[AMPS] 87A Fault 17

[Rich Measures]

>
>For a couple of years, I was getting intermittent fault #17 trips on my 87A
>while operating in contests, mostly on 10M and 15M. The manual describes
>this fault as "Abnormally low or high amplifier gain. Usually due to extreme
>mis-tuning or possibly an RF arc" (don't get too excited, Rich...)
>
>The fault almost always occurred on the first transmitted element or
>syllable. I could prevent it from happening by pulling the power back to
>about 1200W from the full 1500W. I first experienced the problem with a GAP
>vertical, which has been known to arc in wet weather. But later the faults
>continued to occur with a TH-7. I was advised that TH-7 10M traps are
>notorious for arcing in wet weather, so for a while I thought that was the
>cause. Later, I found that I could generate the fault by transmitting into
>the dummy load mounted near the tower. I even considered that my lightning
>suppressors might not have a high enough rating and were arcing internally,
>but the problem didn't go away when I bypassed the lightning suppressors.
>Could it have been the dreaded parasitic oscillation?? Well, I thought that
>was a long shot and there was no physical evidence of arcing inside the amp
>(clean tune cap, clean bandswitch, etc.)
>
>At about the same time this started happening, I began to see momentary
>infinite SWR spikes on the LED meter in the TS950SDX I was using to drive
>the 87A. I could reduce the frequency and amplitude of these spikes by
>increasing the rise time of the CW waveform from 4 ms to 8ms (a
>menu-adjustable parameter on the 950.) It also seemed to me that this
>reduced the frequency of the fault 17 trips, but it was difficult to
>correlate.
>
>I believe I posted a note about this on another reflector, and got a
>response that the problem was probably being caused by RF spikes from the
>950. Actually, I think the person who told me that was Tom, W8JI. I saw a
>more recent post from Tom that he measured 240W spikes on a 950 set for 150W
>output. He said it was a common problem on modern solid state rigs: the ALC
>is off in receive mode and ramps up too slowly when transmission begins. The
>timing diagrams in the 950 service manual seem to bear this out -- the CW
>waveform rises quite rapidly, while the ALC slope has rounded corners and
>rises more gently. You'd think the designers would have been concerned about
>this.

In addition to the ALC problem, Kenwood engineers also designed 
problematic AGC systems.  The TS-430S is a prime example.
>
>It seemed reasonable that spikes were causing the faults, especially since
>increasing the CW waveform rise time reduced the SWR spikes and maybe the
>fault 17 trips, too. But I just couldn't buy the explanation completely
>because the amp has a separate fault status for the overdrive condition. I
>couldn't understand how a 240W spike would generate an abnormal gain fault
>instead of an overdrive fault!
>
>Well, I had a conversation with Karl at AlphaPower that finally cleared this
>up for me. He said they often see fault 17 instead of the overdrive fault.
>Fault 17 is detected by measuring the ratio of the input power and the
>output power. If the gain is too high or too low, the amp trips. Karl wasn't
>sure about the amp's coding on this, but my background in software suggests
>a possible timing problem: the code that checks for fault 17 probably gets
>executed before the code that checks for overdrive. Since the amp does
>report overdrive faults if you manually increase the drive power too much, I
>suspect that the trip level for overdrive is set lower than the trip level
>for abnormal gain. Manually increasing the drive takes place on a time scale
>that is very slow relative to the CPU's clock cycle time, so the software
>sees the trip point for overdrive before the trip point for abnormal gain.
>But on an input spike with a very rapid rise time, the amp could get to the
>trip point for abnormal gain almost instantly relative to the clock cycle
>time, and if that condition is checked first, a fault 17 will be reported.
>(If the fault detection system is interrupt-driven, the same sequence would
>occur if the abnormal gain interrupt has priority over the overdrive
>interrupt.) The easy fix would be to check for overdrive first. A better fix
> might be to leave the order as is, but check for overdrive immediately
>after abnormally low gain is detected and report the overdrive fault
>instead, appending "possible transmitter RF spike" to the error message (or
>perhaps assigning a new fault number to this condition.)
>
>Recently, I happened to catch a fault 17 while monitoring the RS232
>interface. The message said, "Gain abnormally low." Karl wasn't sure if the
>code really distinguishes between high and low gain (i.e., it might display
>the same message in either case), but I think it's more than a coincidence

--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm