[Amps] Alpha 87A Fault 17

Sat Dec 31 13:51:18 EST 2016

This one is a long-shot, but are you sure the amp doesn't have bad tuning parameters? It's conceivable that the user tuning parameters for 20m were for a non-resonant antenna or perhaps got corrupted by CPU power supply disruptions during the many faults this amp has experienced. If you haven't already, go back to using the default tuning and check 20m again. The default tuning parameters should work for a flat antenna, though wouldn't be optimized for max gain. I doubt both the default and user tuning parameters are corrupted, and I don't *think* the defaults can be overwritten by the user without issuing a special command on the RS-232 interface.

Another possibility is that someone disconnected the stepper motors from the variable caps or band switch, then moved the caps and/or band switch before reinstalling the motors. In other words, the optical counters on the motors may no longer be properly aligned with the caps and/or band switch. I can't recall if this is possible or if the optical counter is keyed so this can't happen. If not, a misalignment could cause massive tuning issues that would surely cause fault 17. 

I've had the arc to L1 problem, and I'm pretty sure it will generate a fault 8 or 9 or maybe 1. I never got fault 17. Also, it's likely that the amp would produce some power (more than you're seeing) before L1 arcs and the amp faults. However, I would still carefully check L1 and the vicinity for any signs of arcing. On at least some revs of the 87A (perhaps the early ones), the rubber insulating pad under it disintegrates over time, allowing the coil turns to arc to the PCB and the screw near L1.

I agree that it's not likely to be the band switch alignment if fault 17 occurs on every band, unless the band switch is seriously out of alignment. As described above, that's not likely unless someone disconnected the stepper motor and moved the band switch before replacing the motor.

FWIW, I don't believe the problem with fault 17 is that the amp reports the last fault. As Dick Ehrhorn described it to me, the 87A firmware reads the sensors in its main loop. The sequence of instructions checks for a gain fault before checking for the many other faults that could be causing it. For example, you usually get fault 17 instead of fault 13 when overdriving the amp. Yeah, it's a gain fault, but that's not the problem. The typical way to deal with this would have been to redesign the fault processing to be interrupt-driven, which perhaps the programmer didn't know how to do or there wasn't room in the firmware memory to do it. Perhaps a simpler fix would have been for the code to reset the sensor checking sequence whenever it detects a fault condition. That would allow it to recheck the sensors in a more informative sequence.

Hope this helps. Let us know what you find.

73, Dick WC1M

-----Original Message-----
From: Jerry O. Stern [mailto:jsternmd at att.net] 
Sent: Saturday, December 31, 2016 10:28 AM
To: 'Jim W7RY' <jimw7ry at gmail.com>; 'Charlie Young' <weeksmgr at hotmail.com>; amps at contesting.com
Subject: Re: [Amps] Alpha 87A Fault 17

Fault 17 could be due to several issues, the fault reporting only shows the last fault and sometimes the actual triggering cause may precede the 17.   You will need to hook up the 87A to a PC with hyperterminal or similar and look at fault history.

Three top causes : 1) bandswitch misalignment; 2) GPIO failure and 3) arc to L1 with carbon trace

1) If it happens on every band when manually selected (mid band) then unlikely to be bandswitch unless someone really knocked it around during repair

2) For the GPIO then due to a failed diode (1N5711) in one of the wattmeters. There are Input and Output wattmeters. 

Here are some tests. 

a) Connect PC to serial port with a modem program communicating with the amplifier. 

b) With the 87A in STANDBY, transmit through the amplifier with 100 w from the radio to a dummy load (or antenna with low SWR). 

c) On the PC, type EXT ON. 

d) On the PC, type GPIO (radio still transmitting 100w). This command reports Grid current, Plate current, Input power, and Output power each time it is typed. It should be showing about 100w input and about 100w output. If one of those is not reading correctly, that indicates the wattmeter diodes are damaged.

3) requires removal of L1 and close visual inspection

73 es HNY
Jerry NY2KW

-----Original Message-----
From: Amps [mailto:amps-bounces at contesting.com] On Behalf Of Jim W7RY
Sent: Saturday, December 31, 2016 9:41 AM
To: Charlie Young; amps at contesting.com
Subject: Re: [Amps] Alpha 87A Fault 17

I see a variable capacitor in the wattmeter, the only adjustment.  Is this a calibration adjustment?

This is called the Null adjustment. Which you set for lowest reflected power with the amplifier into a KNOWN good 50 ohm (very low return loss) load. 

Most good wattmeters have them. If you look in any of the later handbooks, or on various commercial wattmeters, you will find the null cap.

73 and good luck with the 87A. (I’ve never owned one).

Jim W7RY/0

-----Original Message----- 
From: Charlie Young 
Sent: Friday, December 30, 2016 9:40 PM 
To: amps at contesting.com 
Subject: [Amps] Alpha 87A Fault 17 

Hello friends, I am in need of some info and/or suggestions regarding a fault 17 problem.

I am helping a friend restore his 87A to operation after he changed the blower.  Immediately prior to the blower change, the amp was inoperable with a series of fault indications.  Unfortunately, I do not know the faults he was getting but he does not think fault 17 was one of them.

I have only been inside one other Alpha 87A which had a loose wire on the Rbias terminal on the HV board.  Prior to this case I had no experience with operating or repairing an 87A.  This is a different animal from my other amps, all of which I maintain myself.

To repair this one, I studied the schematic and did research online about pin diodes, to figure out how the amplifier is supposed to work.  Was able to figure out the bias sources and how the thing goes from receive to transmit.  Next I reviewed many posts in the Amps archive, plus the conversations on the Yahoo 87A board.

Several issues were worked through and corrected, including blowing step start fuses, a blown bypass capacitor on the plate choke, and replacing a control wire that had been cut (apparently during the blower replacement) which was preventing the amp from going into operate mode from standby.

HV is normal.  All the bias sources (-109 volts/ Tbias + or - 30 volts/ Rbias  960 volts ) are there.   The bias sources switch off/on as they should between transmit and receive, and the Tbias switches polarity.

Without drive and keyed with a footswitch, using a terminal program to communicate with the 87A,  the tubes have 60 ma resting current, the same for both sets of tubes I have.

With the radio (FT5K) hooked to the 87A, receive is normal with the amp in the operate position.   My 20M yagi has a flat SWR.   On transmit, the first dit sent causes the 87A to go to 20M (if it is on another band). The tune and load capacitors change position.  On the 2nd dit sent, the amp immediately gives fault 17 with red plate led and goes offline.  Fault 17 is a soft fault, and the message on the computer monitor is severe mistune or very low gain.

Online, I found a reference to input and output wattmeter readings being different causing fault 17 (among other potential causes).   With the amp in standby and the radio keyed, I used the GPIO command in the Alpha software to get a read on the input/output wattmeter.   With the input wattmeter reading 103 watts, the output wattmeter reads 79.

With lower drive, the input wattmeter reads 19 and the output reads 13.

My question to the gurus out there with experience on this amp:   What differential reading on these wattmeters is enough to cause a fault 17?   Does anyone know what the actual tolerance is?

I checked the diodes in the output wattmeter in circuit with an ohmmeter and get a standard diode reading on all 4.

This does not mean they are good, but none are shorted.

I see a variable capacitor in the wattmeter, the only adjustment.  Is this a calibration adjustment?

Knowing if the wattmeter differential is outside of the Alpha operating range would be helpful.  If it is, I can focus on the wattmeter.  If my variation is acceptable, I need to keep looking for the problem elsewhere.

This fault 17 happens so fast, it is hard to tell with the led meters what is going on.  I see the grid current flicker up, so I think the tube is being driven.  What are the symptoms of a failed input pin diode?   If there is a plate current spike, it is too fast for the led's to respond.

There is no noise or arcing detected, with the room lights turned off.

It does the same thing on every band, so it is not band related.

I was wondering if maybe the input bandswitch out of synch during the blower change, but this seems unlikely. I have not tested that yet.

I have not unsoldered one end of the pin diodes and tested them yet.  The receive functions OK so I am pretty sure the rx pins are OK.   I suppose if the TX pins were not turned on or open during transmit, the wattmeter would see no output.

IF the tx pins were open, it would be like transmitting into an open circuit. Maybe the fault system is so fast

it prevents a plate current spike or high voltage arc.

I have had no other fault except 17, after I repaired the cut wire.  Before I fixed that, it gave hard fault 1 when the operate switch was pushed.  The Rbias monitor voltage was missing.

This has been educational.  However, I would really like to give the amp back to my friend in operating condition.   Any help would be appreciated.

73 Charlie N8RR

Charleston WV

_______________________________________________
Amps mailing list
Amps at contesting.com
http://lists.contesting.com/mailman/listinfo/amps
_______________________________________________
Amps mailing list
Amps at contesting.com
http://lists.contesting.com/mailman/listinfo/amps