>Ok Steve I have Also have had bad experience, with the build in choke like >W3RJW have done it. With only 200W out the teflon "cocoon" exploded. > >I have my self put in a feedthrough, with 1/4wave choke in the side wall >of the cavity. I am using af copy of G3sek tretrode boards, so the voltage >for the screen is ok. > >The HAM's getting out lets say 500W. Do they do that without dangering the >tube. The current, at that power will be 600mA. Is that ok???. > >Thanks for reply > > Best regards >[Peter Hansen] Peter Hansen OZ1LPR I have seen several 432 MHz AM-615x amplifiers reliably put out 450 watts with the plate current running at something over 500 mils (the meter needle being against the right peg), and using the Amperex DX-393 tube. I've cautioned the owners that because of the poor efficiency, they should keep that extra power in reserve for "NEW GRID!!!" times, and run the amplifiers at no more than 400 watts output during normal operations. I have installed one 4CX400A about three years ago into another 6155 and that amplifier would put out over 500 watts but with plate current near 600 mils as you say. However, the 4CX400A operates somewhat differently from the 8930/DX-393 tubes in that it usually draws very little grid current and screen current is also very low at those high output power levels, regardless of band. It does show less thermal drift compared to the 8930/DX-393 on 432 MHz. Unfortunately, I and two other US hams have been having great difficulty with six 4CX400A tubes in both modified AM-615x amps on 144 and 432 MHz, and a homebrew K2RIW-type amplifier on 144 MHz. We are still not certain exactly what is happening and so I do not want to go into specifics. But basically, we seem to have observed the tubes suddenly dropping output power with the current readings becoming very different while attempting to tune up the amplifiers at around 300 to 400 watts output. We are not certain whether the tubes have been damaged due to lack of time to perform more tests. The primary symptom, in addition to sudden output power drop, has been a requirement to reduce the grid bias voltage to obtain the same plate idling current. The required drive power may have also increased somewhat. These recent experiences are in contrast to the first AM-6155 that I modified for 432 MHz some three years ago which gave 500 watts output with 10 watts drive. The owner subsequently had an "accident" whereby he greatly overdrove his amplifier and saw over 550 watts before shutting off. The amplifier and tube survived, however, with no apparent damage. My own 4CX400A tube was bought at the same time that I bought the tube I installed in his amplifier, so it probably was manufactured by Svetlana at about the same time as that other tube. All of the other 4CX400A tubes were bought within the last year, however, and all of these new tubes showed differences in the ink labelling on the anode structures and minor color changes in the pink insulators as compared to my own older tube, indicating possible production line changes through the years at Svetlana. But again, I caution that we still do not know for a fact that we have damaged our more-recent 4CX400A tubes, nor the cause, if so. The feedthrough capacitor and choke that you installed should work fine presuming that the choke is sufficient to prevent significant RF current at 432 MHz from passing through the feedthrough, and that the feedthrough itself is capable of passing that UHF current. I didn't use that technique because of possible clearance problems between the feedthrough and the RF drawer's mainframe in the AM-615x amplifiers. It would be very important to orient the RF choke so that it does not magnetically couple to the existing plate line, or it will pick up significant RF current which must then be bypassed at the B+ end by the feedthrough capacitor. The flat-plate capacitor that I described is capable of passing many amperes of RF current without failure; I have no idea what a typical high-voltage ceramic feedthrough capacitor can withstand, particularly at UHF. The 144 and 220 MHz amplifiers easily put out 500 watts CW with less than 10 watts drive, and I've modified one 432 amp that also would do that power level with that low amount of drive. But usually, it takes 12 or 15 watts to drive the amps to 500+ watts with the 432 modification. That power level is so easy to obtain at 144 and 220 MHz that I guarantee those amplifiers to show that performance at 10 watts or less drive power, and 400 watts minimum on 432 MHz amplifiers. When operating from 124 VAC AC mains (my nominal shack line voltage under load), the plate voltage of the AM-615x amplifier mainframe HV powr supplies is about 2450 to 2500 volts no-load, dropping to about 2000 +/- 100 volts at a plate current of 500 mils, which is usually sufficient for 500 watts output. I presently have one RF drawer which is the best I've ever seen in terms of efficiency and output power; I'm actually reading 600 watts output at 9 watts drive on 144 MHz and the plate current is around 525 to 550 mils (just barely pinned on the original meter). So you can see that the efficiency is still rather low although the output power is there. And no, I don't think it's so good to run the tube at 500+ milliamperes, except for those QSOs that are new grids! :o) Of about 15 amplifiers that I've modified, I believe that all but three have had the Amperex DX-393 tube; the other four had Eimac 8930s. One of those 8930s subsequently arced over on 144 MHz at 550 watts output when the owner fired up into a water-filled 9913 feedline which showed high VSWR. Another 8930 was replaced with a 4CX400A immediately and so I have no history on that 8930 (this was the first 6155 that I modified for 432 MHz). My own original 432 MHz AM-6155 had an 8930 which just barely produced 400 watts output. I have not heard anything from the owner of the amplifier with the fourth 8930. I've been running this 600-watt-output RF drawer since September including almost-daily HSCW skeds at 10000 lpm with W8WN plus heavy contest, meteor shower, and some EME activity. I have not yet seen any sign of tube "aging"; but I also do try not to run the tube over 500 watts except when absolutely necessary. I operate 95% CW/HSCW and very little voice. Before then, I used this same tube in another RF drawer, also on 144 MHz, obtaining only about 525 watts but only 475 mils plate current; my 10 watt exciter would not drive it any further. Now, I know that is because the grid circuit in that original RF drawer is not quite as "good" as the grid circuit in this present drawer. I have found most of the amplifiers that I've modified over the past year to have warped flanges between the plate "cavity" cylinder and the plate compartment walls. Several also had warped grid compartment boxes. These all showed any or some of these symptoms: 1. Unstable output power, usually drifting upward; 2. Drop in output power, sometimes appearing to be thermal drift; or 3. Erratic grid tuning. In each case, I was able to "cure" the problem and make the RF drawer stable with output power by sandwiching a piece of thin, solid hookup wire between the two flange mounting faces, bending the wire around the mounting screw holes where necessary. This was particularly effective is greatly reducing the output power drift as the amplifier heated up at 300+ watts output. In addition, I have received more than 6 brand-new AM-615x amplifiers which had loose hardware in the RF drawer, particularly the screws securing the plate cylinder "cavity" to the plate compartment. I did not attempt to operate those drawers before modification so do not know whether they were actually operative. Several of those drawers also had unsoldered joints in the grid compartment, including one with a loose grid bias choke which lead was barely contacting the grid pin on the socket but otherwise was not soldered! I also found my own Eimac 8930 (in my own 432 MHz 6155, long since sold) to exhibit apparent thermal drift which my DX-393 did not show. I have never seen thermal drift using either tube except on 432 MHz. I've set the filament voltage to several 432 MHz amplifiers to 5.5 volts using series resistors; but because I could not anticipate the actual filament voltage at a customer's installation, I stopped doing that. The nominal filament voltage I usually see is 5.9 to 6.1 volts when my AC mains is 124 VAC under load. As for grid current: I have only recently begun including grid current measurement capability in modified AM-615x amplifiers. But with the original or suggested-modified grid bias circuits, one can always see when one was drawing grid current because the grid bias will go more negative very easily unless you are using a variable shunt regulator. Originally, I used the same kind of bias regulator that you mentioned: zener diodes with a 10K pot across the zeners. This was insufficient to provide regulation when the tube draws grid current, as you can see with the simple calculation below. Assume the Eg bias is -100V at the top of the pot, which bottom end is grounded, the wiper of the pot going to the tube grid. Assume you have set the pot wiper to provide -75 volts grid voltage. So the resistance from the tube grid to the top of the pot is now 2500 ohms while the resistance from the tube grid to ground is 7500 ohms. Now, assume that you draw 10 mils grid current through the top of the pot: what is the voltage differential: E = I*R = 0.010 * 2500 + 25 volts!! So the grid will "float" to -100 volts when drawing 10 milliamperes. Of course, one is not "supposed" to be drawing grid current when on SSB, right? So normally, one would not see the Eg bias voltage change with "normal" drive power to the tube where Ig is zero. However, the original AM-615x amplifiers do not provide grid current metering so the only indication one has of any grid current is observing that the grid bias is being driven increasingly negative. At 432 MHz, my experience has been that to drive the 250B-type tubes (of which the 8930 and DX-393 are derivatives) to more than 300 watts output, SOME grid current will be drawn, usually less than 10 milliamperes. This will, using the typical potentiometer circuit, cause the grid voltage to go more negative, which will require more drive power since the tube is being driven further into class C, which drives the grid voltage further negative, requiring still more drive power. So, you can see why I quickly changed from the original tapped potentiometer to the shunt regulator. My own circuit is very similar to that which G3SEK and his G4 friend developed except that the grid regulator doesn't need to have any of the fault detection circuitry that their tetrode screen grid regulator has. I will try to describe it simply, but can also attach a JPEG image of the schematic if you would like. In these AM-615x amplifiers, the main unregulated bias supply is about -120 volts under load of the shunt regulator and around -128 volts with no load. The shunt regulator that I use regulates in the POSITIVE side of the bias supply output; so the positive side of the unregulated bias supply goes into the current-limiting resistors while the negative side of the supply goes directly to the tube grid. The shunt regulator circuit, therefore, uses the negative side of the bias supply as its "ground" reference. At the output of the shunt regulator, the positive side is grounded directly; or if you wish to include grid current metering in the AM-615x amplifiers, you can ground the positive output of the shunt regulator through an 18 ohm resistor, and run an extra wire from the positive output side of the shunt regulator to the front panel metering circuit, which will have to be modified to measure the voltage drop across that resistor. The meter in the AM-615x amplifiers is a 50 microampere movement and is 7000 ohms total resistance; thus, a voltage of 0.35 volts is required for full-scale deflection. With the 18 ohm resistor in the grid bias regulator circuit to ground, full-scale will then occur at about: I = E/R = 0.35/18 = 19.44 milliamperes grid current. Since the shunt regulator circuit is entirely before that resistor, none of the shunt regulator current is shown on the meter, only the actual tube grid current. There are two dropping resistors from the positive output of the unregulated supply: one is 4.7K/5 watts, the other is the main dropping resistor for the shunt transistor and can be from 1.7K to 2.7K at 10 watts. A lower value (1.7K) will provide higher shunt-regulator idling current. The limitation, in the AM-615x amplifier bias supply, is that the bias supply transformer is rated at only 25 milliamperes. But this is a military-rating. At low output voltages, this shunt regulator, unfortunately, does draw more current than the transformer rating. However, this is of consequence only when the grid bias level must be low such as for a 4CX400A, which typically requires about -35 volts compared to around -70 to -80 volts for an 8930/DX-393. In addition, the shunt regulator is turned off during RX mode so the overcurrent condition only occurs while in transmit. Still, this is a concern and I hope to find a replacement transformer. The 4.7K/5W dropping resistor provides low DC operating voltage for a dual op-amp type LM358, which is the same chip as the quad op amp LM324 (I believe the LM748 is the basic single chip but have not used that type). The LM358 is NOT frequency-compensated, one of the differences between my circuit and that of G3SEK. The 4.7K resistor's output is regulated to +24VDC (can be as low as +12VDC and as high as +33 VDC, I think, not important) by a zener diode; I use a 5 watt diode. This is filtered by a 0.01 mF and 4.7 mF capacitor for zener noise reduction, and is the Vcc for the op amp. The +24VDC regulated supply is dropped further through another resistor, 1.8K 1/8 watt, to a 4.7 or 6.0 volt zener diode, 1 watt type. This is also filtered with 0.01 and 4.7 mF capacitors to reduce zener noise, and is the reference voltage for the op amp. This reference voltage drives the op amp inverting input through a 10K, 1/8 watt resistor. There is also a feedback resistor from the op amp inverting input to the op amp output to limit the DC gain, value uncritical, typically 220K to 330K or so. The op amp will oscillate without some feedback. The output of the op amp is current-limited through a 1K, 1/8W resistor to drive the base of a high-voltage NPN bipolar transistor such as a TO-220 package BUX-85. The power dissipation in this shunt transistor is never more than several watts regardless of the set output voltage. The emitter of the transistor (in addition to the "ground" of the op amp) is connected to the "ground" common of the two low voltage zener supplies, which is actually the unregulated bias supply's minus output. This current-limit resistor is necessary since the base-to-emitter voltage of the transistor can never be more than about 0.6 to 0.7 volts. The collector of the shunt transistor is connected to the output side of the previously-mentioned 1.7K-to-2.7K, 10 watt resistor. To provide a sample of the output voltage to feed back to the op amp for comparison against the 4.7V or 6.2V zener reference voltage, I use the original 10K pot with a 75K, 1/4W resistor to the collector of the shunt transistor on the top end of the pot, and a 4.3K, 1/8W resistor from the bottom of the pot to ground. It is only necessary to have an adjustment range across the pot of a few volts centered around the reference zener voltage (which can be your choice, not critical; I use either 4.7 or 6.2 volts depending upon which zeners are readily available). I sometimes change the 4.3K resistor to limit the bias voltage adjustment further. With the 75K-10K-4.3K and 6.2V zener, I can adjust the grid bias between about -35 volts and a most-negative voltage of almost -120 volts. You can also use a MOSFET such as the IRF830 or IRF840 (or any rated at 150 volts and several hundred milliamperes), as the G3SEK circuit does; but in that case, the output voltage to the gate of the MOSFET will have to be around 2.5 volts to a maximum of perhaps 4.05 volts or so, and you may have to use a voltage divider network as the G3SEK circuit does. I've found the IRF840 to conduct very heavily with a gate voltage of just 3.6 volts. Since the output voltage of the op amp can be so low, the Vcc applied to the op amp can be much lower than the +24VDC that I usually use. I use +24VDC because I have 24V/5W zeners available but no lower-voltage, high-power zeners. Remember that this Vcc zener-regulated supply must provide sufficient current for both the op amp and for the reference zener. At the same time, one wants to keep this current draw as low as possible due to the already-existing excessive-current load on the bias transformer. Since no current is required from the reference zener supply, the current through it can be set to 10% or 5% of its maximum rated power dissipation or just a milliampere or so. I set it relatively low to avoid heating the zener unnecessarily and causing voltage drift due to self-heating (which shouldn't be apparent anyway since the zener always draws the same amount of current whether in TX or RX mode). To prevent the circuit from oscillating, it is absolutely necessary to connect at least a 1 mF/150VDC capacitor from the collector to the emitter of the transistor. I found this capacitor to also be necessary when using an IRF840, because I used the LM358 non-compensated op amp instead of the LM741 type in the G3SEK circuit. Bypass capacitors on any other op amp pins or the transistor base only caused more severe oscillation. In addition, a bypass capacitor across any of the three output voltage divider sensing resistors (the 75K/10K-pot/7.5K resistors) mere exacerbated the oscillation problem. If the capacitor is not large enough, the circuit will oscillate at low output voltages (when the shunt transistor is conducting most heavily); if the capacitor is too large, the grid bias will not change level quickly between transmit and receive and there is the risk of a current spike through the shunt transistor which could destroy the transistor. To change the bias voltage between TX and RX values, one can short the base of the transistor, or the gate of the MOSFET, to the emitter, or source. Since both leads are at almost the full negative bias voltage, this requires an isolated relay contact. I consider this to be one of the two major shortcomings of this circuit, that I cannot switch it by simply grounding a lead somewhere. I have a couple of ideas on doing that but have not had time to test them. Another shortcoming of this circuit, and of that of the G3SEK circuit, is that when the desired regulated output voltage is nearly that of the unregulated input voltage, the shunt regulator is necessarily drawing minimum current, and so one cannot *SOURCE* as much grid current as with lower grid bias voltages. I am also working on another circuit which combines a series-pass regulator with the shunt regulator to overcome this problem. At 432 MHz, I have observed the 4CX250B-type tubes to SOURCE current at times rather than DRAW current. This causes the grid voltage to DECREASE rather than increase. Since I only see this happen when the tube is being driven very heavily and with lots of output power, I believe that it may be thermal emission from the control grid, just as the screen grid can do. I believe that this is what caused the one 8930 tube, and a pair of 4CX400As, to arc over; the tubes essentially were driven into "thermal runaway" whereby the bias voltage drops and the plate current rises. If the output power continues to increase so that the cathode emission continues to rise, the grid heats further which drives the grid voltage even lower, further increasing the plate current, which further raises the output power. And so on. I don't believe this happens at lower plate voltages such as below 1500 volts; but I have "killed" some 4X150As operating at 1500 volts when they appeared to go into runaway. I used to think this was a result of the cathode emission increasing as the plate current continued to increase; and perhaps it is. But at the same time, I've noticed the grid current go NEGATIVE, sourcing current to the grid bias power supply. On the AM-615x amplifiers operating on 432 MHz, the plate output coupling is almost at minimum, and often right at the stop. You can obtain a bit more output coupling reduction by removing the rubber washer spacer located inside the gear box which adjusts the plate output coupling. I have found best efficiency and best thermal stability to occur in almost all 4CX250B-type amps when the screen current is at or within several milliamperes of zero. This is not possible to achieve with the AM-615x RF drawer unless that rubber washer is removed. In addition, I have found that replacing the original output coaxial cables which originally connected to the output power sensor and low-pass filter with a single length of RG-213 double-shielded cable reduces the total power loss between the output connector of the plate compartment and the plug-in connector on the rear panel of the RF drawer. In your case, Peter, since you are not using the original mainframe, you can simply disconnect the output cable from the power sensor and filter, connecting it directly to your TR relay. For those with the original AM-615x amplifier mainframes, it is necessary to remove the output coupling gear assembly by unscrewing the output coupling capacitor plate inside the plate compartment, then unbolting the gear box assembly on the outside. The coupling capacitor plate unscrews easily with finger pressure if you loosen the locknet between the plate and the connector. DO NOT USE PLIERS TO GRASP THE CAPACITOR PLATE, or you run the risk of destroying the yellow Kapton insulator. Should the insulator fall off the aluminum disk, I do not know what kind of cement one can use to reattach it. I have enhanced the high-voltage insulation between this output coupling disk and the anode ring coupling surface by cutting a piece of teflon sheet about three inches long (about 4 to 5 cm) and about 1 inch wide (about 1.5 cm), then punching small holes in each end of the sheet. Then I place the sheet with the threaded center conductor of the output assembly going through the holes in the teflon such that the teflon forms a ring of insulation around the original coupling disk. This is one technique which allows very high output power on 144 MHz where the coupling capacitor plate spacing to the anode ring must be very close, usually less than 1/8 inch (about 4 mm). Once you have removed the coupling capacitor gear box assembly, remove the brass gear from the assembly, and loosen the cable lock nut and move it away from the assembly so that you don't have to heat it. Then with a large soldering iron, heat up the assembly to remove the small disk which prevents access to the center pin. Do not worry about overheating the assembly but do not use a flame; a large soldering iron and/or gun is best. With the disk removed, you will have to heat the center pin while pulling the old coax cable out of the assembly. You will notice melted plastic as the cable is pulled loose; this is ALL from the coax itself, not the assembly. Once the cable is pulled out, you can use tools to clean out remaining plastic and solder stuck inside the assembly. In 15 amplifiers, I have never overheated or destroyed one of these assemblies while using just a soldering iron and soldering gun. Once the assembly is clean, you can prepare the new cable; it should be between 15 and 18 inches long (less than 0.5m) but it should not be any shorter or you will have to bend the cable too sharply in order to bring the far end back to the original rear-panel connector. Installation of the new cable into the gear assembly is easy and in reverse procedure to disassembly. However, when reinstalling the small disk covering the hole in the assembly, it is imperative that there be no excessive solder build-up on that face of the assembly because if so, it will tend to rub against the cast metal assembly bracket, making front-panel adjustment very difficult and "sticky". If necessary, use a flat file to smooth excessive solder from that face. Replacement of the original rear-panel push-on connector onto the new cable is just like a normal N-type clamp connector and needs no explanation. 73, Steve K0XP
From Peter Hansen : (PH@Focon.dk) >Hello outhere >I got an Am6155, I have converted to 70Cm. I have used Ron W3RJW method. I >have made some further changes, but only to the better. My problem is the >efficiency, it is very poor only something like 40%. Have anybody any >sugestions wath to do. Mayby someone knows wath efficiency to expect from >this amplifer on 70CM. > >I do only have the amplifier section so I build my onw powersupply. > >My working conditions is. > >Anode voltage 2050V loaded >anode current 400mA >Output 320W hf >Screen 350Vdc (regulated) >Idling current 80mA Efficiency of the AM-615x plate tuned circuit is very poor on 432 MHz. It increases above 40% as output increases beyond 400 watts but never gets as good as 50%. I believe the problem is the extremely short length of plate line remaining within the "cavity" at resonance combined with about half of the total tuned line being composed of the anode connection ring and plate blocking capacitor assembly. For realiable operation at the 400+ watt level, it is necessary to remove the original center-fed plate B+ line from within the tuned plate line tubing, build a new B+ blocking capacitor on the outside wall of the plate compartment using 0.01" (2.5 mm) teflon and a sheet of PC board or other smooth metal tightly compressed against the plate compartment wall by half a dozen through-screws threaded into the wall. B+ can then be fed to the anode via a small 4-1/2 turn choke soldered to a corner of the silver-plated anode ring (drill a small hole to wrap the wire through then solder to the silver on both sides of the ring). Drill out the hole which has a screw threaded into it to plug the hole; this was used on the AM-6154 to mount an extra gimmick plate capacitor but is plugged by a screw in 144/220 MHz modifications. Using a piece of large coax cable center insulation through the hole, a screw is connected to the new external bypass capacitor plate, fed through the insulator and the new anode choke soldered or attached to the screw inside the plate compartment. This modification will withstand at least 550 watts; I've never had it fail up to that power level although I strongly suggest that because of the remaining poor efficiency,output power be limited to 400 watts under normal operating conditions. This mod is detailed and shown on N1RWY's AM-6154/6155 pages at: http://www.umecut.maine.edu/~baack/AM6155/432opq.html There are many other hints and tips on RWY's pages including links to all known web modification pages. W3RJW's grid circuit mod works very well and the plate circuit mod is necessary to get output power above a couple hundred watts consistently. But complete replacement of the original anode B+ plate choke is necessary for reliable operation at the 400 watt level and above. By the way, the 144/220 MHz amplifiers, when putting out 550+ watts CW, show an exhaust air temperature of at least +53 degrees Centigrade at an output efficiency of about 55% with the original high speed blower. Part of this temperature rise is due to relatively high air inlet temperature (probably 5 to 10 degrees C. above ambient) since the blower is mounted inside the mainframe and sucks air past the power supply and other modules before pressurizing the amplifier's grid compartment. It is also absolutely necessary to regulate the grid bias using a shunt regulator; the tube will otherwise drive the grid bias further negative from the nominal -75 VDC for 80-100 mils resting current to over -110 VDC at 10 watts drive on any band. I have a simple circuit similar to G3SEK's screen regulator which I've been using for many years which keeps the grid bias stable within a few tenths of a volt. 73, Steve K0XP
>Hello All, >What is the typical 8930/DX393 plate current on 222 and 432 MHz? >Please reply direct. Thanks & 73, >Tony WA8RJF EN91 QRV: DC, 6 - 2304 & 5760 Tony, way over 500 mils! But unless you've moved the two diodes across the meter itself instead of in front of the 4320 ohm meter multiplier resistor, the diodes will conduct and shunt the plate current, preventing the meter from reading any higher than about 300 mils. This mod was discovered by KF0M in Wichita a few years ago and is not very widely-known. And some remove it because under normal conditions when output power is over 400 watts, the plate meter is pinned and can't be read anyhow! It's easy to check this out using an external meter; simply connect the external meter in series with the thick red wire with the spade terminal that's connected to the rear post of the large oil-filled capacitor in the HVPS. Make certain that you place the external meter on a thick book or something as it will be at the plate voltage of 2400+ volts no-load! And do NOT key up the amplifier if the wire through the meter comes loose; you could blow the screen grid of the tube with no plate voltage. On idling with no RF drive, set the grid bias for 90 or 100 milliamperes resting current. All of these numbers apply also to the Svetlana 4CX400A. 73, Steve K0XP
Date sent: Mon, 08 Mar 1999 04:50:22 +0000
To: cjack cjack@cchat.com
From: Steve Harrison ko0u@dandy.net
Subject: Re: am 6155
>hello
>i am hoping that you will be able to help me with a problem.
>i have modified a am6155 for 432 and it work for a few days. now when i
>key it, it the meter goes up to about 100 watts then starts dropping off
>rapidly. i have tried every mod i can find with little or no change.
I write this paragraph after having written everything that follows. Now
that I think about it, I strongly suspect that your real problem has to do
with the center rivet in the circular disk of the plate line; see below
for more details.
This sounds very much as if something has happened to the plate line. When
modified for 432 MHz, one of the mods involves replacing the original
plate choke, located inside the 1/2" diameter tubing inside the plate line
"cavity", with one wound on a small resistor and then placed inside a
teflon "cocoon". This new choke is then located at the far end of the
plate line from the tube, at the original screw-in plate DC feedthrough
capacitor. here are three (or more) possible trouble areas that have
caused this type problem on amplifiers that I have modified:
1.When I have used the original red teflon-insulated to connect between
the new choke and the feedthrough at the tube end, I have experienced
arc-over between the teflon wire and the inside of the 1/2" diameter
tubing. This would appear, at first, as a drop in output power as you
describe, but eventually would become a solid arc that shorted the high
voltage to ground, blowing the HV fuse. In the worst case, there was no
more than about 25 watts output power! 2. On several amplifiers, I have
experienced a loosely-fitting rivet that the original plate HV line is
soldered to in the center of the original plate blocking capacitor disk. I
tried tightening this rivet using a center punch but the fix did not last,
and I eventually had to completely bypass the rivet assembly entirely by
going to the new DC plate bypass capacitor assembly described below. The
loose rivet does not then affect the RF output power of the amplifier. 3.
Another possibility (although unlikely) is that the original DC plate
blocking capacitor, which is the circular yellow Kapton insulator
sandwiched between the anode ring and the plate line inside the "cavity",
is not tight and so is loosening further with heat as you transmit. This
assembly is held together with two 4-40 screws that pass into two teflon
shoulder washers on the anode ring side of the line. The problem can be
caused by not tightening the two screws sufficiently, or possibly by
having assembled the rebuilt line with the relocated plate choke such that
tightening the two screws does not pull the two round plates of this
capacitor together sufficiently. This problem might also be caused by
broken plastic insulator pedestals that support the center of the plate
line within the "cavity" itself.
I've seen two varieties of these insulators: one is a light-colored
yellowish-tan, and the other is a dark brown color. I've found the
light-colored insulators to be extremely fragile and prone to cracking and
shattering very easily if the plate tuning control is run against the stop
so that the plate line short butts against the insulators (in comparison,
the dark-brown insulators have appeared to merely crack). I've been able
to repair such insulators using ordinary epoxy cement, even building up
missing and shattered pieces with the epoxy. To the best of my knowledge,
the two amplifiers that were repaired in that manner are still operating
but the owners have been repeatedly cautioned NOT to ever allow the plate
tuning to be run against the ends of the tuning range again or the
insulators may again be broken, perhaps irrepairably.
As an aside, a little over three years ago, I read of a couple of
Californians who actually machined new insulator pedestals from solid
teflon blocks. But I've not been able to find that old e-mail posting nor
who the two guys were.
I've found only one of the dark brown insulators to be broken, and that
one was a simple crack such that the insulator was still whole and able to
maintain mechanical rigidity of the plate line; so I did not perform the
epoxy repair. But again, the owner was cautioned not to run the plate
tuning against the stops. In retrospect, I believe that I should have used
epoxy to repair the insulator. But at the time, I had not tried that fix
before and the cracked insulator appeared able to maintain mechanical
rigidity without cement. When I did try the epoxy repair later, it was a
last resort because the two light-colored insulators were totally
shattered and unusable. I had nothing to lose!
The fragility of these insulators requires extreme care in reassembling
the plate line. First, the two screws holding the outer circular disk
against the plate line (with the circular Kapton insulator) are tightened
just beyond finger tight; the objective is to compress the split
lockwashers but no more or the teflon washers will be deformed. Then the
two plastic centering insulators are laid in place, and the outer circular
disk is rotated by hand so that its mounting shelf will fit flat against
the mounting shelf on each insulator. The mounting hardware for the
circular disk to the insulators are tightened finger-tight to that as the
insulator mounting screws are tightened, the circular disk hardware will
slip to allow the disk's mounting shelf to fit snugly flat against the
insulators. There must be absolutely no stress placed on the plastic
insulators whatsoever other than the compression caused by the mounting
screws at each end of each insulator. These insulator mounting hardware
are then inserted and carefully tightened by finger. When all appears to
fit properly, the hardware for the insulators is carefully and alternately
tightened, and then the two screws for the circular disk's mounting shelf
are tightened slightly. If all still appears to fit neatly with no stress,
then all six screws may be snugly tightened, but not so tight as to crack
the insulators.
To install the anode ring, place the ring inside the compartment and then
start the two securing screws into place. Then insert the tube which will
center the anode ring around the tube. I have found that in most
amplifiers, the anode ring does not fit exactly perfectly; the mounting
holes really need to be elongated by perhaps another 0.030" inch to be
perfect. But I've never actually done this as things seemed to be "good
enough".
4. It is possible, but not likely unless you are also blowing HV fuses,
that the Kapton insulator itself has punched through or deteriorated in
some other manner. I have replaced this insulator with ordinary 0.0055"
thick teflon with absolutely no problems whatsoever despite the
experiences of some others, who have said that teflon has caused problems
in their amplifiers. However, I have done this only on 432 MHz amplifiers,
not any for 144 or 220 MHz. The Kapton insulator actually has a higher
voltage breakdown rating than teflon for the same thickness, and is also
apparently as good at high RF current densities such as in these
amplifiers. The dielectric factor of Kapton is also somewhat higher than
teflon which means the existing Kapton insulator has more capacitance than
the equivalent-thickness teflon insulator will. So there may be something
to the claims that replacing the Kapton with teflon causes some
performance deterioration, particularly in 144 or 220 amplifiers.
I have never seen a failed Kapton insulator, although I replaced one
because of a punch-through caused by a tiny piece of foreign material, and
another out of curiousity to see how the teflon affected the amplifier
performance (no difference that I could tell on 432 MHz). I see no reason
to automatically replace the original yellow Kapton material unless it has
a high-voltage punch-through.
The following is a description of how I replaced the original plate line
choke and plate DC bypass feedthrough capacitor with one that turned out
to work very nicely, even at 500 watts continuous output power from a new
Svetlana 4CX400A on 432 MHz.
Initially, to cure arc-over from the original red wire through the center
of the 1/2"-diameter plate line and the inner surface of the line itself,
I replaced the wire with a length of center conductor from 1/4" teflon
coax such as semirigid cable. Because the center conductor of this coax is
magnetic steel (although silver-plated), I also pulled out the original
steel wire and fished solid copper magnet wire in its place through the
center hold in the teflon dielectric. This cured the arc-over, but the
amplifier still experienced severe output power droop under load. Some
hours later, after intense scrutiny of every part of the plate line, I
finally discovered that the center rivet through the aluminum disk that
forms one end of the plate blocking capacitor, was loose in the disk. This
rivet is brass and apparently was originally pressed into the aluminum
disk. The rivet, in this case, became loose apparently with repeated
high-temperature soldering operations while replacing the original plate
choke in the plate line. As described later, attempts to tighten the rivet
did not succeed and so more-drastic measures were required to bypass the
rivet altogether.
But in addition to the loose-rivet problem, I have also had the choke
assembly itself burn up at only 400 watts output power. This original
choke is located within an area of extremely high RF voltage gradiant
between the original red teflon wire and the inside surface of the 1/2"
diameter plate line. If the red wire is connected between the circular
disk at one end of the line, and the plate voltage feedthrough capacitor
at the front of the "cavity" such that the wire is essentially centered
within the 1/2" plate line, then arc-over problems will be minimal. But on
432 MHz, this long plate line apparently has RF voltage minima and maxima
all along its length, and it is common for the red teflon wire and/or the
relocated plate choke itself, to arc to the inside of the plate line. The
complete cure requires eliminating these two potential weak points
entirely.
To this end, what I have done on all further 432 MHz conversions is to
relocate the plate choke at the anode and build a low-inductance plate
voltage bypass capacitor on the outside wall of the anode compartment.
This is done by installing a small choke that is soldered to an edge of
the original tube anode ring. This new choke then connects to a homebrew
high voltage plate bypass capacitor which is made by sandwiching a sheet
of single-sided PC board (I usually use PTFE-glass board scraps) against
the outside flat surface of the tube anode compartment, insulated from the
anode compartment with a sheet of 0.0055"-thick PTFE (available from
McMaster-Carr, but not cheaply!).
A screw through the board and then through an enlarged hole in the side of
the anode compartment (on original AM-6154 RF drawers, this hole was
originally used to mount a small aluminum plate against the inside wall of
the anode compartment to provide a small amount of capacitance in parallel
with the tube anode for better efficiency) projects into the anode
compartment and the new, very small anode choke is soldered to a solder
lug which is sandwiched between two nuts on the end of the screw. The
original red HV wire that was connected to the large HV plate feedthrough
at the front of the "cavity" is then connected to the screw on the outside
of the anode compartment. The total surface area of the new bypass
capacitor plate is about 6.0 square inches. The plate is 3.25" wide by
2.625" high with a rectangular area cut out of one corner that is about
1.50" wide by 0.75" high. See the drawing below:
|<--------------3.25"---------------->|
|<--- 1.50" -->| |
_________ _______________________________________
^ ^ | / / / / / / /| o o |
| | |/ / / / / / / | |
| | | / Cut Out / /| |
| 0.75" |/ / / / / / / | O < Existing |
| | | / / / / / / /| hole through |
| _v_ |______________| anode comp. |
| | wall |
2.625" | o o o |
| | |
| | |
| | |
| | o o o |
_v________ _______________________________________
Note: Outside view shown. Board is single-sided PC board, either
PTFE-glass or ordinary FR-4 (G-10) board, 0.030" or greater thickness.
Board is sandwiched to wall of anode compartment using the eight screw
holes shown above by the small letter "o". This requires that copper be
cleared at least 0.20" around the edge of the screw holes.
Copper-clearance absolutely MUST be with no burrs whatsoever as any burrs
WILL result in punch-through of the teflon insulating sheet.
Construction Technique:
1. Cut out a board to the dimensions shown above.
2. Remove the tube and anode ring BEFORE attempting to enlarge the
existing hole through the side of the anode compartment. The hole may be
enlarged to 0.375" but should not be made any larger. 3. Carefully deburr
BOTH edges of the new, enlarged hole. Use a fine-grade piece of sandpaper,
at least 200-grit, on the outside of the anode compartment wall to smooth
any burrs or projections. If there are casting imperfections on the wall
(these are common and can be seen fairly easily), try to smooth them with
a fine-tooth, wide file. Try to keep metal filings and dust from the
inside of the anode compartment. 4. Center the board against the anode
compartment wall; it should fit neatly and "feel right"; i.e., it should
feel as if it is fitting flat against the wall. If not, determine the
reason and using the file, try to smooth any problem areas on the wall. DO
NOT FILE ACROSS THE PC BOARD! 5. With the board centered on the wall, use
a sharp short scribe or pencil and from the inside of the anode
compartment, scribe the newly-enlarged hole onto the inside surface of the
board. Locate the exact center of this scribed circle on the PC board. 6.
With the board still held against the wall, carefully mark on the board
where to locate other mounting holes that will be used to hold the board
against the wall. Drill out these holes in the PC board, including the
newly-scribed centered hole, using a small drill bit such as 1/16". The
smaller this bit is, the more precisely you can locate the mounting holes
through the anode compartment. Drill out the centered hole large enough
for a 4-40 screw. 7. Use a long 4-40 screw, mounted through the board and
compartment wall, with a set of large washers and a nut inside the anode
compartment, to securely hold the board against the outside of the wall.
Make certain that the screw is centered within the anode compartment hole.
With the board securely held against the wall, carefully place a drill bit
inside each hole of the board and drill through the anode wall. I used
2-56 screws to hold my boards, so I drilled the anode compartment holes so
that I could thread the holes for 2-56. 4-40 will also work but will take
away a little more area of the board's copper surface, slightly reducing
the total bypass capacitance. On 432 MHz, this is not significant; but if
the amplifier is to be used for 144 or even 220 MHz, the capacitance could
become relatively small. 8. Thread the anode compartment holes for your
chosen hardware size. If unable to thread the holes, you may try drilling
the holes to clear the screw threads. In this case, all the mounting
hardware located within the anode compartment must be non-magnetic, such
as brass (I have not tried using stainless steel hardware inside the
compartment but am reluctant to try that). If you MUST do clearance holes,
then the location of the holes must be carefully chosen so that hardware
inside the anode compartment will fit flush against the inside wall; this
is why I threaded the holes rather than try to to use such hardware inside
the compartment. 9. If necessary, drill out the board holes to allow easy
clearance of the screws. In my case, I had pieces of 1/4" OD Teflon rod
available from which to make shoulder washers. I drilled the holes in the
board to 1/4" diameter, then drilled a centered hole into about 1" of the
teflon rod to pass the 2-56 screws. Then I sliced thin washers of the rod
which I placed in the 1/4" holes in the board. These washers provide some
extra high-voltage insulation between the grounded screws and the copper
of the boards. 10. You can try using mylar for an insulation sheet, but it
should be at least 0.010" thick. Cut the insulating sheet at least 1/4"
larger than the board. I had some 0.0055" thick teflon sheet which I have
used. Lay your insulation sheet against the board and carefully, with a
sharp scribe, punch holes into the sheet where the screws will pass
through (this is easy to do if laid against a soft wooden block). Using an
Exacto knife or other small circle punch, enlarge the holes to pass the
screws with no excess material folded back when the screws are inserted. I
cut X marks centered on the holes and then cut out each tiny triangular
piece with the Exacto knife. In another case, I found a piece of brass
hobby tubing about 1/8" diameter and sharpened one outer end with a file
then used the tubing as a punch to cut clean holes through the teflon. 11.
Using the flat file, very carefully wipe the file across the copper at the
edges of the PC board. The idea is to completely eliminate any sharp edges
which might cut into the insulating material. Finally, using 400-grit
sandpaper or steel wool, carefully clean the copper side of the PC board.
When done, you should be able to rub your fingers across any copper edge
without feeling any burrs or sharp edges whatsoever. ANY such sharp edges
WILL punch through the insulating material the first time that you turn on
the high voltage. Experience speaks! I have yet to make one of these
capacitors and not have it punch through the first time. Several have
required three or four complete disassemblies and clean-ups before they
did not punch through. 12. Perform a preliminary assembly of the board
against the anode compartment wall WITHOUT using the insulation and make
absolute certain that everything seems correct, including that the screw
projecting through the wall is centered within the enlarged hole. This is
a good time to choose the final assembly hardware. The idea is to use
screws that do not project more than one or two threads into the anode
compartment; otherwise, they could become points at which a high RF
voltage gradiant will form and where corona can form under high output
power. To trim screw lengths, either use a very sharp screw thread cutter
or sharp dikes, then use a file to clean up the messed-up threads. By
placing a nut over the screw before cutting it, the nut will also tend to
straighten the threads as it's backed off. 13. It is now time to make
hardware with which to ensure that the screw to which the plate choke will
be connected will contact the PC board copper. I try to use 2-56 hardware
for this, also, but usually do not have a screw long enough. In any case,
I find a screw long enough to pass through the anode wall and the board
(with the insulation in place) and which projects pass the inner edge of
the anode compartment wall by up to 1/8", and through the PC board
material on the outer side by about 1/4". I cut the head off the screw so
I have a threaded stud. Then I use a washer, solder lug and nut on the
outside wall with the smallest nut I can find against the PC board copper
surface. This holds the threaded stud in place and contacts the copper
surface. Then I use another short length of the teflon rod with the hole
drilled through, perhaps 3/16" or 1/4" long, and place it over the
threaded rod where it projects through the enlarged hole in the anode
compartment wall. If I can find it, I then use a very small washer against
the end of the threaded rod, a lockwasher and very small nut to tighten
the teflon rod against the board. Then I usually use a very small solder
lug and another very small nut to hold the solder lug in place. The choke
is soldered to the solder lug after assembly is finished. It will be
necessary to leave off the solder lug and last nut until the board is
assembled against the wall for the last time. Sometimes, I cut out a small
rectangular piece of teflon sheet and work it into the anode compartment
wall's hole around the threaded stud to help provide still more high
voltage insulation for the washer, lockwasher and nuts that are inside the
compartment. 14. When all the hardware is ready, the board and anode
compartment wall are cleaned up with steel wool once more, and the
insulating material is carefully cleaned to make certain there are no
particles that will punch through the insulator, then final assembly can
take place. If you have any sort of high voltage hi-pot tester, you can
use this before soldering the choke to the solder lug to find any trouble
spots, which will show up with an arc, leaving soot or carbon tracks. This
saves repeatedly blowing HV fuses and possible other damage. The very last
thing to do is to carefully locate and bend the inside solder lug so that
the choke can be soldered. Usually, the solder lug is located so that it
is about 1/8" from the plane of the inside of the compartment wall, and
pointed up toward the outside of the compartment (where the top cover is
attached). 15. For a choke, I use something like #20 solid hookup or
magnet wire wound over a 1/8" or 3/16" diameter drill bit shank, about 4
or 5 turns (not especially critical but the more turns, the better;
however, there is not much room for the choke so the restriction is
usually around 4 or 5 turns). I drill a 1/16" hole in the corner of the
anode ring, very carefully deburr the hole without cutting silver plating
away, and then attach one end of the choke to the hole by fishing it
through and wrapping the wire around and over the top edge to make as
secure a mechanical connection as possible. Then I carefully solder the
wire to the silver plating of the ring. There must be NO sharp points,
either the wire itself or solder, on the connection. The choke is arranged
so that it roughly points upward toward the solder lug. Then the choke is
soldered to the solder lug itself.
You can perform a final hi-pot test of the assembly by leaving the tube
out.
Before reinserting the RF drawer into the mainframe, you must lay the
outside solder lug back against the PC board. As you slide the drawer into
the mainframe, you will have to jiggle and slant the drawer so as to clear
the new screws and solder lug on the outside of the wall. The new
capacitor plate is located on the top of the anode compartment wall, and
by removing the square plate on the RF drawer side of the mainframe, you
can see how much clearance you have between the new capacitor and the
inside top of the RF drawer area. There is no clearance problem here
assuming you are able to fit the drawer into the mainframe to start with.
>my grid current is at or above 200 ma when keyed.
I presume that you mean plate current. There is another modification by
KF0M which relocates the diodes that are originally wired across the meter
so that the diodes do not conduct. As originally wired, one of the diodes
is wired on the far side of the meter multiplier resistor and will begin
to conduct when the meter reading reaches slightly more than one-third
scale. By the time that the meter reading, in the plate current position
at any rate, is past center-scale, a large portion of the meter current is
being shunted by the diode and so the meter is actually reading very low.
I've found that when the indicated plate current is only 300 milliamperes,
the ACTUAL plate current is well past 500 milliamperes! When the meter
protection diodes are rewired so that they are connected across the meter
itself, they do not conduct and the meter reads correctly. This
meter-reading anomaly occurs on other scales besides plate current.
From: ko0u To: Wayne Heinen Subj: AM-6154/6155 PLATE FEEDTHRU CA The AM-6154/6155 contains two plate B+ feedthrough capacitors, both made by Spectrum Control, Inc. (that's what the "SCI" means on the side of the hex mounting surface!). One is located on the power feedthrough plate mounted on the side of the RF drawer compartment while the other is on the front of the plate "cavity". As many of you have told me, the one on the cavity often blows at high power on 432 MHz; and to date, everybody who has lost it seems to have put the cavity aside for parts. Newark stocks some of the more popular SCI feedthru caps, but I don't see this particular one listed. HOWSUMEVER!! I DO see a 2500VDC, 1500 pF cap, made by Murata (even though it's listed on the same page as SCI, the part number identifies it as being a Murata product) listed as P/N 1280-060 (Newark P/N 19F689). The illustration in the catalog is nearly impossible to make out, but it looks to me like it shows the thread as being 3/8"-32, which is, I believe, what the SCI feedthrough actually is. The bad news is the cost: $24.03! I'm waiting for a response from my local SCI distributor on availability/cost of the original SCI feedthrough. W6GGV made the point to me recently that these large, high-voltage tubular feedthrough capacitors no doubt do not work very well as a purely-capacitive AC blocking capacitor at 432 MHz. I intend to measure the S11 and S21 response of one on the network analyzer soon to see where the resonances are. Hank suggested replacing the feedthrough with the usual flat-plate bypass cap using teflon as the dielectric. Unfortunately (of course), the gears on the front of the 6154/6155 plate cavity get in the way of making that capacitor as large as it should be (I'll have to calculate its required size vs. teflon sheet thickness tonight). But that may still be a way to replace the original feedthrough for you guys not wanting to enrichen Newark by a hundred+ two-bit pieces. 73, Steve Ko0U/1
> > The AM-6154/6155 contains two plate B+ feedthrough capacitors, both made by > Spectrum Control, Inc. (that's what the "SCI" means on the side of the hex > mounting surface!). One is located on the power feedthrough plate mounted on > the side of the RF drawer compartment while the other is on the front of the plate "cavity". As many of you have told me, the one on the cavity often > blows at high power on 432 MHz; and to date, everybody who has lost it seems > to have put the cavity aside for parts. >
If I remember correctly I took a 2" piece of solid polyethylene dielectric
(not foam) stripped from a piece of RG-8. Drill small holes through the
center and end. Wind about 20 turns of #20 enameled wire on one end, strip
the insulation of the ends and poke the ends through through the holes in
the PE. Release the plate line from the B+ end of the plate cap and
thread one end through up the center of the dialectric through the middle
hole (tricky) along with the center (plate) end of your choke, solder.
Cover the choke end of the assembly with shrink-tubing and force screw the
assembly into the existing FL1 mounting hole. Size and resolder the B+
line to the plate blocking cap. Bypass the B+ end of the choke with an
appropriate cap (I used 1000pf at 3Kv) and attach the HV line. Use
appropriate care in dimensioning and insulating the assembly with due
consideration of the voltage involved. This unit has been in use for years
without trouble. This is, of course, for 144Mhz. Another possibility is
to install the choke inside the plate line tube (as per K0TLM?) Hope this
helps.
W8KX
The "bypass capacitor" at the top of the cavity is actually a filter assembly. I broke one once and found there were several ferrite beads inside. In this case I replaced it with a conventional plate choke coil with a bypass cap wound on a protruding section of RG-8 polyethylene screwed into the existing threaded hole. (For 144Mhz) W8KX
tonycash@worldnet.att.net wrote: Well I think I got the grid circuit in. Does anybody know the approximate setting for the turns counter on the cavity and the approximate spacing for the plate tuning capacitor when the AM6155 is tuned for 432.100. Thanks in advance for any info. 73 de Tony, KD4K Cumming, GA. Trying to make this thing work for the January Contest. Tony, The turns counter ought to be around 155 and the plate tuning capacitor is out to the max. Which grid circut did you use? I've had a 6155 for several years and have used the 5/8 wave helixcal input and am now using a 1/2" donut inserted in the grid cavity to shorten it. I haven't made up my mind which works best. I had the amp doing 350 watts out and decided to go for more. It hasn't been the same since! Currrently, I've got an arc over problem in the plate blocking capacitor. I'm going to change the insulating material to teflon and try again, as soon as I get some more fuses! Let me know how it works. I'd like to compare notes. Jeff - WB5YWI em25 landman@ok.azalea.net Bands: 6,2,432,1296
From: KO0U
It's doing about 350W at between 40 and 45% efficiency and about 400W at 40%. Doesn't seem to want to get any better. This's about what mine did. Had to do more work on the plate cavity to get the efficiency up a bit, same as I did on mine. Richard's didn't get these mods and did this performance anyway. But then, he got a brand new 4CX400A (for $120) which may have had something to do with the extra output. He said he made a mistake one night and kicked it up to 500+ watts accidentally; it survived and still works fine. I don't recommend putting in a 400A until this tube goes south. But it's a straight drop-in other than having to bend the finger stock slightly because the anode is 1/16" smaller in diameter. Another problem sometimes shows up on 432 amps that are run at too high power, usually 400+ watts, but I've not heard of it occurring in the lower frequency amps as yet. This problem is invisible and can only be located either by knowing that the amp used to put out lots of power on 432 and suddenly does not (or the output power suddenly drops to almost nothing after you exceed some output level such as a couple hundred watts), or by disassembling the cavity completely to inspect things inside. What happens is that a piece of wire that passes through the center tube of the cavity and which provides HV to the tube (the end of the wire also passes through the center of the round aluminum DC plate blocking capacitor plates and is soldered into what looks like a rivet that you can see right in the center of the outside round plate) heats up and can arc over, inside the center tube, which usually, but not always, blows the HV AC fuse (you can tell when this happens because the HV bulb will not come on when you flip the toggle switch). The fix has to both replace the original wire, and allow higher-power output on 432. It isn't complicated but is involved. The original wire is replaced with some ordinary bus wire which has a small coil wound in one end to put more inductance in the wire, which is actually what's called the plate choke. In addition, the wire and choke have to be heavily insulated with teflon tubing to prevent it from arcing inside the center aluminum tube. I hope you don't have this happen, because it's a real bear to fix, taking all of a day and some serious testing to ensure that it's really fixed!
Well, after many hours-days of off and on diddling with the beast I believe I have reached the point of no further gain. With 10W of drive it now puts out 410W into the Bird load, key down CW using the 8930. I do however have a few questions and concerns. I am using the WA5VJB input circuit which is a 3" piece of brass bent into almost a U shape. My concerns are: The coupling cap from the input connector is a 270pf silver mica.....the article says use 200 to 400 pf. Isn't this a bit large for 432 ?? I would think 5-10pf would be more reasonable. I moved the input tap around until I obtained a 1.2:1 input VSWR....a real pain to do. The input tuning and power output vary almost 100W depending on the length of the coax from the rig to the amp. This tells me the amp really needs a input tuning capacitor. If I use the ARR preamp (RF relay switched) in the line from the IC-451A to the input coax I can only obtain 220W out ; I have to put the preamp in the amplifier by-pass line for things to work. Has anyone used an input trimmer, Johanson maybe, with the WA5VJB input? Has anyone actually set one of these amps up with a Network Analyzer? With all the jumper cables and relays in place the output is 390W so maybe I shouldn't complain but I just don't like the idea of selecting cable lengths to do the input tuning. Once I get everything finalized I plan to run IMD tests. The metering circuits leave a lot to be desired so I am not 100% sure if I am still in pure AB1. I plan to outboard a bunch of meters for the tests and also to calibrate the internal " mess". Looking for comments, suggestions, etc. Tnx Carl KM1H
From: "Dr.Gerald N. Johnson P.E. Subject: Re: AM-6155 on 432 MHz Carl, best I recall, I used a 5 or 10 pf trimmer coax connector to grid in the Collins TU-9 with the original grid inductance which has to tune from 220 to 400 in the original arrangement. I probably tuned it for maximum power transfer, neglecting SWR. I had too much drive from the FT-726R so I put 3 db attenuation between the 726 and the final (a length of RG-58). The output tank in the Tu-9 is lossy drug up to 432 with way too much circuit C so the loaded Q is too high. I get only 150 watts out by my measurement and thats after some changes to get the loading under control. I need to make some more changes to try for higher efficiency, but I have other projects in mind. This works and 150 watts is a lot better than the 8 watts the 726 produces. 73, Jerry, K0CQ