The AM Forum
March 28, 2024, 01:52:21 PM *
Welcome, Guest. Please login or register.

Login with username, password and session length
 
   Home   Help Calendar Links Staff List Gallery Login Register  
Pages: [1]   Go Down
  Print  
Author Topic: Where did the power go? RE: GS-35B pair amp  (Read 17724 times)
0 Members and 1 Guest are viewing this topic.
W1IA
Member

Offline Offline

Posts: 778



« on: June 13, 2011, 06:27:47 PM »

Ok...after fighting with the input circuit on 10 meters Mark (KA2QFX) and I managed to get the input circuit to work on 10. We had to resort to a T-match configuration to deal with the extremely low input impedance of the pair of GS-35Bs.
The efficiency is dog squezings; With 3600 volts on the plate and 800 mils of  current (2880 watts) were getting aprox. 250-275 watts of carrier...about 1/3 of the expected output. The tank is a plug in design and seems to be correctly resonated per our spread sheet calculation. So where did my power go?

Brent W1IA
Logged

Run What Ya Brung!
WQ9E
Member

Offline Offline

Posts: 3282



« Reply #1 on: June 13, 2011, 06:56:11 PM »

Brent,

That sounds like a great shack heater-if it were winter.  My gues is excessive loading in conjunction with insufficient drive is killing your efficiency.
Logged

Rodger WQ9E
W1IA
Member

Offline Offline

Posts: 778



« Reply #2 on: June 13, 2011, 07:14:09 PM »

Brent,

That sounds like a great shack heater-if it were winter.  My gues is excessive loading in conjunction with insufficient drive is killing your efficiency.
It seems easy to drive...only 17 watts to excite the tubes to 800 mils...I will look at the loading
Logged

Run What Ya Brung!
Opcom
Patrick J. / KD5OEI
Contributing
Member
*
Offline Offline

Posts: 8308



WWW
« Reply #3 on: June 13, 2011, 08:57:12 PM »

The same kind of thing happened with my class-C 4-1000 when I first got it put back together and running. I had the LC ratio wrong for the voltage and current I was running. If this is a new amplifier, could you have the same issue or be mis-tuned to a harmonic?
Logged

Radio Candelstein - Flagship Station of the NRK Radio Network.
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #4 on: June 13, 2011, 09:23:17 PM »

Yup, LC ratio is off. Might be too much tube plate C. to get around it, put a little "L" between the plate coupling cap and plate tune cap.  Check the value of the loading cap. If it isn't where you think it should be I.E. lower than expected then there is an issue.  Another trick check the efficiency at 1 KW input then check it at 2 kw input. If the efficiency is higher at 1 KW then you have too much L. If it is better at two kW then it wants to run higher power or add some L. A trick I use is a bolt and 2 large washers between two turns of the coil to short them. another clue.
A good first step is the thing didn't take off.
Also another problem could be too much L in the parasitic choke. I was burned by that one and I thhink I pointed out that it looked like you had a bit much.
Logged
W1VD
Contributing
Member
*
Offline Offline

Posts: 401



« Reply #5 on: June 13, 2011, 10:18:37 PM »

Possible plate choke issue?

I'm with Frank on the parasitic suppressors ... looks like a lot of L. For the pair 8873 amplifier I used 1/2" copper strap formed into a hairpin with resistors located near the open end. See attachment. This improved efficiency on 10 meters vs. conventional multi turn type. Use minimum L consistent with tame amplifier behavior.     



* parasiticsuppressor.jpg (2.35 KB, 193x149 - viewed 624 times.)
Logged

'Tnx Fer the Dope OM'.
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #6 on: June 13, 2011, 10:20:11 PM »

Brent,

As Jay and Frank suggested, try smaller parasitic chokes first.

Also:  Do you get a sharp power peak when you tune C1 or is it a broad peak compared to 15M?  Is C1 all the way out?

I agree with Frank that you may have too much L or even stray capacitance from the tube's internal plate to ground or circuitry. After all, two tubes has a lot of infrastructure.

The idea of using some L at the pi input will lesson the C1 requirement too.

Heck, I even had too much stray C on 20M with my big amp. I didn't even bother putting it on 10M. Though, the input L trick may be just what I need.



T

Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
W1IA
Member

Offline Offline

Posts: 778



« Reply #7 on: June 14, 2011, 12:18:32 AM »

Thanks for the tips guys...I will run some tests today and repost with the results. As for the parasitic suppresors; they were changed to a single turn that the pictures on the web don't show. Tom, the dip is sharper than other bands suggesting I'm too litely loaded on C2. Thanks again everyone

Logged

Run What Ya Brung!
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #8 on: June 14, 2011, 09:55:53 AM »

I run 1 turn of 1/2 inch strap about 3/4 inch ID. Sharp dip watch the RF choke for heat damage. Sharp dip means high Q somewhere.
Logged
AB2EZ
Member

Offline Offline

Posts: 1722


"Season's Greetings" looks okay to me...


« Reply #9 on: June 14, 2011, 10:39:21 AM »

Brent

With the amplifier set up for 10 meters, with the input circuit in place... and keyed on... what is the resting plate current with 0 volts rf input? (i.e. remove the r.f.  cable from the output of the exciter, and replace it with a 50 Ohm dummy load at the exciter end of the cable)

Stu

Logged

Stewart ("Stu") Personick. Pictured: (from The New Yorker) "Season's Greetings" looks OK to me. Let's run it by the legal department
W1FRM
Guest
« Reply #10 on: June 14, 2011, 02:50:14 PM »


Brent:  Your answer to my question below is probably the same as the age old question “Why does a dog lick his BA’s?”

Question:  Why would anybody want to run that much power on 10 meters?

Either that or maybe you plan on entering one of those Good Buddy Band power corntests.  In that case you don’t need no stinkin harmonic suppressors.
Logged
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #11 on: June 14, 2011, 03:37:00 PM »

running juice on 10 is a lot of fun speccially if you have a good antenna
Logged
AB2EZ
Member

Offline Offline

Posts: 1722


"Season's Greetings" looks okay to me...


« Reply #12 on: June 14, 2011, 04:18:06 PM »

Brent

I am a bit puzzled by this statement in your earlier post:

"It seems easy to drive...only 17 watts to excite the tubes to 800 mils..."

If the tubes are running in class AB1, and if the average plate current is 800mA, then: the peak plate current will be around 2400ma; and the amplitude of the plate current waveform at the fundamental frequency will be around 1200mA. [I.e. more than the average plate current but less than pi/2 x the average plate current].

If the B+ is 3600V, the the cathode bias for AB1 will be around -30V; and the peak rf drive voltage (to drive each of the two tubes in parallel to 1200mA) will be around +20 volts (i.e. for a total of 2.4A of peak plate current).

http://gs35b.com/gs35b/index.html

Therefore the amplitude of the rf drive signal will have to be 50V

Therefore, the drive power will be larger than the 0.5 x the amplitude of the rf drive voltage x the amplitude (at the fundamental frequency) of the plate current (which is less than the cathode current) = 0.5 x 50V x 1.2A  = 30 Watts. I would expect that the actual required drive power would be somewhat more than 30 watts

To drive my own 2-tube GS-35b transmitter, when running as an AB1 linear amplifier, at 3.8 MHz, with 3200 Volts of B+ and 400mA (not 800mA) of average plate current, takes about 30 watts from my Ranger.

I'm wondering if your transmitter is biased too far toward class A.

Separately, assuming that the resting current (no rf input, but keyed on) is less than 200mA (or even if it us less than 400mA), it sounds to me like (as others have suggested) the impedance looking into the tank circuit at the fundamental frequency is too low. I.e. the tank circuit (including the tubes' plate-to-ground capacitance, and the effects of the parasitic suppressors and the rf choke) is not resonant at the fundamental frequency.

One way (among many) to check this is to turn the amplifier completely off, and then (temporarily) place a carbon film resistor of value (roughly): 0.64 x 3600V/400mA  = 5760 ohms  from the plate of each tube to ground. Then (with the power still off), use your antenna analyzer to see if the impedance looking into the amplifier's output connector can be made to equal 50 ohms at the fundamental frequency (e.g. 29 MHz) as you adjust the tuning and loading controls on the amplifier. If the output tank circuit is working properly, then it should look like 50 ohms from the output connector looking into the amplifier (with the pair of temporary 5760 ohm resistors in place) ... just as it should look like 5760 ohms/2 = 2880 ohms from the parellel tubes looking out toward the output connector when there is a 50 ohm load at the output of the amplifier.

Don't forget to remove the resistors when you are done. If you do forget, they will remove themselves when you turn the transmitter on :-)

Best regards
Stu


* Amplifier currents.jpg (47.02 KB, 960x720 - viewed 618 times.)
Logged

Stewart ("Stu") Personick. Pictured: (from The New Yorker) "Season's Greetings" looks OK to me. Let's run it by the legal department
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #13 on: June 14, 2011, 05:00:56 PM »

yes Stu, they really will remove themselves. I have done the same test with a scope and signal generator. Then you can sweep it to see where the network is actually tuned.
An exciter that tunes outside the ham bands is also helpful but you might want to use a dummy load or a broadband antenna like a LPDA
Logged
W1IA
Member

Offline Offline

Posts: 778



« Reply #14 on: June 15, 2011, 07:04:30 PM »

Brent

I am a bit puzzled by this statement in your earlier post:

"It seems easy to drive...only 17 watts to excite the tubes to 800 mils..."

If the tubes are running in class AB1, and if the average plate current is 800mA, then: the peak plate current will be around 2400ma; and the amplitude of the plate current waveform at the fundamental frequency will be around 1200mA. [I.e. more than the average plate current but less than pi/2 x the average plate current].

If the B+ is 3600V, the the cathode bias for AB1 will be around -30V; and the peak rf drive voltage (to drive each of the two tubes in parallel to 1200mA) will be around +20 volts (i.e. for a total of 2.4A of peak plate current).

http://gs35b.com/gs35b/index.html

Therefore the amplitude of the rf drive signal will have to be 50V

Therefore, the drive power will be larger than the 0.5 x the amplitude of the rf drive voltage x the amplitude (at the fundamental frequency) of the plate current (which is less than the cathode current) = 0.5 x 50V x 1.2A  = 30 Watts. I would expect that the actual required drive power would be somewhat more than 30 watts

To drive my own 2-tube GS-35b transmitter, when running as an AB1 linear amplifier, at 3.8 MHz, with 3200 Volts of B+ and 400mA (not 800mA) of average plate current, takes about 30 watts from my Ranger.

I'm wondering if your transmitter is biased too far toward class A.

Separately, assuming that the resting current (no rf input, but keyed on) is less than 200mA (or even if it us less than 400mA), it sounds to me like (as others have suggested) the impedance looking into the tank circuit at the fundamental frequency is too low. I.e. the tank circuit (including the tubes' plate-to-ground capacitance, and the effects of the parasitic suppressors and the rf choke) is not resonant at the fundamental frequency.

One way (among many) to check this is to turn the amplifier completely off, and then (temporarily) place a carbon film resistor of value (roughly): 0.64 x 3600V/400mA  = 5760 ohms  from the plate of each tube to ground. Then (with the power still off), use your antenna analyzer to see if the impedance looking into the amplifier's output connector can be made to equal 50 ohms at the fundamental frequency (e.g. 29 MHz) as you adjust the tuning and loading controls on the amplifier. If the output tank circuit is working properly, then it should look like 50 ohms from the output connector looking into the amplifier (with the pair of temporary 5760 ohm resistors in place) ... just as it should look like 5760 ohms/2 = 2880 ohms from the parellel tubes looking out toward the output connector when there is a 50 ohm load at the output of the amplifier.

Don't forget to remove the resistors when you are done. If you do forget, they will remove themselves when you turn the transmitter on :-)

Best regards
Stu
Thanks Stu,
Mark is going to correct my post and give an update on the specifics. As it turns out, as of today the power problem was an improperly loaded amp. We are seeing aprox. the same efficiency on 10 that we have on any other band. Again, I will let Mark update the post.
Logged

Run What Ya Brung!
W1IA
Member

Offline Offline

Posts: 778



« Reply #15 on: June 15, 2011, 07:13:34 PM »


Brent:  Your answer to my question below is probably the same as the age old question “Why does a dog lick his BA’s?”

Question:  Why would anybody want to run that much power on 10 meters?

Either that or maybe you plan on entering one of those Good Buddy Band power corntests.  In that case you don’t need no stinkin harmonic suppressors.

Funny part is Guy...I will most likely never use 10 meters; but when you have spent 2 years building an amp it became a moral imperative to have all the bands working. Besides, some good buddy might offer me a pile of money if it worked on the Good Buddy band HA!! Grin Lips sealed
Logged

Run What Ya Brung!
W2PFY
Contributing
Member
*
Offline Offline

Posts: 13312



« Reply #16 on: June 15, 2011, 09:53:56 PM »

Quote
As it turns out, as of today the power problem was an improperly loaded amp. We are seeing aprox. the same efficiency on 10 that we have on any other band.

You sure know your stuff Brent. I'm happy it all turned out well!
Logged

The secrecy of my job prevents me from knowing what I am doing.
KM1H
Contributing
Member
*
Offline Offline

Posts: 3519



« Reply #17 on: June 16, 2011, 10:09:22 AM »

running juice on 10 is a lot of fun speccially if you have a good antenna

Amen to that. There is nothing like a KW and a big antenna to not have to waste time in a pileup plus it also keeps the riff raff away when in a rag chew Grin I make no apologies for kicking the amp on when needed.....on any mode.
Logged
k4kyv
Contributing Member
Don
Member

Offline Offline

Posts: 10057



« Reply #18 on: June 16, 2011, 01:20:11 PM »

I suspect a lot of hams would be disappointed if they measured their rf output with a  real power measuring instrument on any band. My HF-300 rig, when  loaded up to a KW DC input class-C, indicates only about 3.5 amps with a thermocouple ammeter working into a 50Ω dummy load... which calculates to a little over 60% efficiency. A far cry from the 75-80% you are supposed to get with class C.  And that's with about 50% extra grid bias voltage, driven to full grid current, well beyond the point where additional grid current results in any increase in rf power out. I have tried various L-C ratios in the tank circuit, and that didn't change anything. Also tried several different thermocouple meters just to make sure it was not a meter error. After letting the thing run for several minutes and then turning everything off, and feeling all the components, the only thing that appears warm to the touch is the tubes and the dummy load. That eliminates a loose or high-resistance connection somewhere, or miserable efficiency in one component.

But that's not too surprising, even for 160 and 80, let alone 10m, considering that neither the tank circuit nor the matching circuitry could be expected to be 100% efficient. Plus, when running "leen-yar", the efficiency can be expected to be even less, and with slopbucket, full amplifier efficiency occurs only at the crest of voice peaks.

My Hammy Hambone Mirage "wattmeter" indicates more rf power output (into the flat dummy load) than I am running DC input. If that meter were accurate, I would now be wealthier than Bill Gates, since I would have inadvertently run across a solution to the world's energy crisis.

Once you consider tank circuit losses, ATU losses and transmission line losses, rf power efficiency comparing DC input to the final to RF input to the radiating part of the antenna, is likely to be much less than the ideal figures given in the transmitting tube manuals.
Logged

Don, K4KYV                                       AMI#5
Licensed since 1959 and not happy to be back on AM...    Never got off AM in the first place.

- - -
This message was typed using the DVORAK keyboard layout.
http://www.mwbrooks.com/dvorak
KM1H
Contributing
Member
*
Offline Offline

Posts: 3519



« Reply #19 on: June 16, 2011, 03:32:11 PM »

A little loss each in multiple components ends up to going from 70 to 60%, 75-80? for Class C is a myth unless everything is perfect that day and you won the lottery.

The HF-300 is not exactly a robust 10M tube either and there is considerable loss in long internal leads and vintage design grid structures.

In a 2500W 8877 amp (not mine but I built it for 2500, runs cool at 1500 contest duty) Ive seen a 10% difference on 10M going from #12 interconnect and bandswitch wiring and a 3/16" tank coil to 1/2" copper strap (roofers sheathing) and 1/2" edgewound coils made from transformer material; that shape is more efficient than round. Thats something I discovered only about 2-3 years ago after reading discussions on another forum about tubes with handles.

BTW, did you find that Gates info yet?

Carl

Logged
k4kyv
Contributing Member
Don
Member

Offline Offline

Posts: 10057



« Reply #20 on: June 16, 2011, 10:25:26 PM »

Yes, I have found the xfmr info.  Now just need to get off my arse and bring it to the house where the computer is. Glad you reminded me.

I don't try to use the HF-300 on 10m.  Just 160-80-40.  Nevertheless, when the tube was first released by Amperex in the mid 1930s, its selling point was that it worked well on higher frequencies (hence the name "HF"-300). It was basically sold as a "high frequency" version of the 204-A, whose performance began to crap out above about 3 mc/s. Even so, you couldn't expect it (or any other tube) to function just as efficiently on 10m as on 160m.
Logged

Don, K4KYV                                       AMI#5
Licensed since 1959 and not happy to be back on AM...    Never got off AM in the first place.

- - -
This message was typed using the DVORAK keyboard layout.
http://www.mwbrooks.com/dvorak
K1JJ
Contributing
Member
*
Offline Offline

Posts: 8893


"Let's go kayaking, Tommy!" - Yaz


« Reply #21 on: June 17, 2011, 11:58:15 AM »

Though most AM plate modulated rigs can be made to work on 10M and 6M with some careful work, I think a clean exciter and linear is a great way to go on these bands. The idea of having a rig parked up on 6 or 10M ready to go - and the ease of use and stability is hard to beat with a linear. When the bands go away in the future years, then the lashup can be rowed to a lower band.

It is important, especially on 6M, to have an ultra-clean rig. (Both TX and RX) The atmospheric noise is very low, so that side crud really stands out on a signal.  When the noise flooor is S1 and a guy is S9 +40 over, even a moderately clean signal can be heard up the band as "splatter" much more than on HF.  Guys simply hear a PW signal and if they get trashed by an adjacent clean signal, they usually point the finger.

I've heard some ssb signals running dirty linears that were up the band 15kc. Sometimes they will be there every night and no one says anything. Other times they get lynched... :-)

Even though most everyone is in the same boat running these -30db IMD riceboxes, on AM we tend to have wider signals to boot.

Just something to be aware of.  We need to cut each other more slack on the higher bands when it comes to legit "splatter" because of the lower noise floor.  Striving for a clean rig will give us an edge to work more seemlessly with the crowd, especially if we are QRO and/or using directive antennas.

T
Logged

Use an "AM Courtesy Filter" to limit transmit audio bandwidth  +-4.5 KHz, +-6.0 KHz or +-8.0 KHz when needed.  Easily done in DSP.

Wise Words : "I'm as old as I've ever been... and I'm as young as I'll ever be."

There's nothing like an old dog.
WA1GFZ
Member

Offline Offline

Posts: 11152



« Reply #22 on: June 17, 2011, 12:03:23 PM »

As more people run FFT spectrum displays the crap boxes stand out a lot more.
Logged
KM1H
Contributing
Member
*
Offline Offline

Posts: 3519



« Reply #23 on: June 17, 2011, 12:07:16 PM »

I bet one of those crap signals is from Bristol Tom!

A properly converted SB-200 is a cheap way to get 600W CW/SSB on 6 and enough AM for 50.4 to be useful since most run a lot less. With 125W from the Zeus I can work anything I can hear.
Logged
K5IIA
Member

Offline Offline

Posts: 400



« Reply #24 on: June 19, 2011, 04:23:16 PM »

i hear ya Tom. when i went from a 20 foot high dipole to my array at 80ft. all of a sudden a few guys started mumbaling the W word. its neat that no one, running a piss weak s5 or so, running a rice box wide open is wide.

but a nice 40db carrier with a properly modulated rig can be attacked within minutes of the first transmission.
Logged

Be kind, for everyone you meet is fighting a hard battle.

73, Brandon K5iia
Pages: [1]   Go Up
  Print  
 
Jump to:  

AMfone - Dedicated to Amplitude Modulation on the Amateur Radio Bands
 AMfone © 2001-2015
Powered by SMF 1.1.21 | SMF © 2015, Simple Machines
Page created in 0.085 seconds with 18 queries.