Friday, August 7, 2015

Amplifier Woes! Instability at Low Drive Levels! (Video)



I have to keep reminding myself:  This is not "plug and play." These are not appliances. 

After I got my 40 meter problems squared away,  I was doing some testing on my beloved 17 meter BITX.   I noticed something weird:   With the CCI EB63A amp feeding my 17 meter Moxon antenna, as I raised the output of the BITX17 driver, at one point (at about half the max input power) the SWR would suddenly spike. Then, as I raised the drive level above that point, the SWR would go back to normal. 

I looked at it on the 'scope.   I can see the signal go very ugly at the mid-level drive point.  In the FFT display, I can see that there is a strong signal at around 435 kHz.  The 18 MHz signal seems to be riding along on top of it.  Take a look at the video above.

Additional clues: 

I see no signs of the 435 kHz signal at the output of the BITX 17.  It seems quite clean.

This problem disappears if I replace the Moxon with a dummy load. 

This problem does not show up if I feed the EB63A with my almost identical BITX20.   And I use the same LP filter on both 20 and 17 in the CCI amp.

Any suggestions?  Has anyone had this kind of problem? 

Allison and Pete have been helping me with this.  Thanks to both of them.  

Our book: "SolderSmoke -- Global Adventures in Wireless Electronics" http://soldersmoke.com/book.htm Our coffee mugs, T-Shirts, bumper stickers: http://www.cafepress.com/SolderSmoke Our Book Store: http://astore.amazon.com/contracross-20
at

6 comments:

  1. dougzAugust 9, 2015 at 3:08 PM

    OT, but I thought it might be of interest to you and your podcast audience -- Incredibly useful headmount magnifier for $8 -- http://http://boingboing.net/2015/08/05/incredibly-useful-headmount-ma.html

    ReplyDelete
  2. caulktelAugust 9, 2015 at 4:38 PM

    Bill, what IF did you use on your Bitx17, I used a 10 MHz IF and I have a loud internally generated birdie on 18.136, but otherwise very clean. I don't notice the same problem you do.

    Joel
    KB6QVI

    ReplyDelete
  3. Rupert GoodwinsAugust 10, 2015 at 10:28 AM

    Have you seen the eHam thread about instability with the CCI EB63A at http://www.eham.net/ehamforum/smf/index.php?action=printpage;topic=89500.0 ? Doesn't seem like quite your problem, but I've had linears honk away at particular combinations of drive/input/output connections that would have defeated Spock's powers of logical analysis. Because it's LF instability rather than VHF, it might be power-side issues, with some high-impedance-at-400kHz path putting the finals (or driver - I don't know the CCI circuit) into their unhappy place until enough drive changes their effectve impedance and moves whatever ad-hoc network it is out of the danger zone. That's the fun of RF - at least you haven't got the sort of VHF squegging that can fry a part at a speed proportional to its cost. (A friend of mine at a big aerospace company tells me he blew "several times my monthly wage"s worth of hexfets during one black afternoon detonation-debugging a satellite power bus...)

    ReplyDelete
  4. AnonymousAugust 10, 2015 at 2:22 PM

    Because the 435 KHz signal does sometimes pop up on QRP, I suspect the BITX output stage. Please make sure it is dead stable before you go on.

    Chip V.

    ReplyDelete
  5. KA7OEIAugust 14, 2015 at 6:32 PM

    While I do enjoy hombrewing rigs, it is times like this that an additional piece of "test equipment" (say, a "store bought" radio) can come in handy to rule things in or out.

    If you can borrow, say, an FT-817 for long enough to see if the same thing happens with it at a similar power level, under similar conditions.

    Also, I take it that the BITX and the amplifier are NOT using the same DC supply? Who knows what evil lurks on a a common power bus... Adding an extra 100-470 uF of capacitance (using a 105C low-ESR electrolytic) very close to the RF blocking choke on one or both of the PAs can sometimes quash the gremlins before the make their way out of the circuit.

    Finally, it might be worth trying a 1 dB pad between the BITX and the amplifier. This would increase the return loss by 2dB and "change the map" without causing too much loss - plus adding a bit of resistive loss in the mix can sometimes do wonders if it is, in fact, strange reactances that prevent the two piece of gear from "playing nice" with each other.

    ReplyDelete
  6. AnonymousAugust 17, 2015 at 5:17 PM

    Maybe it's the stopband impedance of the LPF at 455kHz that’s causing the oscillation. Generally, filters block signals by creating a bad match in the stop band, usually by going towards an open or a short. There are exceptions, but filters can’t be relied on to be 50 Ohms in the stop band. While not the same situation, I had a highpass filter present a low impedance (almost a short circuit) to the harmonics of a square wave signal I was trying to clean up. Shorting out the 3rd, 5th ,etc. harmonics caused excess current and heating in my case. Changing the filter design to have a high impedance in the stop band fixed my problem.

    ReplyDelete