Bruce, thanks for the thoughtful response.
Yes, I was incorrect about the wattmeters being the same. Last night at dinner
I thought about the CT ratios and checked that after coming home, but missed
the difference in the voltage divider resistors after the diodes. Thanks for
pointing that out.
I have two pairs of tubes, 1990 date code and 2012 date code. The current
condition of these tubes is unknown to me but the 2012 pair was being used by
my friend. He acquired the amp from a SK estate a few years ago. He has not
tried the 1990 pair in the amplifier.
The first thing I did after fixing the fault 1 problem that would not let the
amp go into operate mode was to run the amp with HV on the tubes and no RF
drive. Yes, I have checked the B+ path from the tube plate back to the HV
board. All 4 tubes were static tested with HV and no problems. HV was set for
the high position and appears normal.
Next, I keyed the amp with a footswitch , still with no drive, and tested the
various bias voltages to make sure they were proper and that the amp was going
from receive mode to transmit mode. All appeared normal. No faults.
Next, I used the computer terminal to monitor the Ip and HV in the transmit and
receive state. 0 Ip on receive and 60 ma
Ip idle current in transmit, with no drive. HV drops a bit in transmit from
the standby state due to the load from the tubes pulling idle current. This
all looked normal to me. Both pairs of tubes were tested in this mode with the
same idle current result.
At this point I was ready for RF, or so I thought. Hooked the FT5K to the amp,
and the antenna switch matrix to the amp output.
First up, with the amp in the operate mode, I confirmed the receive was normal
with no attenuation. The rx pin diodes were working.
The amp was on 17M. I put my very low SWR 20M yagi on it with the RF output on
the rig set all the way down, about 10W. The first dit caused the amp to go to
20M, and both the tune and load capacitors moved to a preset position.
The next dit caused a fault 17 with red plate led indication, and reversion to
standby from operate.
If the amp is on a different band from the rf drive source, it stays in standby
until retuning/band change is done. It does not go to operate mode until the
band and tuning have been reset.
After the fault 17 on 20M, I tested the amp on some other bands, including
80/40/30 and 17 meters. All of these antennas have very low SWR. With about
10 watts drive, the result was fault 17 on each band on the 2nd dit, after the
retuning/band change was complete.
Next, I used the computer terminal program to communicate with the CPU and read
the input and output wattmeters with the GPIO command, with the amp in standby.
The input wattmeter appeared to be accurate and reads approximately the same
power as the exciter. The output wattmeter reads consistently low, checked at
drive levels of minimum to 100W from the FT5K. For example, with the rig set
for 100W, the input wattmeter read 103 and the output wattmeter read 79, a
23.3% disparity. With reduced drive, the input wattmeter read 19 with output
at 13, a 31.5% difference.
Last night, on the microprocessor board, I measured the DC voltage output from
both the input and output wattmeter, directly at the plug. With minimum drive
from the transmitter, somewhere in the approximate 10W range, the output
wattmeter read .13V and the input wattmeter read .49V. Because of the
different CT ratio and, as you pointed out, the different output voltage
divider resistance, these readings were expected to be significantly different
from each wattmeter.
The next test was to disconnect the microprocessor board loads from the
wattmeter outputs to see what, if any, loading effects the microprocessor board
electronics might be having on the voltage readings from the wattmeters. With
drive level untouched, the output wattmeter voltage remained completely
unchanged at .13V but the input wattmeter voltage increased to .55 from .49
volts with the microprocessor board disconnected.
We know the input wattmeter is approximately accurate compared to the drive
level, so the increased voltage with no load is likely the norm. Perhaps the
fact the output wattmeter was unchanged from load to no load conditions is the
sign of an anomaly with the op amp circuit on the microprocessor board. Not
sure.
I am going to do some more testing. The test setup is a 1KW 50 ohm dummy load
on the amp output, to verify the wattage levels. I will test the wattmeter
output voltage levels at the microprocessor board for various drive levels,
from about 10 to 200 watts and see how the power ratio holds with the internal
computer measurement. I will repeat these tests with the microprocessor board
disconnected to see what the internal load does to the wattmeter output voltage.
I will attempt to measure A/D:0 and A/D1 voltage levels and see what that shows.
No, I have not verified the input circuit is OK, other than to check and makes
sure there is continuity in the RF path through the input band switch. I would
like to get around this wattmeter ratio question first if possible, because we
know if the difference between input and output power is too great, it triggers
the fault 17. We just don't know how great this ratio difference must be per
the design of the amp.
73 Charlie N8RR
________________________________
From: bruce@bubble.org <bruce@bubble.org> on behalf of Bruce W2SE <W2SE@QSL.net>
Sent: Sunday, January 1, 2017 2:24 AM
To: Charlie Young
Subject: Re: [Amps] Alpha 87A Fault 17
On 31-Dec-16 23:25, Charlie Young wrote:
Happy New Year Bruce.
Ditto
The 87A has two built in wattmeters, one on the input and one on the output.
Only the output wattmeter reads on the front panel LED display. The input
wattmeter is used for the CPU to monitor the amplifier gain. If the CPU does
not see an output level it believes is commensurate with the input level, it
immediately puts the amp into a fault condition (17).
As soon as the first dit is transmitted, the amp faults giving a low gain fault
17 indication. So, I have yet to make RF with it that can be measured.
The two built in wattmeters can be read with an external computer terminal
plugged into the CPU, with the amp in standby. This monitor says the output
wattmeter is reading 23% less than the input wattmeter, with 100 watts applied.
The wattmeter transformer ratios are different as are the voltage dividers
after the D1 & D2 1N5711 rectifier diodes for each wattmeter. A/D is done by
the CPU (2 multiplexed analog inputs). Software makes all the decisions.
Try again with 50W and see what the percentage difference is as reported on the
external PC interface async serial link.
Can you measure the DCV outputs on the microprocessor Analog I/O board in
standby with 100W and 50W drive? microprocessor schematic sheet 6 right side
of page for the output signals: A/D0 and A/D1 ?
I think that you may have some of this data already.
I can think of one way to dummy-up the input and/or out-watts voltage to help
you figure out the ratios for fault 17 but it will mean a temporary circuit mod
with a small pot.
By design, the amp will kick out on fault 17 if there is too great a disparity
in what the CPU thinks the readings should be. So far, I have been unable to
find the factory parameter that is programmed in to trigger a wattmeter related
fault 17.
Certainly OutWatts > InWatts for normal and correct operation. It has
to be a ratio since input drive can vary.
I would expect #17 generation is disabled (in software) during tuning.
I would guess that #17 should expect to see an A/D0 : A/D1 voltage ratio
that equates to at least 6 to 10db amp gain, that's just a guess. Anything
less is either: soft tubes, bias issue, low HV, input issue (which you have
probably ruled out), output tuning/matching issue. and of course wattmeter
issue. I have no idea what this means for the A/D0 and A/D1 voltage ratios and
values which would require some testing.
With the amp turned off and unplugged, and the HV bled down and fully
discharged, have you tested continuity between the HV supply and the plate
caps? (this tests DC continuity of the chokes feeding HV into the plate
circuit, which the amp does not measure). Be sure to check the HV R34 10 ohm
25W glitch resistor.
So far, I have run 100W through the amp in standby with a dummy load on the
output, and measured the voltage output directly at each of the two wattmeters.
The wattmeters are
NOT
identical but fed with a different ratio current transformer on each input. I
have run the calculations and they confirm the readings on the output wattmeter
are consistently low compared to what they should be. This confirms what the
computer monitoring shows.
The amp does this on every band.
I am about ready to pull the output wattmeter board and check all the
components.
Fixing the wattmeter may not solve the fault 17 issue, but in order to rule out
the wattmeter as a problem, I have to get it reading up to normal level. So
far, no one has been able to tell me how much difference in wattage reading
will trigger the CPU fault.
It has to be in an algorithm in the software for the CPU.
If you want to brain-storm via Echolink or telephone let me know. I like
puzzles.
73
Bruce W2SE
I am optimistic we will find the problem, but the fault 17 may or not be the
wattmeter. It certainly could be.
I will post on the Amps Reflector what is found.
Thanks for the response.
73 Charlie N8RR
________________________________
From: bruce@bubble.org<mailto:bruce@bubble.org>
<bruce@bubble.org><mailto:bruce@bubble.org> on behalf of Bruce W2SE
<W2SE@QSL.net><mailto:W2SE@QSL.net>
Sent: Saturday, December 31, 2016 7:40 PM
To: Charlie Young
Subject: Re: [Amps] Alpha 87A Fault 17
Hi Charlie,
I haven't worked on an 87a yet but ...
Just a thought: Did you check/compare the amps' power output to a known good
wattmeter connected to the 87A's output?
Both standby and operate.
Did you get any key-down current and voltage measurements during key-down that
confirm that you are getting normal output power?
If you can verify normal power output, the first place that I would look would
be the output wattmeter circuit.
Good luck.
73
Bruce W2SE
_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps
|