Re: {Collins} KWM-2 no output power PROBLEM SOLVED!



This is how the problem with erratic drive and output on my KWM-2 was finally solved. 

First I tried the classic remedies such as loosening and tightening screws, cleaning switches, cleaning tube sockets, and changing tubes. I then measured voltages and resistance from V3 to V10. The measurements indicated that the problem was at the early stage of the transmit circuit, but it was only when I started tracing the signal with an oscilloscope from the BFO to V4 and through the mechanical filter to the 1st mixer (V5), that I managed to locate where the  troubles began. With a scope and a SG set at 450 kHz I checked the mechanical filter, and it was fine, so the problem could be narrowed down to V4 and/or the balanced modulator. 

When I resoldered the four diodes in the balanced modulator, and the RF signal came back, I assumed that badly soldered diodes were the cause of intermittent / erratic drive. Some colleagues in the local club suggested that the heat from the solder iron had revitalized the diodes. But as the problem came back after a few days, we proved wrong. I suspected the diodes, but before they were replaced, the problem was finally solved. While observing the erratic reflections on the grid meter in TUNE mode and gently poking at the various leads around V4 and the balanced modulator, I found the culprit: a bad solder where the conducting capacitor (C8) from the BFO connects to the wiper of the carrier balance pot (R15). This time there was no doubt. Here was hard core evidence and the explanation of the no output problem that had been haunting my KWM-2 for a long time. I resoldered the cap, and the curse was finally lifted!

Thanks to all of you for your great help.

Peter OZ8CTH

-----Oprindelig meddelelse-----
Fra: Peter Ravn [mailto:fasan5@xxxxxxxxx] 
Sendt: 16. juli 2016 15:09
Til: 'Gary J FollettDukes HiFi'; 'Bill Carns'
Cc: 'COLLINS@xxxxxxxxxxxxx'
Emne: SV: {Collins} KWM-2 no output power

When I came back after a week’s holiday, for some mysterious reason I could tune up my KWM-2 to 60 watts, and I had my first QSO. But the joy was short lived. To increase output power I did the PA neutralizing per manual, and after that new troubles began: little drive and output about 30 watts, on some bands no output at all. This is the situation  right now.

Here is the 2nd update on the initial output problem.

1.The present PA neutralizing circuit with the infamous ceramic trimmer (C120) seems ok. When lifted R127, C180 and C183 are according to specs. During neutralizing the trimmer produced the expected dip on a scope connected to antenna jacket (J1). When mounted and adjusted I see 28 pF across C120. That being said, the ceramic trimmer will be replaced as soon as I have got the cap kit from W7SKG.

2.The PA tank circuit seems fine too. C149 doesn’t look toasted, nor is it shorted. Capacity is according to specs. L17 is 7 ohms and C148 is according to specs. There is continuity between the junction of C149 and C150 and the antenna jacket when the antenna relay (K3) is closed. So there is no relay issue either.

3.But there is definitely a drive problem. With MIKE GAIN at 12 o’clock and a scope on the control grids of the finals, I only see 10 V peak. There is also 10 V peak on pin 6 of the driver (V8), so there is no loss because of a bad C127, which by the way is according to specs. In transmit voltages on the plate, screen and grid are according to specs, but the cathode voltage is almost non-existing in transmit. According to the manual it should be 4 V (?). L23 in the driver cathode should be 22 ohms; when lifted it is 1.1 ohms and thus actually shorted. Could he be the culprit? The driver valve is new, and C260 and R106 are fine.

Peter OZ8CTH

-----Oprindelig meddelelse-----
Fra: Gary J FollettDukes HiFi [mailto:dukeshifi@xxxxxxxxxxx]
Sendt: 11. juli 2016 17:11
Til: Bill Carns
Cc: fasan5@xxxxxxxxx; COLLINS@xxxxxxxxxxxxx
Emne: Re: {Collins} KWM-2 no output power

One more thing: If the plate blocking capacitor were “open”, a condition that would present the issue you see,  it likely would show some physical damage because the nature of the capacitor does not lend itself to “open circuit” failure by means other than catastrophic damage, which is usually visible. In any case, the RF generator test would show this condition. A 14 MHz applied signal to the plate cap of one of the 6146’s will pass through this capacitor with virtually no attenuation regardless of the status of the plate and load capacitors (unless one to the other is shorted via bent plates).

If that blocking capacitor were “shorted”, the power supply fuse should open because L21, 2.5 mH RFC, would pull the resultant 800 volts on the plate coil to ground. 

Also, if the plates of the load capacitor were bent or had arc damage, this would cause a condition in which no output would appear at the antenna output. Open this capacitor all the way (fully clockwise) and then determine if you can get output on one of the higher bands (20 meters, preferably) when tuning the plate capacitor. You won’t get a lot of power but you should get some power output.

If the plate RF choke had many  shorted turns (most of them) from damage from the arcing tubes, you would expect to see a LOT of plate current as the RF path would be through the choke L17 to the plate bypass capacitor C148. Likely this is to your problem.

Gary




> On Jul 11, 2016, at 9:11 AM, Gary J FollettDukes HiFi <dukeshifi@xxxxxxxxxxx> wrote:
> 
> Many have asked how much drive there is on the grids of the 6146 tubes. However, that level would need to be very low to produce ZERO output at the antenna jack. Determine if the output is ZERO by putting the oscilloscope at the antenna jack with a 50 Ohm load connected. Determine if going through the tuning range of the plat capacitor causes a peak in the output (leave the load capacitor fully meshed and the drive control set to about 11 o’clock and preselector properly tuned in receive). You should see a very strong peak in the output, however small, in this test, and you can do this test even if the output is too low to see with the platesetter or an external wattmeter. The frequency you read from your scope should agree with the frequency selected with the PTO and band setting.
> 
> Depending on the quality of your scope, I would work on 20 meters as this does not challenge most scopes and yet still would allow SOME signal to pass through capacitors that might be bad.
> 
> I have experienced this failure mode twice (two different failed parts) and use the same method to find it. The radio had already failed the test above.
> 
> The method I employed to find it was to disconnect the power supply altogether from the radio, remove the 6146 tubes, and insert a signal generator signal right at the plates of the 6146’s. I then traced the signal through the RF tank circuit with an oscilloscope (you need to force the antenna relay inside the PA cage closed to do this). Then you look for output at the antenna jack with not load connected. There should be SOME output seen and tuning the plate and load capacitors should produce a peak, even though it will not be in the right place on the scales due to impedance mismatches. 
> 
> One time I found a condition wherein the double pole double throw antenna relay was not closing even though the other two main T/R relays were closing just fine. If that hidden relay has an open coil, the receiver will work completely normally but you will see not TX output at all.
> 
> The other time I found that the neutralizing capacitor was shorted, 
> but only in some positions. Its plates were bent, causing this 
> behavior. I had a new one and that corrected the problem
> 
> In this case, that capacitor appears to be OK. as evidenced by the measurements, but a shorted C180, C183 or C270 would produce the same result.
> 
> A shorted C180 or C270 would be unlikely as this would apply 800 volt plate voltage to the cathode of the 6CL6 and this would demolish L22 and/or R106 and put the 6CL6 cathode at 800 volts. I think you would have other problems if this were the case.
> 
> In order, I would check the DPDT antenna relay, then C183. If neither of these are at fault, use the signal generator/scope tests.
> 
> Gary
> 
>> On Jul 11, 2016, at 8:39 AM, Bill Carns <wcarns@xxxxxxxxxxxxx> wrote:
>> 
>> Good deal Peter…   If you have a grid dip meter,  see if that network is working by running the dipper on it.  That will give us another indication.  I am pretty sure that something is shorted in there. That output stage is acting just like it is completely healthy from a DC standpoint, but shorted out from an AC or RF viewpoint. Either that, or the blocking capacitor is completely open.
>> 
>> 
>> 
>> Open the network at the juncture of C149 and L17 and then:
>> 
>> 
>> 
>> 1.        Take another close look at L17 and make sure the DC resistance is 5 ohms or thereabouts
>> 
>> 2.       Measure C149 (blocking and make sure it is still OK and then, while disconnected from the PA Stage, run the dipper on the network and make sure it will resonate.  If it will not, then the culprit is definitely in there somewhere. 
>> 
>> 3.       Take a critical look at those loading padder capacitors (C152 thru C154).
>> 
>> 
>> 
>> Did you ever tell me if you had checked the PA on all bands??  Is it doing this on all bands and did you tell me what the RF grid voltage is on the PA driver (V8) at full up RF drive??
>> 
>> 
>> 
>> Regards,
>> 
>> 
>> 
>> Bill
>> 
>> 
>> 
>> From: fasan5@xxxxxxxxx [mailto:fasan5@xxxxxxxxx]
>> Sent: Monday, July 11, 2016 3:03 AM
>> To: Glen Zook <gzook@xxxxxxxxx>
>> Cc: 'Alan Chandler' <alanchandler2@xxxxxxxxxxxx>; 'Bill Carns' 
>> <wcarns@xxxxxxxxxxxxx>; collins @ listserve . com 
>> <collins@xxxxxxxxxxxxx>
>> Subject: Re: {Collins} KWM-2 no output power
>> 
>> 
>> 
>> I have ordered the KWM-2 neutralizing kit from W7KSG and will let you know the result when it is installed. Until then I will continue the search for an RF shortage in the PA circuit. As already said, the infamous ceramic neutralizing capacitor is neither shorted or open and capacity is according to specs.
>> Peter, OZ8CTH
>> 
>> ----- Original meddelelse -----
>> 
>> Fra: Glen Zook <gzook@ya <mailto:gzook@xxxxxxxxx>   That final stage is acting like it is working DC wise but completely shorted out from an AC or RF hoo.com>
>> Til: Peter Ravn <fasan5@xxxxxxxxx <mailto:fasan5@xxxxxxxxx> >, 'Alan 
>> Chandler' <alanchandler2@xxxxxxxxxxxx 
>> <mailto:alanchandler2@xxxxxxxxxxxx> >, 'Bill Carns'
>> <wcarns@xxxxxxxxxxxxx <mailto:wcarns@xxxxxxxxxxxxx> >, 
>> collins@xxxxxxxxxxxxx <mailto:collins@xxxxxxxxxxxxx> 
>> <collins@xxxxxxxxxxxxx <mailto:collins@xxxxxxxxxxxxx> >
>> Dato: Lør, 09. jul 2016 16:45
>> Emne: Re: {Collins} KWM-2 no output power
>> 
>> Unfortunately, the ceramic trimmer type neutralization circuit units should NEVER be used with the 6146B/8298A or 6146W tubes made after 1964.  There is an extremely good chance that the trimmer will fail if those tubes are used.  Collins had to come up with the air variable neutralization circuit when the military changed the specifications for the 6146W tubes from that of the 6146A/8298 to the 6146B/8298A in 1965.
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> When the neutralization circuit was changed, Collins did have a field modification kit to change the neutralization circuit to the air variable circuitry.  There are 3-components that have to be changed when the air variable neutralization circuit is installed.  Checking the schematic of the earlier version with a schematic of a later version will show the changes.  The value of the air variable is not that critical and one with 10 pf maximum will work and those variables are considerably more common than the value that was used by Collins.
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> Since 6146B/8298A tubes were in the unit when it failed, I would suggest looking at the neutralization circuit for problems and to update that circuit to the "new" configuration.  When the new neutralization circuit is installed, then the original 6146, the 6146A/8298, 6146B/8298A, 6146W, or 6293 tubes can be used.  You can "mix" an original 6146 with a 6146A/8298 or 6293.  However, do NOT mix a 6146B/8298A or 6146W tubes with any of the other versions.  You WILL have all sorts of problems if that is done.
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> The air variable rotor has to be insulated from the chassis and there are various ways of doing this.
>> 
>> 
>> 
>> 
>> 
>> 
>> 
>> Glen, K9STH
>> 
>> Website: http://k9sth.net
>> 
>> 
>> 
>> 
>> _____
>> 
>> 
>> From: Peter Ravn <fasan5@xxxxxxxxx <mailto:fasan5@xxxxxxxxx> >
>> To: 'Alan Chandler' <alanchandler2@xxxxxxxxxxxx 
>> <mailto:alanchandler2@xxxxxxxxxxxx> >; 'Bill Carns'
>> <wcarns@xxxxxxxxxxxxx <mailto:wcarns@xxxxxxxxxxxxx> >; 
>> collins@xxxxxxxxxxxxx <mailto:collins@xxxxxxxxxxxxx>
>> Cc: 'Peter Ravn' <fasan5@xxxxxxxxx <mailto:fasan5@xxxxxxxxx> >
>> Sent: Saturday, July 9, 2016 3:45 AM
>> Subject: Re: {Collins} KWM-2 no output power
>> 
>> 
>> Thank you gentlemen for your fast response to my worries. Here is an 
>> update on my trouble shooting:
>> 
>> 1) There is a jumper in the back for the screen voltage, and in 
>> transmit the screen voltage on pin 3 is according to specs, so the 
>> big R148 screen resistor isn't open.
>> 2) Checked the small neutralization ceramic capacitor with an 
>> ohmmeter and an L/C meter. Capacitor neither broke or open. Capacity 
>> according to specs and varies when the trimmer is turned. Note: The 
>> now deceased finals were also Chinese 6146B's.
>> 3) No coax or other leads in the PA cage pinched or shorted.
>> 4) The big plate choke not toasted. Looks as before and not open.
>> 5) - 60 VDC on the control grids in receive and transmit.
>> 6) Antenna relay contacts will be checked next.
>> 
>> 
>> 
>> 
>> 
>> *** You are subscribed to Collins as dukeshifi@xxxxxxxxxxx. If you 
>> wish to unsubscribe, or modify your preferences please visit 
>> http://mailman.listserve.com/listmanager/listinfo/collins ***
> 
> *** You are subscribed to Collins as dukeshifi@xxxxxxxxxxx. If you 
> wish to unsubscribe, or modify your preferences please visit 
> http://mailman.listserve.com/listmanager/listinfo/collins ***




This archive was generated by a fusion of Pipermail (Mailman edition) and MHonArc.