Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

Resources
Calendar
Classifieds
Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement
About eHam.net

donate to eham
   Home   Help Search  
Pages: Prev 1 [2]   Go Down
  Print  
Author Topic: Yaesu FT101EE power glitch on PTT release  (Read 15455 times)
VK6HP
Member

Posts: 553




Ignore
« Reply #15 on: July 29, 2017, 01:03:33 AM »

While you're looking at timings, and hopefully following things back down the transmit chain, have a look at the grid bias to the PA.  In most rigs switching to "receive" hard biases the tubes well beyond cut-off (tens of volts negative, typically), more or less guaranteeing the PA becomes inoperative.  See if you can see that happening, and when.

By the way, is the PA properly neutralized?  You can check and adjust that quite well with your test setup. 

The glitch on the "receive" line at the end of the burst is also interesting but doesn't seem to be accompanied by any momentary switch back to "transmit", at least at the test point you're looking at.
Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #16 on: July 30, 2017, 11:40:42 PM »

We were thinking along the same lines. Here is a capture of the timing between the "transmit" line ( blue trace ) and the grid supply voltage to the finals ( measured at the regulator board ).



This capture shows the timing of the power glitch ( blue )  relative to the grid bias supply voltage ( yellow ). The glitch happens after the finals are biased off.



This capture shows the actual voltage at the grids of the finals ( yellow ) vs. the grid supply voltage at the regulator board ( blue ).



This capture show the voltage at the plate of the driver vs the grid supply to the finals. Notice the Vpp of the glitch is the largest at the plate of the driver



This capture is the voltage at the grid of the driver ( yellow ) vs. the RF output after the 40 dB attenuator ( blue ). It appears that the 5 pF load of the probe quenched the RF glitch. Notice how slow the driver bias returns to receive mode. I wonder if it were to return to -20 VDC quicker, the glitch would vanish or if the driver is not well neutralized. There is mention that the neutralization capacitor for the driver needs to be changed if American tubes are used. Page 4-12 of the service manual "Driver tubes ( 12BY7 ) are less tricky, but if you use G.E. tubes, reduce the neutralizing capacitor ( C-123 ) to 3 pF, or increase it to 5 pF when using RCA tubes. This FT101EE has a 12BY7A with the Raytheon label and says "12BY7A/12BV7", "registered Japan". Initially I though it said made in Japan but maybe this is the problem

Logged
VK6HP
Member

Posts: 553




Ignore
« Reply #17 on: July 31, 2017, 01:00:42 AM »

You've had a busy weekend Smiley  It looks like you're on the right track and the only additional thing I'd do is to make absolutely sure that the burst isn't originating a stage or two before the driver when the radio is keyed with no audio input and using the MOX.  I hadn't heard of the 12BY7 issues for the FT101, although the PA (and its neutralization) is known to be picky with respect to the origin and era of the sweep tubes.  As a matter of interest, what PA tubes do you have, and are they both the same brand and internal construction (as they should be)?

It looks like the swing at the plate of the driver is so big it's effectively overcoming the negative grid bias and driving the PA into operation.  With regard to the timing of the bias voltage step, it'd be worth having a look at the dc grid return paths around the PA and driver stages (resistors etc).  That said, the transition is not really that long and tube circuits are not usually as critical as their solid state counterparts.  But I'd look nonetheless.

With the radio operating normally, I'd also do the PA and driver neutralization as set out in the service manual and just check for any anomalies. If you can't get that right, look further into the tube type issues and modifications.  By the way, some radios of the FT101 era advised neutralizing by getting a symmetrical plate current dip, and coincidence of the dip with maximum power output.  That should happen with good neutralization but you can do better with your setup.  The technique of minimizing the PA output with the HT removed (or slightly less kosher with just the filaments off) will get you a more accurate result. 

I guess you already know about the Fox Tango web site and the many FT101 owners' sites. If you do discover problems with the neutralizing and tubes, it'd be worth trawling those sites.  I don't want to show my Kenwood colors too much (being also the owner and restorer of a nice FTDX560!) but it is pretty clear that PA stability is something of an Achilles heel for the 101 series.  As in other radios, time has seen us substituting brands and epochs of tubes never really envisaged by the designers.

Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #18 on: July 31, 2017, 02:14:56 AM »

My 12BY7A tube looks exactly like this one right down to the markings and the halo getter and plate color so it appears to be made by Matshushita.

http://www.ebay.com/itm/Vintage-Raytheon-12BY7-NOS-NIB-Tube-1970s-Halo-Getter-Matsushita-Japan-Tested-2-/112484397440?hash=item1a3097b180:g:KtwAAOSwD8BZa~Tu

I will be checking if the bias for the driver is behaving as it should since the bias on the finals recovers much faster.

It has US made Sylvania finals. I did the capacitor modification because I found the adjustable neutralization capacitor set for minimum ( no plates of the air variable meshing ) when I bought it. Replaced a 100 pF mica cap with a 10 pF NPO in series with the air variable. It neutralized easily then.

As I recall there was a small amount of leakage from the mixer before the driver to the driver in receive mode. I think we are getting close.
« Last Edit: July 31, 2017, 02:19:43 AM by HAMHOCK75 » Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #19 on: August 02, 2017, 03:46:14 PM »

I did an analysis of the bias supply for the 12BY7A. It seems to be behaving just as theory suggests is should. About -3.5 VDC in transmitt and -20 VDC in receive. The 12BY7A is driven by a mixer which uses the crystal bank to produce the output frequency. I pulled the crystal for the band being tested but the output glitch is still there.

I plan to get a handful of 1 pF capacitors to add to the 2 pF neutralization capacitor for the 12BY7A. Let's see if that helps.

As a further test. The 12BY7A looks a bit like a tuned grid, tuned plate oscillator circuit. I tuned TC5 which adjusts the tuned grid circuit frequency back and forth a bit. By doing so, the glitch could be made to disappear but the output is low when it does disappear.
« Last Edit: August 02, 2017, 04:08:33 PM by HAMHOCK75 » Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #20 on: August 08, 2017, 04:14:17 PM »

I tried adding a parallel 1.5 pF capacitor to the existing 2 pF neutralization capacitor for the 12BY7A to no effect. In the meantime I noticed something unusual.

Below is a capture of the drain voltage of the input mosfet vs. the collector voltage of Q3 the second transmit mixer. The collector voltage is supposed to be higher than the drain voltage to turn on a diode in receive so Q1 is connected to the pre-selector circuit with T102. During transmit, that voltage is supposed to be below the drain voltage of Q1 so the diode is reverse biased thereby disconnecting Q1. The schematic shows that in transmit the collector of Q3 is 11 volts. That does not appear to be happening.

The yellow trace is the collector of the second transmit mixer which has RF present during transmit.



Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #21 on: August 09, 2017, 05:14:54 PM »

Today to take a break from the glitch, I decided to use a spectrum analyzer to view the output at 3.8 MHz. VR1 on PB1181 where the glitch seems to be centered is to be adjusted at 3.8 MHz RF out to minimize a spurious output at 3.840 MHz.

First I determined where the spurious signal originates. It is  ( 4 x VFO ) minus ( 2 x xtal osc for 80M ) = ( 4 x 5.720 MHz ) - ( 2 x 9.520 MHz ) =  22.88 MHz - 19.04 MHz = 3.840 MHz.

Adjusting VR1 nulls the 3.84 MHz by -65 dB relative to the desired output. The voltage at the base of Q3 is almost exactly 1.1 VDC as shown in the service manual. I moved the frequency to 3.795 MHz with the result that the spurious moved up to 3.860 MHz as expected.

VFO = 5.725 MHz with Xtal Osc = 9.520 MHz = 9.520 - 5.725 = 3.795 MHz

( 4 x 5.725 MHz ) - ( 2 x 9.520 MHz ) = 3.860 MHz

The diode that connects the input mosfet to T102 is D1. To see if the leakage in this path causes the glitch, I un-soldered the anode end of D1. The glitch vanished. That might explain why the manual RF gain control which is in the source of the mosfet can also make the glitch disappear when the RF gain control is lowered to about 8 from 10.
« Last Edit: August 09, 2017, 05:24:41 PM by HAMHOCK75 » Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #22 on: August 10, 2017, 04:43:10 PM »

I think I have found the source of the power glitch. It is not a neutralization problem but a leakage of RF into the receive portion of the FT101EE in the PB1181B board, “high frequency unit”. The schematic is below. If the voltages at the various measurement points are correct then this circuit should and does work properly. These voltages only apply in the CW mode. In LSB or USB we have a problem.

The collector of Q3, the second transmit mixer is shown at 11 VDC during transmit. The drain of Q1, the input dual gate mosfet during transmit is at about +13 VDC because the connection to the RF gain control is open circuit by RL-1. +13.5 VDC though resistor R7 ( 1.8K ) and R24 ( 220 ohms ) then passes through the fet acting as a switch so that voltage shows up on the drain. When this happens, D1 has +13 VDC on its cathode and +11 VDC on its anode so it is reverse biased. This isolates the receiver input from the transmit second mixer output. This works well in CW mode because the finals are biased off. The second mixer always has output at the transmit frequency because it is continuously being driven by the crystal oscillator, Q4 though emitter follower Q5, and the 5.520-6.020 MHz output of the first transmit mixer. It is this Q3 output that causes the collector of the second transmit mixer to drop to +11 VDC thus reverse biasing D1.

Unfortunately, that does not happen in the SSB modes. In the SSB modes, the finals are biased on. There is no drive to the second transmit mixer from the first transmit mixer when nothing is being said into the microphone. The second transmit mixer has no RF input at its base, only the input from the crystal oscillator at its emitter. This is the RF that is visible in the oscilloscope image above during transmit with no microphone attached. The problem is that the RF from the crystal oscillator does not bring the collector of Q3 down sufficiently to reverse bias D1 so the crystal oscillator signal can leak into the receiver. This leakage creates the power glitch.



D1 is a 1S1555 diode with a zero bias capacitance of 1.5 pF. I replaced D1 with a 1.5 pF capacitor. The glitch was gone. To determine how much isolation is needed, a 0.1 mfd capacitor in series with a resistor was tested in place of D1. The resistor values started at 1k, then 2.2k, then 4.7k. It took 4.7k to eliminate the glitch although the 2.2K had reduced is significantly.

Not sure what to do about this since it seems to be a design problem.

« Last Edit: August 10, 2017, 04:47:47 PM by HAMHOCK75 » Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #23 on: August 13, 2017, 01:52:16 PM »

The following tests have been done since my last post

  • 1. Diode D1 in the schematic above was removed. The glitch disappeared entirely. A 0.1 mfd capacitor with a series resistor was used to replace D1. The R value was increased until the glitch disappeared. The glitch was down considerably with 2.2K and gone with 4.7K
  • 2. Diodes were added in series with R51 which is on the chassis. R51 supplies the collector voltage to Q3, the second transmit mixer on PB1181B via pin 11 of the schematic above. This lowered the voltage on the anode side of D1 allowing it to be reverse biased more. Three series diodes reduced the full power glitch to a 0.65 watt glitch

I next tried to isolate why this problem occurs. To do that parts were removed until the only parts left were the ones needed to produce the glitch.

  • 1. The crystal for the band was pulled so there was no drive to the second transmit mixer. The glitch remained.
  • 2. Q3, the second transmit mixer on PB1181B was removed. The glitch remained.
  • 3. D2, the connection from the input mosfet amplifer in receive to the first receiver mixer, Q2, was removed. The glitch remained.
  • 4. The 12BY7A was removed. The glitch was gone.

I did a timing test to see when the antenna changeover relay RL-2 actually changed to receive. RL-2 is a pair of SPDT contacts only one of which is used to change between transmit and receive. The center arm of the unused SPDT switch was connected to +13.5 VDC from the power supply with a jumper. In receive, the closed contact had +13.5 VDC and in transmit that voltage became zero. That voltage was monitored on the oscilloscope relative to the "R" line. RL-2 did not disconnect from the finals until 7 msec after the unit was in receive so the glitch occurs well within the time RL-2 was connected to the 50 ohm dummy load.

These further tests were done to isolate the problem.

  • 1. The voltage on gate 2 of the input mosfet, Q1, was measured. It jumped to +6 VDC during transmit and slowly dropped back to about +3.44 VDC in receive. That means that the gain of the input stage with the RF gain control at 10 ( max ) is even higher than normal upon return to receive. It also means that with D2, and Q3 removed, Q1 acts as a very high gain amplifier and is the only device now connected to the 12BY7A grid via T102. In effect, Q1, the 12BY7A, and the finals are all in series with Q1 at max gain and due to timing the 12BY7A and finals are biased off
  • 2. The receiver input at RL-2 was shorted to ground with a jumper. The glitch remained.
  • 3. C1 at the input to Q1 in the schematic was removed. The glitch disappeared.

C1 connects gate 1 of Q1 to pre-selector T101A which is a high impedance node. It appears that the receiver input pre-selector picks up noise or a glitch during the transition from transmit to receive, applies that to Q1 which is operating with very high gain, enough to create a signal large enough to overcome the bias of the 12BY7A and finals with this huge glitch the result. More later
« Last Edit: August 13, 2017, 01:54:18 PM by HAMHOCK75 » Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #24 on: August 14, 2017, 04:59:32 AM »

These tests were done today.

  • 1. Replaced C1 on the PB1181B board
  • 2. Tried the RF attenuator which is before the pre-selector. The attenuator stopped the glitch which suggests it originates not in the pre-selector but nearer the antenna
  • 3. Before the attenuator is a filter board PB1116A which contains a trap and diode detector, D4, for the power out reading of the front panel meter. I disconnected the output from this board to the attenuator. The glitch disappeared. 
  • 4. After reconnecting the output of PB1116, the input was disconnected. The glitch remains

It appears that the source of the glitch is PB1116A. More later.
Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #25 on: August 16, 2017, 03:29:15 AM »

Today I removed the PB1116A board with the trap for 5.94 MHz

    1. Then connected the coax output of PB1116A to the coax input effectively bypassing it. The glitch reappeared
    2. I disconnected the two coaxes again and placed a 12" jumper wire on the coax leading to the preselector to sniff what location might be creating a glitch. The jumper just drapped over the wiring did not result in a glitch, however, as it approached the pi-network coil or L25 which is right after the pi-network coil, the glitch returned.
    3. Below is what was found at L25 when there should have been no activity.  The blue trace is the R line. Low is in receive, high in transmit.



    4. The 12BY7A driver was pulled. The glitch was gone.
    5. The PB1181B board was pulled and the 12BY7A replaced. The glitch was gone.
    6. The PB1181B board was replaced but one by one every component that was connected to pin 11 output to the 12BY7A was disconnected starting with Q3, D2, D1, and C5 in the schematic above. The glitch remains.
    7. It seems that this glitch is picked up by the PB1116A board and the wiring in front of it which almost touches L25, gets amplified by the receiver FET, Q1, which makes it  large enough to turn on the tube stages creating the full output power glitch in my first post.

It will be interesting to find out if the glitch originates on PB1181B or if it leads elsewhere.
« Last Edit: August 16, 2017, 03:32:00 AM by HAMHOCK75 » Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #26 on: August 26, 2017, 05:46:20 AM »

The above problem was traced to a poor solder connection at C7 on the chassis. After the C7 issue was eliminated, RF was found at a low level during transmit along with some 60 Hz. This test was at 27 MHz. Using a spectrum analyzer, the spurious signal moved down 30 Khz as the output frequency was increased 10 Khz to 27.01 MHz. Since the output frequency is equal to f out = 33.02 MHz – ( VFO – 3.1785 MHz ) then VFO = 33.02 + 3.178.5 – f out = 33.02 + 3.178.5 – 27 = 9.1985 MHz.. If f out = 27.010 MHz, then VFO = 33.02 + 3.178 – 27.01 = 9.1885 MHz. The frequency leaking out is the third harmonic of the VFO at 3 * 9.1985 = 27.5955. Wondered if the selectivity of T102 and T103 should be able to eliminate this spurious. T101 Q was measured by pulling PB1181B and measuring the RF at pin 8 while using a signal generator at the input. Q was found to be about 8-10 with a 3 dB bandwidth around +/-1.4 MHz. So the pre-selector cannot remove this spurious. Yellow trace is the R line.



The examined the AGC signal at pin 9 of PB1181B. It appears to be slow attack, slow release. Yellow trace is the R line.


Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #27 on: September 24, 2017, 02:20:12 AM »

Tried adding a ferrite bead by Fair-Rite Products Corp at the input to PB1116A, no effect. Added the ferrite bead after PB1116A with a strong effect. Ferrite bead was p/n 2673000801 with minimum impedance of 44 at 10 MHz and 70 at 25 MHz. P/n 2643000801 was also tried with minimum impedance of 50 at 25 MHz and 92 at 100 MHz ( no spec. at 10 MHz ). It did not work as well. The glitch was absent with p/n p/n 2673000801 in all bands but 15 M. 88. Tested the sensitivity of the receiver with a HP synthesizer set for 0.1 uV but the bead did not seem to make any difference.



I should probably get ambitious and put PB1116A in a shield can and see if that gets rid of the glitch.
« Last Edit: September 24, 2017, 02:36:41 AM by HAMHOCK75 » Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #28 on: September 24, 2017, 01:26:32 PM »

I found a better solution that works well for now. I disconnected the wires from the RF gain control. The RF gain control is not a linear potentiometer. Maximum gain is position 10.

Position   Resistance ( ohms )

0            928              
1            903
2            533
3            261
4            220
5            179
6            146
7            100
8              54
8.5           30
9              6.3
10            0.4

I added a 36 ohm resistor in series so that the resistance when the potentiometer is a zero ohms does not go below 36 ohms. It eliminated all RF spikes completely in every band. I also checked the receiver sensitivity before and after. There was little difference. I could plainly hear a 0.1 uV signal from an HP synthesizer with or without the 36 ohm resistor.
« Last Edit: September 24, 2017, 01:32:05 PM by HAMHOCK75 » Logged
HAMHOCK75
Member

Posts: 640




Ignore
« Reply #29 on: September 30, 2017, 01:55:20 PM »

With a solution to the high power RF power glitch that occurred going from transmit to receive, I went back to see why just prior to the PTT being released there is a smaller power glitch associated with the use of the PTT on the microphone but which does not occur if the MOX ( manual transmit ) on the front panel is used. I bought a used Yaesu VE-7A microphone which I understand is the microphone that normally came with the FT101. Here is a schematic and photo from another web site.

http://www.oh1sa.net/data/mirrors/rg4wpw-microphone_connections/ye7a.html

Notice the schematic shows only one switch. The FT101E instruction manual shows the same on page 4. However the YE-7A I received has two switches as shown below with one of them in series with the microphone element.



The contacts are arranged to be staggered so the lower contacts connect before the upper contacts in the photo. The lower contacts are the PTT switch, the upper the microphone switch. So when pushing the PTT in, the FT101 goes to transmit first, then the microphone is connected. Upon release of the PTT, the microphone is disconnected first, then the FT101 leaves transmit mode. Needless to say, the YE-7A microphone showed no output power glitch when it was released.

I then examined the original microphone that showed the glitch. It also has two switches just like the YE-7A  but the design is different.



Notice on the switch there are two arrows one slightly ahead of the other. So this switch does the same as the Yaesu. So why didn't it work? The switch had been wired incorrectly so it was working essentially the opposite of the YE-7A. After correcting the wiring error there is no more power glitch with either microphone upon release of the PTT switch.
Logged
Pages: Prev 1 [2]   Go Up
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!