开云体育

开云体育

开云体育

On Mon, Jan 15, 2024 at 12:26 AM, W7WRX wrote:

<I got all the parts in and am building the board.

<I installed all the relays, the 10 meter A and B trimmers and the wound a toroid. 7 turns of 20 AWG

<With a dummy load on output of 50 ohms and the analyzer on the input:

<with 10 meter relay closed, I can tune the trimmers to get a dip of 1.0 at 56 ohms 28.500.? ?I am using LESS cap than the Chart says. It says 100PF, 7 turns and 100PF.? I am at 25 to 35 PF on each in and out cap.

<Question

<What is the expected behavior if I roll the Analyzer down to 20 meters?? ?

<I do NOT SEE A BAND PASS for the 10 meter position. I only see matching and a HIGH PASS.? If I try to go past 40MHZ the SWR shoots up high.

<If I go LOWER in frequency the impedance will move but its fairly flat

<Is this expected behavior????

<C

## I design and run the PI tuned inputs through software every other week...then verify on the bench.? For a overall network Q? of '4'? (input Q = 2....and output Q =2)? ?on 10m,(28.5 mhz) when matching 50 ohms from xcvr to 50 ohm tube input Z....?it should be 223.4 pf? - .22 uh? - 223.4 pf.?

## For an overall network Q of? ? '1.8'? (input Q =.9......and output Q = .9) on 28.5 mhz, it's then 100 pf - .28 uh - 100 pf.?

?

##? if ur caps are only 25-35 pf, then ur Q is through the floor...and the coil has way too much uh.?

?

## either remove a turn or maybe 2, and? / or spread the turns? a bit.? Tweak the coil till both caps are in the 100-140 pf range.? Ok, at that point, you will see the 1:1 swr @ 28.5 mhz..... which should remain flat between? 28.0? and 29,0...then rise when u go below 28 mhz...and above 29 mhz.

?

##? with the Q through the floor, it all goes to hell...and you won't get a? proper bandpass response.?

?

## Have u tested al the other bands yet ?? ? IE: installed caps and toroid's for the other bands...and tested other bands.?

?

##? Note on the above,? only when matching 50 ohms to 50 ohms? (or any 2 x equal Z's)? is the C1 + C2 caps both the same value.? On software, we only tweak the overall network Q.? ?The C1 input Q? ?+ the C2 output Q = overall network Q.?

##? For a PI network for the kw output side, (High tube Z? down to 50 ohms), typ an Overall network Q of....'12' is used....which consist of an input Q of 10...and an output Q of 2.??

##? In the old ARRL books (and also Eimac's? 'care + feeding'), they only used the INPUT Q...which is why they said? to use a Q of? '10'.? ?They did the same thing with PI tuned inputs.? When they said to use a Q of..'2'? ?what they really meant was to use an overall network Q of 4.?

?

##? For tuned inputs that are tweaked once ..in the middle of the band,? you want low swr? across the entire band.? ?500 khz on 80m is like 1mhz wide on 40m..and 2 mhz wide on 20m..and 4 mhz wide on 10m.? This is where you can do some creative tweaking.? Each band is different....with 160m only 200 khz? wide, 80m = 500 khz, 40m = 300 khz, 30m = 50 khz,? 20m = 350 khz, 17m = 100 khz, 15m = 450 khz, 12m = 100 khz, 10m = 1700 khz.? ( 6m = 4000 khz).
Then factor in if the tuned input has to cover? 10+12 m...... or 15 + 17m.? ?The best u can do there is to tweak it so it's flat swr at the midpoint between 10+12m.... and ditto with 15 + 17m.?

?

## here is the other thing u gotta worry about.? The tube's input C? is directly in parallel with the C2 of the tuned input. Say a 3CX-3000A7.... with it's 38 pf input C (grid to cathode C).? Now you have to subtract 38 pf from the calculated C2 values of ALL the various tuned inputs for each band.?

##? here's the other thing to deal with.... the XL of the bifilar fil choke (since it's RF grnded at it's cold end...and hot end of fil choke goes to the cathode) is also directly in parallel with the tube's input cathode to grid C.?

##? On a SB-220, they used a 9 uh (per winding) fil choke.? ?XL @ 3.75 mhz =? 212 ohms.? ?You now require 212 ohms of XC to cancel out the XL.? ? 200 pf @ 3.750 mhz = 212 ohms of XC.? On an SB-220, the C2 cap of the 75m tuned input has to be increased by 200 pf? vs it's calculated value.? But since the tube's input C on a pair of 3-500Z's is 16.6 pf,? it's then 200-16.6 pf =? 183.4 pf.

## The fil choke primarily affects the lower bands. On the upper bands it's less of an issue.? ? 9 uh is 1611 ohms of XL @ 28.5 mhz.? ? ? 3.465 pf = 1611 ohms of XC.? ? ?That amounts to dick on 28.5 mhz...and the pair of 3-500Z's still has 16.6 pf...so now it's 16.6 - 3.465 pf = 13.135 pf.? 13.135 pf has to be subtracted from the calculated C value of the tuned input.

?

##? If you are doing say a PI network for a 3CX-3000A7,? wire the relay switched tuned input? into the cathode.....and temp wire a 50 ohm resistor between one cathode lug....and chassis....and also plug the tube into the socket.? With no fil or B+, all voltages off....and TR relays activated, put ur analyzer on the input...and tweak both arco's for each band for 1:1 swr.? ?50 ohm resistor removed.? Apply B+ and fil V.... and test the amp.? ? It will be sorta close, and perhaps some minor C1 + C2 tweaking required.? That depends on drive level, bias used,? loaded B+ used, amp tuned to max PO.... and perhaps loading increased till PO drops 1-2%, and cable length between xcvr and input to amp.?

##? if the relay switched, tuned input board is below the chassis, near the cathode, it may or may not be possible to tweak all the arco caps.? ?The board would have to be mounted to rear apron, with holes to access the caps..or to chassis..pita.

If amp is fan cooled, you might be able to remove bottom lid.? If a chimney is used, removal of bottom lid means? zero airflow.??

?

##? This is why I went to the pair of broadcast caps + tapped coil, manually tuned, tuned input setup on the hb amps.?

?

Jim? ?VE7RF

I got all the parts in and am building the board.

I installed all the relays, the 10 meter A and B trimmers and the wound a toroid. 7 turns of 20 AWG

With a dummy load on output of 50 ohms and the analyzer on the input:

with 10 meter relay closed, I can tune the trimmers to get a dip of 1.0 at 56 ohms 28.500.? ?I am using LESS cap than the Chart says. It says 100PF, 7 turns and 100PF.? I am at 25 to 35 PF on each in and out cap.

QuestionL

What is the expected behavior if I roll the Analyzer down to 20 meters?? ?

I do NOT SEE A BAND PASS for the 10 meter position. I only see matching and a HIGH PASS.? If I try to go past 40MHZ the SWR shoots up high.

If I go LOWER in frequency the impedance will move but its fairly flat

Is this expected bahavior????

C

The resonating oil cap value is easily measured with any digital lcr meter.

The real problem is measuring the choke.? I went through this with dahl / hammond a couple of years ago. A standard digital LCR meter won't read the inductance correctly, but on the low side.

My B+ K 875A / and also my 875B? will read from .1 uh? up to 200H.?

When clipped onto my dahl? 4 H? choke ( 127 lb hypersil, rated at 3 amps CCS)... it reads like 3.26 H or a bit more.

The only way to measure the inductance on those chokes accurately per dahl / Hammond was to use rated? DC current flowing through the choke, while conducting the test.? They sent pix of the 2 units that are used for that job.? One is 20 amp rated, the other is 100 amp rated.? ?I will have to dig out the emails, but it used a bridge circuit and other items...along with the 20 amp dc supply, to arrive at the actual inductance.?

The problem with the resonant choke setup is.... you have TWO currents flowing through the choke.? The normal rated? DC current to the amp...and the 2nd current is the circulating current that flows through the resonating cap and choke.?

What was tested with resonated chokes? was 1H, 2H, 4H, 8H.? The smaller the choke, the bigger the resonating cap has to be.....and the more the circulating current.? On a 4H choke, the resonating current was as much as the actual DC output current.? Ok, the light bulb went on.? For resonant choke use, the choke has to be rated for a helluva lot more than? your dc plate current.?

You can use the LC calculator online to see what value cap required to resonate any choke to any freq.? ?Typ it's aprx .44 uf to resonate a 4H choke at 120 hz.? On the Collins? 30S1, the B+ supply is resonant at 120 hz.? ?On the resosnat choke setup for the screen supply, they resonated it at 165 hz.?

?

John Lyles, K5PRO..back when he designed FM broadcast tube amps for BE ( broadcast engineering)..... this is on there small 5 kw output FM? PA..... John used a pair of .9 uf? @ 5 kv oil caps wired in series to make up the resonating cap.

The resulting .45 uf cap assembly was then sent to the choke manufacturer....and they designed the choke such that it resonated exactly at 120 hz....when rated normal plate current drawn.? This is on the 5 kw,? ?240 vac /? 60hz / single? phase B+ supply used in that PA.? ? So it's just a resonant choke...followed by a pair of paralleled oil caps, typ standard setup.?

Interestingly enough,? he said there is a spec for FM PA's...called an ...'AM noise spec'...and to pass that spec, extra filtering in the B+ supply was required.... to kill ALL the harmonics of 120 hz ripple freq.? ?The resonant choke only kills the 120 hz, and the main pair of oil caps are not big enough in value to kill all the harmonics of 120 hz (240-360-480 etc).? ? To pull that off, a 2nd choke was used( not resonated), followed by some more oil caps.? The 2nd choke + oil caps is a simple LC filter....which has 12 db attenuation? per octave.? Depending on the value of the 2nd choke and caps... will determine the staring point or corner freq, where attenuation 1st starts.?

?

I got a hold of the 4 H dahl choke from a ham across town.? It uses the same pair of hypersil cores? as the dahl 10 kva plate xfmr.? several of em were used at the now defunct satellite uplink? station? 50 miles north of me.?

I'm using it in an experimental? ?C-L-C? config..... with a spst vac relay across the choke terminals? ( and also an adjustable spark gap across the choke).? Choke is in series with the B+, and is insulated from grnd with 4 x huge glastics.? ?So far, so good, with 120 hz ripple down to just 20 mv? ( -121 db) on a 7 kv loaded B+ supply...loaded to 2.5 amps CCS.

?

Software only shows P-P ripple at 120 hz. BUT the C-L-C continues attenuating above 120 hz.....at the rate of 12 db per octave. IF I have done the calcs correctly, it's 1st modelled as plane C filter, (just C1) the 120m hz ripple is down to -54db. Then the LC is added to the mix. Then it's down to -119 db....(an additional 65 db) It works out the corner freq is just 2.9 hz......then attenuates at the rate of 12 db per octave above that.

If u only have X amount of money to spend, spend it wisely.? Money is better spent on a bigger hypersil plate xfmr...with a plane jane C filter...like a string of lytics... vs a plate xfmr + resonant choke.?

The problem with the FWB + resonant choke setup is.... the sec HV has to be a lot bigger.? Since the B+ is only 90% of the ACV...... vs? 41% more on a FWB? + high C filter.?

On Wed, Jan 10, 2024 at 06:29 PM, Peter Voelpel wrote:

RG393 will do it easily, it is specified 25kW at 10MHz, 9.5kW at 50MHz.

?

73

Peter, DJ7WW

I have found at least 1/2 oz specs for RG-393..... and they are all over the map.? ?393 only has a? small 12 ga center conductor.....(.0808").? The issue is....whatever the puny center conductor is connected too, is gonna have one helluva lot of heat on it. IE: vac relay contacts etc.? Relay contacts are sitting at 150-200C.... just from the heat from the coax.? Try running 9.5 kw cxr on 50 mhz with 393.......... been there, done that.? It's not big enough.

?

SFT-600 is the Teflon version of LMR-600UF.? ?SFT-600 has a .159" center conductor? (6 ga).? The length? required from the vac load cap to the vac relay assy is very short in most cases.? It's readily available, cheap on ebay etc...and doesn't run stupid hot. Connectors are readily available, like 7-16 din etc.? ?SFT-600? works superb for stuff like rotor loops etc.?

The biggest they make in teflon? coax is SFT-900.

The fellow's website is interesting.? 100 kw pep output on ssb? from a 4x20,000C...with a 30 kva plate xfmr...dream on.?

?

On that 4x10 on his site, that uses air variable tune and load caps, he uses a string of 330 uf lytics on the single phase 230 vac supply.? Those 330 uf caps only have a 1700 ma ccs ripple current rating.? Typ ripple current on the B+ supply is 2.5? X the dc plate current. You can see that in software...and also easily measured.? Those caps wont handle it, not big enough.? ?The larger value lytics always have progressively higher ripple current ratings.

?

Something else to note, the actual VA power drawn from the mains is quite a bit higher than the DC input.. like a LOT higher.? But u need a true rms ammeter to measure it. If a non true rms meter is used, it will read? 20-40% on the low side when a C filter is used.?

On Wed, Jan 10, 2024 at 06:34 PM, W7WRX wrote:

If your 8k Ultra sags down to 4800 volts you have something wrong.? Possible weak AC line, or failed Oil caps.? I used a Fluke HV probe and mine sits at 5900 volts.? panel meter shows 6000.? ?At rated 5kw output it sags to 5500.?

?

The 8K has a very small power supply. Its only a 1 AMP CCA transformer with 2-amp surge.? It was designed for 5kw pep on SSB and no more than 1 amp plate current in Data modes. That is RIGHT IN the manual.? ?Its rated at 8k INPUT to the power supply.? NOT output.?

?

This is a good example of my exact point.? The amp will make 7000 plus pep with video all over the net showing this.? It does this on SSB with a 2 amp transformer That I can pick up and walk away with.

?

Which of the 7 versions of the 8K ultra are we talking abt ?? ?The 8k manual I have sez...." 5400 dc no load.? Sags to 4800 vdc with a 1.25 amp CCS? cxr".? On the manual I have, they used? a 6000 vac xfmr...and a resonant chole input. No load B+ is .9? X 6000 = 5400 vdc.? The manual I have sez 3.5 kw pep out...on the higher B+ position..... and 2.5 kw out on the lower B+ position ( data modes).

One thing u gotta be careful with on the 8k..... it uses 5 x 20K? @ 100 watt? wire wounds in series for bleeders.? Each wire wound dissipates 60 watts? (300w in total).? If any one of em goes? open, all hell breaks loose.? Then it turns into a capacitor filter, and B+ is now 6000? X 1.414 =? 8484 vdc.? Something to be aware of.? In normal operation, 300 watts of bleeder is used to make the choke look like a choke.?

Are u saying ur 8k sags to 5500 vdc.... while talking on ssb ???

Jim,

Thanks for the explanation.? I understand the principle of what you are describing, eg the energy in the tank circuit causes the voltage to rise on the anode on the positive cycle, before dropping down on the negative cycle of the RF.? With these Russian tubes having such variability as you described, i guess its safer staying with a lower plate voltage and keeping the swing lower.? I don't have the equipment to hi pot test the tubes, but was planning on following the instructions on the internet around "reactivation" of the tube.? Detail copied in below

To avoid flashover in a new or a long time unused tube, it is prudent to prepare ("reactivate") it for his task. There are various opinions and solutions how to do it, but with a GU-43 it can be relatively simple. A DC of 10 to 15 V is sufficient for the tube to draw 900 mA as the grids are connected to the anode. Heat the tube 30 minutes with a filament voltage of 12.6 V. Then supply a "high voltage" of about 15 V and set the voltage to a current of 900 mA. If possible use a current limiter, because during reactivate the current can increase suddenly so that continue monitoring could be necessary to maintain 900 mA. Usually I will not reactivate longer than an hour or so. Soon you will find out at which voltage a good tube drawn 900 mA (or 800 mA/12-12.5 V) so you have an indication if another tube is better or not.

Alex,

Thank you. That makes sense.? His ratings seem way to high to me.? Thanks for the confirmation.

开云体育

开云体育

开云体育

RG393 will do it easily, it is specified 25kW at 10MHz, 9.5kW at 50MHz.

?

73

Peter, DJ7WW

?

?

?

-----Original-Nachricht-----

Betreff: Re: [ham-amplifiers] Building my first amplifier. Where to start?

?

On Tue, Jan 9, 2024 at 11:15 PM, W7WRX wrote:

Its not hurting the tube.? You need to remember that this is Single side band with a duty cycle of 15 to 20 percent.? Even with heavy processor is not past 25%.??

?

What will the amp be used for???

?

This is a very common mistake made all over amplifier groups.?

?

Another one is Blower CFM and pressure.? All the time I see these guys saying you need X CFM or you need 300CFM!!!? That blower is to small....? I got advice from a guy that suggested I get a blower TWICE the size of the tube data sheet, Just for good measure.? LOL

?

?But those ratings are for 24/7 full power at rated Plate dissipation for 8000 hours.? A ham using SSB will never need anywhere near those ratings.? Even if you ran A digital mode you still would never need broad cast ratings.

?

Another one is transformer current.? If the amp is for Ham use it will not need the full rating of the tube spec sheet in transformer.? Ever.?

?

?Its comical people buying 175lb transformers only to use the amp on side band and never even pull HALF the current its capable of.?

?

It should be built for its intended?purpose.. If it is meant to run Data modes at high power with long operation of a contest, I get it... over build the thing. But most of the time, the amp is just going to be used on SSB with Data or CW less frequently.

?

?

C

The bigger plate xfmr typ results in less B+ sag on ssb / cw.? The 8K ultra sags from 5400 vdc, down to 4800 vdc.? Mine sags just 50 vdc.? ?5500 down to 5450 vdc.? The ebm papst blowers I got a while back (EC type) have variable speed on em....via an external 10 k pot.? I can drop the rpm in 1/2.? You are correct though, on ssb, less air required....and on extended RX, even less air required.?

?

On the dc input side, my average plate current is exactly 1/2 of key down cxr.? This is with ssb and a lot of processing.? I call that a 50% duty cycle...on the input side.? On the output side, the average PO is only 29% of the pep out.?

?

I have yet to experiment with data /? FT-8 modes.? FT-8 destroys typ amplifiers.? I factored that in, and designed it, so slightly reduced ratings are used on data modes.?

On Tue, Jan 9, 2024 at 07:36 PM, W7WRX wrote:

This guy runs his at 3450volts with 3kw out with 60 in.

?

That fellow is well know for his total bs ratings. On that GU43B amp, he uses a pulsed 1.5 khz tone fed into the mic jack, with xcvr on SSB mode.? ?1 on, 4 off, pulsed at 30hz.? ( 20% duty cycle)? ?This is on the DC input side. On all my amps, on SSB mode, with processing, the average plate current is exactly 1/2 of the key down value. Meanwhile on the output side, the average PO is only 29% of the pep value.?

?

Note on all his other amps, esp the bigger ones, he uses a 1 on and 9 off (10% duty cycle), again using a pulse generator at 1.5 khz/30 hz, fed into the mic jack of the xcvr...on SSB mode.?

?

I borrowed the HP pulse generator from work, and also dragged home the Tek 4 x channel scope.? It's amazing how low a duty cycle you can achieve.? With real low duty cycles on the input side, the plate current is not much higher than normal idle current.? Average PO is through the floor, yet pep out on the scope is sky high.?

?

Look at his 4CX-20,000C.? He rates it at 100 kw pep out on ssb.? ?Again using the 1 on, 9 off, 10% duty cycle.? ?There is no way in hell that? amp would put out 100 kw pep on ssb.? His plate choke setup is fubar.? RG-393 coax from load cap to TR relay, wtf ?? ?I don't think so.

On Tue, Jan 9, 2024 at 11:03 PM, desmond auld GI0UTE wrote:

Thanks for the link, im reading that as 3450v UNDER LOAD!? So unloaded he could be at near 4KV dc on a 3 KV rated tube.? Is "over volting" tubes this much typical?? Any rules of thumb to consider here i guess is my question,? Also 3KW PEP from a tube with a 1Kw dissipation must be burning the candle at both ends.? Interesting data point!? thanks,

Listen carefully. ONLY when the tube is driven full bore, the RMS? RF voltage is aprx .6? to .63? X loaded B+ voltage.? Then add 41% for the peak? RF voltage.? ?That peak RF voltage (only when driven full bore)? is in addition to the? loaded? B+ voltage. Do the maths, and the end result is the total voltage on the anode is a bit less than double the loaded B+ value.? This total value is what the tube must stand off.

IE: say 3450 vdc loaded. RMS RF voltage is then .62? X 3450 vdc = 2139 V rms.? Peak RF voltage is 2139 X 1.414 = 3024 V peak.? ? Total voltage on the anode (only when driven) = 3450V + 3024 V peak = 6474 V peak.? ?The tube must be able to stand off at least? 6474 V.?

?

This is why you always, always hi-pot test the tubes before using them.? ?An Eimac 3-500Z? will easily hi-pot test to 16 kv.? An Eimac 3CX-3000A7 will hi-pot test to 20 kv.? An Eimac 3CX-6000A7 will hi-pot test to 25 kv.? The 3x10 and 3x15 will hi-pot test to 30 kv.?

Russian tubes are notorious for hi-pot testing all over the map, so beware.? ?I have seen GS35B's that arc at 3.5 kv....and GS35B's that run just fine at 4.5 kv.?

In most cases, we could care less what the UN loaded B+ is on RX.? What counts is what is the loaded B+.? ? I can easily apply 15 kv on an Eimac 3-500Z..and not draw any leakage current between the anode and the control grid.(grid bonded to chassis).??

开云体育