure the actual carrier output when the R1 carrier null pot is set fully one way or the other (unbalanced) on 40m. You could also use the RF probe included in the K2 kit, and then convert its rms RF readings to power (square the reading, then divide by the 50 ohm dummy load resistance). A properly working K2 will have 1w or more of carrier with the carrier null pot set to min or max.

At U3 pin 7 of the preamp/compressor IC on the KSB2, there should be 100-200mV rms of mic audio signal when you whistle into the mic or shout loudly.  Set SSBA to 3 and SSBC to 4-1 (both maximum settings) to do this test.

If not already done, reheat all connections on the KSB2 even if they look OK. Be sure there are no poorly tinned leads on the toroids, and they are all soldered very well.

Be sure the wires on RFC1, RFC2 are not touching and shorting against the sides of the crystals.

U1-2's 6V static ALC level will not go down unless the RF output starts to exceed what is set by the Power knob. If there is too low of RF output for whatever reason - low mic output or low BFO level to the U5 balanced modulator, the ALC may never work. In this case, you will have difficulty determining if Q1 or U1 are damaged. Q1 will only conduct and pull down its collector and the U1-2 level, when the DC level on its base exceeds what is set on its emitter by U2-5 by about 0.7VDC (normal BE junction barrier voltage). So, at lower Power knob settings, Q1 will start to conduct sooner if the VRFDET level from the transmitter RF output increases the Q1 base level by that amount.

Be *sure* T1 and T2 are correct, as I see lots of K2s in service that are not. Use the photos or the KSB2 manual to verify this.

--
73, Gary AB7MY


9 Jul 10
Hi Gary,

Thanks for your suggestions. Re your questions:

I've tried another MH2 - the behaviour was the same.

There seems to be plenty of carrier output, but I'm still not sure I'm measuring this right. What mode (CW or SSB) should I be doing this measurement in?  Does the filter or output power setting matter? If in SSB, do I talk into the mike or not?

U3-7 shows 200mV output across all bands when I whistle into mike.
All KSB2 soldering is good.  RFC1, RFC2 are not touching and shorting against the sides of the crystals. T1 and T2 are correct and match the pictures you sent, ie green turns and orientation.
VRFDET varies, ALC Threshold varies.

I reset all the BFO settings. They match the manual.

I checked the mic plug voltages and the bias resistor - all seem OK.

Where next?  The balanced modulator?

Mike
VK1OO


10 Jul 10
For testing the carrier level, you must be in either USB or LSB and on 40m, then press the PTT button.

When done, redo the R1 carrier null adjustment for minimum carrier.

If that looks OK, then use a scope or an AC voltmeter to see if there is mic audio coming out of U4 pin 7 to the input of the balanced modulator U5 pin 1. Anything from 50mV to 100-200mV when you whistle into the mic shows there is audio there.

Then look at U5 pin 6 in USB or LSB with PTT closed, and see if there is around 150mVrms of BFO carrier injection.

Then look at U5 pin 4 or the top of D4 and see if there is 100-200mVrms of signal when you whistle into the mic. This is DSB at this point, or double sideband suppressed carrier before the DSB signal passes through the crystal filter and is stripped of the opposite sideband. You can follow this signal through that path with a scope, RF probe, or by listening on another rig with a wire from its antenna held near the signal path, though it does get weaker and weaker as it goes through the crystals due to their normal signal loss. Once the SSB is formed it then goes into Q2 to be buffered and sent to the U10 TX Mixer on the RF board.

--
73, Gary AB7MY


10 Jul 10
Gary,

Thanks. OK, at one extremity of R1 with a CRO I measure the RF output to be 40mV and at the other 75mV both peak. Clearly not at least 1W. This seems to be the same regardless of what output power I set.
Separately, when whistling into  the mike there does appear to be 200mV of audio from U3 pin 7, the output of the speech compressor.

Does this mean there is a carrier injection problem and what next?


Mike


13 Jul 10
Did you measure with the scope at U5 pin 6 to see if there is BFO signal there, in SSB transmit with the PTT button closed?

"Then look at U5 pin 6 in USB or LSB with PTT closed, and see if there is around 150mVrms of BFO carrier injection."

--
73, Gary AB7MY


13 Jul 10
Hi Gary,

The frustration continues, but thanks for your ongoing help.
I had a friendly VK2 amateur come over on Sunday afternoon.  He's an ex Army radio tech and fairly knowledgeable.  We reworked through the circuit and he couldn't find anything obvious either.  Everything was broadly working as you have indicated it should in past emails, other than with the carrier unbalanced there isn't 1W of carrier output.

Here's the latest measurements around U5.

All voltages are peak

The following measurements for U5 pin 6 with carrier inject balanced:
40m and 20m USB = 300mV, LSB = 600mV
15m and 10m USB = 600mV, LSB = 300mV  (ie reversed from lower freq bands)
Voltages on the BFO crystals on the main K2 board have a similar ratio

The following measurements were for 40m, output power set to 10W, constant whistle into mike:
U5 pin 1 =  200mV, U5 pin 4 = 1Volt,  RF Output = 20V on LSB, 5W indicated on external power meter.

The following measurements were for 10m, output power set to 10W, constant whistle into mike:
U5 pin 1 =  200mV, U5 pin 4 = 0.5Volt,  RF Output = 0.8V on LSB,  no indication on external power meter.

If I remove the ALC diodes and keep the output power set low will I be able to see if the ALC is the issue?
If this is a useful approach, what is the maximum output power that I should set.


Cheers
Mike


14 Jul 10
The ALC is OK if there is not a reduction in the DC level on Q1's collector when the power is set to maximum and you whistle into the mic but the RF output is still low.

Instead, look for an incorrect or damage cap in the crystal filter, or a bad connection in that area. All of the crystals on the KSB2 are a matched set and they cannot be mixed with other crystals in the kit.

The RF voltages are too low to measure in the IF path from the output of the U5 balanced mixer through the crystal filter. But, you can often probe along that path with a wire or a test probe used as antenna to pick up noise or stray signals, to see if there is a point where the signal is suddenly lost by a great deal, instead of the normal small loss of signal through each crystal. Or use a noise generator or signal generator set to 4.915MHz to do the testing.

In many, many, cases, I have fixed KSB2 problems just by cleaning off excess solder and flux, then reheating resoldering all of the connections. That should be done if you have not tried it yet. It takes less time than you think since all of the parts are now on the board compared to when you were building it. Clean off excess solder with desoldering braid, and clean the board with cotton swabs moistened with acetone or alcohol.

It is important to set the BFO in CAL FIL too, so the BFO is set properly in relation to the passband of the crystal filter so there is not too much loss in USB or LSB transmit. Use the numbers on the envelope that the KSB2 crystals came in, as on page 10 of the KSB2 manual. Or experiment some with the FL1 BFO freq. and see if there is an improvement.

It is normal for there to be slightly less BFO injection in LSB or USB due to the way the varactors load the BFO. 300-600mV is enough at U5 pin 6. The BFO freq. is reversed on 15-10m due to how the mixing scheme works.

--
73, Gary AB7MY


17 Jul 10
Hi Gary,

I think its fixed.  I did as you suggested and sucked off all the solder starting around T1 and T2 and resoldered. Back to life.  I still have no idea why I would need to have done this after 8 years of happy operation. But at least it seems solved.
Thanks for your careful and insightful help.

Mike
VK1OO