Print

[VE7RF]

10 kw CCS on 6M..using the 3CX-6000A7...Part 5.

https://www.youtube.com/watch?v=EFgPEcMdmyQ

Here the very unique method of constructing the filament choke is shown.  Instead of paralleling 2 x identical 40 amp fil chokes..... (made from 8 ga magnet wire on 8" long x .5" type 43 ferrite rods), a new method is used.
The new method uses a string of Fair-rite ...... 'type 43 full flat cable cores' .   These are oval shaped cores, that are  1.53" wide x 1.024" tall x  1.126" long.   Inside dimensions of the oval shaped hole are 1.03"  wide x  .51" tall.
The max size pair of power cables that will fit inside is  1-0  RW 90 power cable.  (170 amp CCS, and 600 vac rated).  I verified this already months ago, and the pair of 1-0 cables just slides in comfortably.
For this 6M amp, double insulated 4 ga, and fine stranded wire was used (600 vac rated).  4 ga wire will easily handle the 79 amps of fil current.


Each bead has a Z of aprx  125 ohms @  50 mhz.   5 x cores are used which =  625 ohms, which is ample, considering the 3CX-6000A7's input Z is 50 ohms.   Typ you want the Z of the fil choke to be quadruple the input Z. (Heath's SB-220 used a 9 uh  fil choke = 197 ohms of XL @ 3.5 mhz).
For lower freq operation, more cores can be used.

The Z of each oval core per Fair-rite is as follows.

2mhz -       25 ohms
4 mhz -      30 ohms
10 mhz -    75 ohms
21 mhz -  100 ohms
30 mhz -  115 ohms
50 mhz -  125 ohms

For operation down to 160m,  a minimum of 8 x cores are used, which equates to  a total length of 9".  For larger tubes like the 3CX-10,000A7 / 15,000A7, those tubes have a lower input Z, aprx 32-35 ohms...and  a minimum of  6 x cores are required.
More is better of course.   Typ the fil choke is oriented across the width  of the chassis, between the input fil connectors at one end of the rear apron...and the tube at the other end. This method works good on typ  19"/ 24" wide rack setups.
With this 6M amplifier, the 5 x beads are flipped 90 degs, and the length to the socket is very short.  Fil V is measured at the socket (via a pair of RF chokes).  The 7.0 vac fil V at the cold end of the fil choke will be rectified into 8 vdc for powering the opto-isolator, used to key the myriad of vac TR  relays.  Fil V will be controlled via a variac, located in lower box #2.

Fair-rite info. https://www.fair-rite.com/.../miscellaneous-suppression.../

[SWMAN]

 Do you really need 10 kw ? I thought we are only allowed 1.5 kw ! Maybe the rules have changed and no one told me.

[W9IQ]

The Z of each oval core per Fair-rite is as follows.

2mhz -       25 ohms
4 mhz -      30 ohms
10 mhz -    75 ohms
21 mhz -  100 ohms
30 mhz -  115 ohms
50 mhz -  125 ohms

Perhaps you already ran the numbers but keep in mind the quoted figures are the impedance magnitude. A good portion of the complex impedance is resistive (a low Q core). This resistance causes core heating. Make certain you are within temperature and saturation limits of the core and that your wire insulation can handle the ΔT.

- Glenn W9IQ

[KT4WO]

 "Do you really need 10 kw ?"

I will admit, that was my first thought also.
(Actually it was-  "But why")  ha.

[KM4AH]

Maybe by the end of the QSO the hamburger will be done.

[VR2AX]

What are the teflon (?) wires for? Seem oversize for filament voltage measurement, and I did lose the thread at that point (do they connect to the 'banana' connectors?)

[KD6VXI]

Credit where credit is due.  This is W8JI design, yes?

I still maintain that running glastic isolation on the filament transformer and the choke in the input side of the xformer is easier.

On a tube like this you're looking at a fils choke that would need to pass 6A on 120v.  If a 240 primary is used, that's 3 amps (allowing for losses in both cases).

This is how I have built 10s, 15s, 20s, single and multiple tube amps.

--Shane
WP2ASS / ex KD6VXI

[VE7RF]

What are the teflon (?) wires for? Seem oversize for filament voltage measurement, and I did lose the thread at that point (do they connect to the 'banana' connectors?)

The Teflon wires at the cold end of the fil choke will be used to feed the input of a FWB rectifier assy.  The rectified vdc will be used to power the opto-isolator.  The opto-isolator will be used to key the 4 x vac relays.

Actual fil V  will be measured at the socket....via RF choke in each leg....then wired over to a pair of test jacks..... for DVM true rms measurement.   

[VE7RF]

Credit where credit is due.  This is W8JI design, yes?

I still maintain that running glastic isolation on the filament transformer and the choke in the input side of the xformer is easier.

On a tube like this you're looking at a fils choke that would need to pass 6A on 120v.  If a 240 primary is used, that's 3 amps (allowing for losses in both cases).

This is how I have built 10s, 15s, 20s, single and multiple tube amps.

--Shane
WP2ASS / ex KD6VXI

Running a pair of power cables through ferrite beads is real old news...but the damned things are not vey thick, but some are.  JI discovered some oval types....but of course wouldn't allude to where he found em.  After 4 hrs I found em on fair rites site, buried under misc suppression cores.

Sure, the  fil xfmr can be isolated from chassis   with  glastics ..and a small choke in the 120/240 primary.   BUT that only works if the fil xfmr in on the RF deck chassis..... and not a mile away on a lower chassis.   Also a choke is required in the CT of the secondary of the  fil xfmr.... to keep drive RF out of CT / bias / metering etc.

I have not measured the uh  between  CT and outer ends on a fil xfmr.  It might be enough when > 25 mhz.   

[KD6VXI]

Credit where credit is due.  This is W8JI design, yes?

I still maintain that running glastic isolation on the filament transformer and the choke in the input side of the xformer is easier.

On a tube like this you're looking at a fils choke that would need to pass 6A on 120v.  If a 240 primary is used, that's 3 amps (allowing for losses in both cases).

This is how I have built 10s, 15s, 20s, single and multiple tube amps.

--Shane
WP2ASS / ex KD6VXI

Running a pair of power cables through ferrite beads is real old news...but the damned things are not vey thick, but some are.  JI discovered some oval types....but of course wouldn't allude to where he found em.  After 4 hrs I found em on fair rites site, buried under misc suppression cores.

Sure, the  fil xfmr can be isolated from chassis   with  glastics ..and a small choke in the 120/240 primary.   BUT that only works if the fil xfmr in on the RF deck chassis..... and not a mile away on a lower chassis.   Also a choke is required in the CT of the secondary of the  fil xfmr.... to keep drive RF out of CT / bias / metering etc.

I have not measured the uh  between  CT and outer ends on a fil xfmr.  It might be enough when > 25 mhz.

It's a point that's (at least should be) low with rf.  I've built 20 meter amps like this as well.  They didn't require anything special.

Works good.  And even if you need a center tap choke, it's at a low current point.  You only need tk build it for cathode (grid + plate) current. 

Then you have the cathode tubes.  Even easier.



I have some ferrite rings here I could wind with #4 and get enough turns even for 160.  They work a treat for HF chokes.  Maybe 4 in he's across inner diameter.  I use one on my vertical feedline at the end of the radial field.  Feedline on the ground for another 50 feet then maybe 5 feet into the Lp100a.

I have another one, maybe when hurricane season ends I'll try it where the coass plugs into the coupler.....?


--Shane
WP2ASS / ex KD6VXI

--

[KD6VXI]

"Maybe 4 in he's across"

Should have been

"Maybe 4 inches across"

--Shane
WP2ASS / ex KD6VXI

[VE7RF]

Credit where credit is due.  This is W8JI design, yes?

I still maintain that running glastic isolation on the filament transformer and the choke in the input side of the xformer is easier.

On a tube like this you're looking at a fils choke that would need to pass 6A on 120v.  If a 240 primary is used, that's 3 amps (allowing for losses in both cases).

This is how I have built 10s, 15s, 20s, single and multiple tube amps.

--Shane
WP2ASS / ex KD6VXI

Running a pair of power cables through ferrite beads is real old news...but the damned things are not vey thick, but some are.  JI discovered some oval types....but of course wouldn't allude to where he found em.  After 4 hrs I found em on fair rites site, buried under misc suppression cores.

Sure, the  fil xfmr can be isolated from chassis   with  glastics ..and a small choke in the 120/240 primary.   BUT that only works if the fil xfmr in on the RF deck chassis..... and not a mile away on a lower chassis.   Also a choke is required in the CT of the secondary of the  fil xfmr.... to keep drive RF out of CT / bias / metering etc.

I have not measured the uh  between  CT and outer ends on a fil xfmr.  It might be enough when > 25 mhz.

It's a point that's (at least should be) low with rf.  I've built 20 meter amps like this as well.  They didn't require anything special.

Works good.  And even if you need a center tap choke, it's at a low current point.  You only need tk build it for cathode (grid + plate) current. 

Then you have the cathode tubes.  Even easier.



I have some ferrite rings here I could wind with #4 and get enough turns even for 160.  They work a treat for HF chokes.  Maybe 4 in he's across inner diameter.  I use one on my vertical feedline at the end of the radial field.  Feedline on the ground for another 50 feet then maybe 5 feet into the Lp100a.

I have another one, maybe when hurricane season ends I'll try it where the coass plugs into the coupler.....?


--Shane
WP2ASS / ex KD6VXI

--

Shane, you gotta be extremely careful when inserting a small bifilar in the pri of the fil xfmr...esp on 6m....or lower bands.
It ONLY works if the fil xfmr is located on the RF deck, NOT on a shelf below the RF deck, nor in box#2, below the RF deck....otherwise, the interconnecting leads would be hot with radiating drive RF.   The interconnecting leads would also have add a ton of stray C to the metal cabinets, etc, adding to the tubes input C.

The biggest reason the fil xfmr is not installed on the RF deck is WEIGHT.   A 80 amp fil xfmr is no light weight, nor a 100/160 amp fil xfmr.  And none of em will fit below the chassis. IF I added the weight of the fil xfmr to my 3x3 RF deck, I would not be able to get it in /out of the rack by myself.

And if you put them on the top of the chassis, they have to be well shielded.   If not shielded, the 10-30 kw of RF  flying around the upper PA compartment can easily couple into the (hot with drive RF) fil xfmr.   Then you end up with an osc, with output power being coupled back into the input. And when you shield the fil xfmr, depending on spacing of shield to fil xfmr, you add C between them.... which adds to the tube's input C.  The top shield on the fil xfmr also has to be ventilated.

In most cases, there never is  any room on the top side of the chassis to locate the massive fil xfmr.

To really do it correctly, all 3 terminals on the secondary...and both input terminals of the fil xfmr, need .01uf bypass caps...to the metal frame / shell / casing of the fil xfmr.

I checked my old notes.  Back in the 70's, 'low capacitance' fil xfmrs were used on GG triodes with 50/80/100/160 amp fil requirements.   Essentially it was a single C core, with  primary wound on left side..and secondary wound on right side, then the top installed, and steel bands installed, and cinched tight.  (Imagine taking a Dahl plate xfmr, and tossing both pri + sec windings, and also tossing one of the 2 x C cores, leaving just one C core left).   Even on that setup, a choke was installed in the CT of the secondary, to keep drive RF out of the bias/cutoff bias, and metering / and the B- of the HV supply.

[VE7RF]

The Z of each oval core per Fair-rite is as follows.

2mhz -       25 ohms
4 mhz -      30 ohms
10 mhz -    75 ohms
21 mhz -  100 ohms
30 mhz -  115 ohms
50 mhz -  125 ohms

Perhaps you already ran the numbers but keep in mind the quoted figures are the impedance magnitude. A good portion of the complex impedance is resistive (a low Q core). This resistance causes core heating. Make certain you are within temperature and saturation limits of the core and that your wire insulation can handle the ΔT.

- Glenn W9IQ

Yes, I factored that RS is lightly less than the Z across their entire graph.   The completed choke measures 8.9 uh.
8.9 uh =  2803 ohms of XL  @  50.125 mhz.   Anything > 200 ohms of XL is ample.  (SB-220 used a 9 uh fil choke, with 197 ohms of XL  @  3.5 mhz).  Rule of thumb is you want the XL to be at least quadruple the input Z.

The fil choke is still bifilar, even with 2 x cables.  The 60 hz, 80 amps flows up one cable, through the cathode...then back down the  2nd cable.  The 60 hz magnetic fields are always opposing....and simply cancel each other out.  Ditto when the AC sinewave flips polarity.  Done this way so the ferrite cores don't heat up at all.  Cables are rated for 105 deg C + 600 vac  (3 kv test).

500 watts of drive RF =  160 Vrms ....(= 226 V peak).   160 Vrms / 2803 ohms of XL = .057 amp  at the cold end of the fil choke assy...which is bypassed to the chassis via .01 uf bypass caps.   Drive RF goes to the cathode...and a miniscule amount  flows through the pair of cables.  But we still have DC cathode current flowing through the fil choke assy, like 1.8 amps flowing up each cable...when driven.

John,  K5PRO asked about the stray C between the choke assy..and chassis....since the dielectric constant of type 43 is greater than air, and in this case, the heatshrink is butted up to the chassis sidewall.  It measures exactly 50 pf.  This 50 pf is in addition to the 42 pf of input C on the tube itself. (tube input C is measured between the grid and cathode).

That's now 42+50 = 92 pf.  This 92 pf has to be subtracted from the C2 variable cap on the PI tuned input.   IE: it simply gets absorbed by the C2 cap.  (C2 cap is closest to the cathode).  Good thing we measured it....so it can be factored in. A different C2 cap is going to be used, with much lower minimum C.

Raw tuned input values are as follows.   ( 50 .125 mhz..and a 50 ohm input Z)

For a total network Q of 4  (input Q = 2.   output Q = 2)

C1 =   127 pf
L =   .13 uh
C2 = 127 pf

So the 127 pf on the C2 air variable will have to be reduced by 92 pf.   So the C2 cap is tweaked for just  127-92 = 35 pf.

The tuned input is being built as I write this, using 2 x air variables +  1/8" OD   silver plated tubing coil.  Only the back end of each cap is bonded to the chassis....closest to the cathode.  Front end of each cap is insulated from the chassis...and also the shaft. Either 1/4" delrin rod / fiberglass will be used through the front panel.... or a shaft isolator will be used. 

[KD6VXI]

The 3 x 15,000 deck (4CX15000J) has all three fils xformers in the rf deck.

It's not light, but a west coast winder will make them any size they can, if you spec it.

Years ago they where making 7.5@55A xformers.  They fit in a medium size flat rate box.  I'm sure USPS didn't like them, but we shipped tons of them out to people.

--Shane
WP2ASS / ex KD6VXI

[VE3WGO]

Peter Dahl transformers are now made by Hammond, and the 7.5V@85A for YU-148/3CX6000A7, partnum PWDF13050 weighs 21# . I guess they count on voltage drop in the choke.

73, Ed