Call Search
     

New to Ham Radio?
My Profile

Community
Articles
Forums
News
Reviews
Friends Remembered
Strays
Survey Question

Operating
Contesting
DX Cluster Spots
Propagation

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

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

donate to eham
   Home   Help Search  
Pages: Prev 1 [2] 3 Next   Go Down
  Print  
Author Topic: Parasitic Suppressor Resistors in Amplifiers.  (Read 29132 times)
W8JI
Member

Posts: 9748


WWW

Ignore
« Reply #15 on: September 28, 2011, 04:49:00 PM »

  Parasitic suppressors are anti-resonant/low-Q  devices, not resonant ones.  Since higher anode RL means more amplification and lower RL means less amplification, VHF parasitic suppressors perform their job by decreasing the VHF RL presented to the anode by the parasitic resonance in the anode circuit. 


One way to test a theory is to take it to the limits the theory sets.

If low "RL" is the goal, the lowest possible RL comes from no suppressor at all. A dead short.

If high "RL" means highest gain, an infinite impedance load to the anode would offer highest gain.

Obviously you are missing something. Your statement is a gross oversimplification, like all 3-500Z filaments should be set for 4.8 volts.

The optimal suppressor impedance depends on the impedance of the anode path in which it is inserted at the problematic frequency. While one answer to all problems sounds great to some people, it is not often correct.

Quote
This resonance is formed by the anode-C, the L of the leads and DC-blocker C between the anode and the L and the C of the Tune-C plus the L of the return path through the chassis and the tube.  Example:  A TL-922:  The total anode C is c. 10pF, the L of the conductors is c. 190nH, and the Tune C is typically 30 to 120pF.  Thus the resonant freq. is c. 120MHz.
 

That is the system resonance, and not the suppressor resonance.

The suppressor is an independent system in series with the anode lead to the tank. The suppressor independently has an equivalent impedance from terminal-to-terminal. Much like the trap in a trap antenna, it can be resonated to peak at a problem frequency and shift more current into the resistor at the problem frequency. This is commonly done in applications where grid resonance that primarily determines frequency of instability approaches operating frequency.   

Quote
The factory stock parasitic suppressors in a 922 have a Q of 5.5 at 100MHz.  The presents a higher than desirable VHF-RL to the anodes of the 3-500Zs - which produces more VHF amplification - so much so that occasionally there will be a 120MHz parasitic oscillation.

Occasionally? One a week, one a month? What makes it stop and start?

Quote
The fix is to lower the VHF O of the suppressors in order to lower VHF gain.


What is VHF "O"?

Quote
There are two ways of doing this:  1.  Add more L to the suppressor inductance and increase  the resistance of suppressor R.  Unfortunately this increase the heat generated in the R exponentially** at 28MHz so that a much higher dissipation R is required for 28MHz operation.  or 2.  Increase the resistance of R,  use low-L resistors capable of dissipating more heat, and use resistance wire to lower the Q the inductor.  When this is done with double-suppressors per tube the VHF-Q can be reduced to 1.5 at 100MHz, thereby lowering VHF amplification  -- which in turn lowers the ability to sustain oscillation. 
 
 
......and of course the best or only cure for this "oscillation", the one that occasionally or very randomly pops up and damages a part and then goes away,  is the one you sell.     :-)
Logged
W1BR
Member

Posts: 4189




Ignore
« Reply #16 on: September 28, 2011, 05:25:59 PM »

My question to Tom was  if you know the self resonant frequency of the
tube in question, would using a parallel resonant supressor with a high C to L ratio be something
that is worth investigating.
  
The goal would be to minimize reactive losses at 30 MHz and below by using a LC tuned circuit that
would provide a low impedance path at HF, and a very high impedance path at the parasitic frequency.
this would force the VHF energy through the resistor, killing the Q at the self resonant, possible
parasitic requency.

Using nichrome wires in the path seems to be counter intuitive to minimizing HF loss.

« Last Edit: September 28, 2011, 05:27:41 PM by K1ZJH » Logged
W8JI
Member

Posts: 9748


WWW

Ignore
« Reply #17 on: September 28, 2011, 08:45:59 PM »

My question to Tom was  if you know the self resonant frequency of the
tube in question, would using a parallel resonant supressor with a high C to L ratio be something
that is worth investigating.

Absolutely. It is a good method for minimizing loss on desired frequencies while maintaining effective suppression at a prolematic frequency.
  
Quote
The goal would be to minimize reactive losses at 30 MHz and below by using a LC tuned circuit that
would provide a low impedance path at HF, and a very high impedance path at the parasitic frequency.
this would force the VHF energy through the resistor, killing the Q at the self resonant, possible
parasitic requency.

There will be a certain optimum dampening resistance. My parasitic suppressor page goes into all of this.


Logged
G3RZP
Member

Posts: 1280




Ignore
« Reply #18 on: September 29, 2011, 11:10:01 AM »

If you look at older ARRL Handbooks, you will see that one of the suggested parasitic suppressors was a resistively loaded tuned circuit coupled to the plate through a small link.

A straight RLC network was used in a 80kW tx for this purpose in the 1960s.
Logged
AG6K
Member

Posts: 1




Ignore
« Reply #19 on: September 30, 2011, 10:15:17 AM »

  Parasitic suppressors are anti-resonant/low-Q  devices, not resonant ones.  Since higher anode RL means more amplification and lower RL means less amplification, VHF parasitic suppressors perform their job by decreasing the VHF RL presented to the anode by the parasitic resonance in the anode circuit. 


One way to test a theory is to take it to the limits the theory sets.

If low "RL" is the goal, the lowest possible RL comes from no suppressor at all. A dead short.

  SO IT MIGHT SEEM, BUT THE PARASITIC SUPPRESSOR IS NOT CONNECTED FROM THE ANODE TO GND.  Murphy was right about things being more complicated than they first look. 

If high "RL" means highest gain, an infinite impedance load to the anode would offer highest gain.

  Correct Tom but since the VHF parasitic suppressor is not in parallel with the anode the engineer must do a series to parallel transformation to find out what the VHF RL is.  This procedure is covered in "Calculating Power Dissipation in Parasitic-Suppressor Resistors", March, 1989 QST, page 7, 'Finding Impedance by Solving for Admittance'. .® The American Radio Relay League, Inc. - note - Series/parallel transformations require considerable thinking and minimal drinking.   

Obviously you are missing something. Your statement is a gross oversimplification, like all 3-500Z filaments should be set for 4.8 volts.

  In "Care and Feeding ..." Eimac recommends operating  Th-W filament tubes at the lowest potential (+1%) that provides no decrease of peak emission - but Not at a potential below the Eimac's recommended minimum filament V.  The minimum recommended  for the 3-400Z and the 3-500Z filament is 4.75V.   What do you find wrong with 4.80V Tom?

The optimal suppressor impedance depends on the impedance of the anode path in which it is inserted at the problematic frequency.

 say what ?

While one answer to all problems sounds great to some people, it is not often correct.

  A Low VHF Q  parasitic suppressor produces a relatively low VHF-RL at the anode.  Lowering VHF-RL produces lower VHF amplification.   Lower VHF amplification =s less chance of VHF oscillation.   - note - osc. is due to the unavoidable feedback path between the input and the output of all tubes. . 

Quote
  This resonance is formed by the anode-C, the L of the leads and DC-blocker C between the anode and the L and the C of the Tune-C plus the L of the return path through the chassis and the tube.  Example:  A TL-922:  The total anode C is c. 10pF, the L of the conductors is c. 190nH, and the Tune C is typically 30 to 120pF.  Thus the resonant freq. is c. 120MHz.
 

That is the system resonance, and not the suppressor resonance.

  suppressors aren't, they are low-Q anti-resonant devices that produce low parallel-equivalent R at the anode. 

The suppressor is an independent system in series with the anode lead to the tank. The suppressor independently has an equivalent impedance from terminal-to-terminal. Much like the trap in a trap antenna, it can be resonated to peak at a problem frequency and shift more current into the resistor at the problem frequency. This is commonly done in applications where grid resonance that primarily determines frequency of instability approaches operating frequency.   

Quote
  The factory stock parasitic suppressors in a 922 have a Q of 5.5 at 100MHz.  The presents a higher than desirable VHF-RL to the anodes of the 3-500Zs - which produces more VHF amplification - so much so that occasionally there will be a 120MHz parasitic oscillation.

Occasionally? One a week, one a month? What makes it stop and start?

  Who knows what made the fuel pump crap out on my 1973 Dodge Maxivan?  How come the original brakes are still good?  Forrest Gump said "Shit happens".  Maybe the brakes have the good kind and the fuel pump had the bad kind?  IME intermittent VHF parasites are somewhat mysterious but I feel like I know how to discourage them.

Recommended reading:  Fyler, G. W. ''Parasites in Transmitters'', Institute of Radio Engineers journal. Sept. 1935

Quote
 The fix is to lower the VHF O of the suppressors in order to lower VHF gain.


What is VHF "O"?

  X/R at some frequency between 30MHz and 300MHz.  At 100MHz a SB-220's VHF parasitic suppressors have a Q of c. 5. 

Quote
  There are two ways of doing this:  1.  Add more L to the suppressor inductance and increase  the resistance of suppressor R.  Unfortunately this increase the heat generated in the R exponentially** at 28MHz so that a much higher dissipation R is required for 28MHz operation.  or 2.  Increase the resistance of R,  use low-L resistors capable of dissipating more heat, and use resistance wire to lower the Q the inductor.  When this is done with double-suppressors per tube the VHF-Q can be reduced to 1.5 at 100MHz, thereby lowering VHF amplification  -- which in turn lowers the ability to sustain oscillation. 
 
 
......and of course the best or only cure for this "oscillation", the one that occasionally or very randomly pops up and damages a part and then goes away,  is the one you sell.     :-)

  chortle.  If an owner of a SB-220 is uncomfortable knowing that his amp has suppressors with a Q of 5, he can buy one of our retrofit kits, do some 221ºC Ag-Sn soldering, and decrease the suppressor Q to 1.5.  If he/she thinks that the amplifier is not more stable with low-Q parasitic suppressor he/she gets a refund cheque from yours truly.    cheers
Logged
N2EY
Member

Posts: 5083




Ignore
« Reply #20 on: September 30, 2011, 12:50:02 PM »

In "Care and Feeding ..." Eimac recommends operating  Th-W filament tubes at the lowest potential (+1%) that provides no decrease of peak emission - but Not at a potential below the Eimac's recommended minimum filament V.  The minimum recommended  for the 3-400Z and the 3-500Z filament is 4.75V.   What do you find wrong with 4.80V Tom?

I don't know about Tom, but here's my take on it:

If I had an amplifier with a genuine Eimac 3-500Z or 3-400Z, and it could be guaranteed that the filament voltage would never, ever drop below 4.75 volts if set at 4.8 volts, I might consider running the tube that way. Maybe.

But such conditions are simply not realistic in 99.99% of amateur radio situations. Here's why.

To meet the above conditions, you need:

1) A true-RMS AC voltmeter with accuracy better than 1%. Otherwise, when you set for 4.8 volts you could actually be at 4.75 volts or lower.

2) A setup that can measure the filament voltage at the tube pins while under full-power operating conditions.

3) A line supply and filament transformer that are regulated to better than 1% under all conditions so that the filament voltage doesn't ever drop below 4.75 volts

4) Test equipment to measure the IMD, power out, etc. as recommended by Eimac, to be sure that 4.8 volts really is OK.

IOW, to run at 4.80 volts when the lower limit of the specification is 4.75 you need a filament supply and voltage measuring system where the total combined errors from all factors is less than 1% under all conditions. Otherwise there will be times when the minimum voltage specification isn't met, the tube runs too cold and useful life is actually reduced.

The plain and simple fact is that 99.99% of hams, including me, just don't have the resources to run that close to the edge. Not even close.

So what 99.99% of hams should do is to aim for the design-center voltage - 5.0 volts under load. That way the chances of being out of specification are minimized and tube life is maximized.  

Which is the whole point of the filament-voltage discussion.

Of course it is counter-intuitive that running too cold can be as bad as running too hot. But that's how it is. Transmitting tubes with thoriated-tungsten filaments are not the same things as light bulbs.

73 de Jim, N2EY
« Last Edit: September 30, 2011, 12:55:51 PM by N2EY » Logged
W8JI
Member

Posts: 9748


WWW

Ignore
« Reply #21 on: October 01, 2011, 05:04:44 AM »

AG6K wrote:
Quote
SO IT MIGHT SEEM, BUT THE PARASITIC SUPPRESSOR IS NOT CONNECTED FROM THE ANODE TO GND.  Murphy was right about things being more complicated than they first look. 


The common suppressor is indeed connected from anode to ground in a grounded grid amplifier. The path is anode through the anode leads through the blocking capacitor and through the tuning capacitor to ground, which is also where the grid is.

The suppressor is normally, although it does not have to be, in series with the tank connection.

I said:
If high "RL" means highest gain, an infinite impedance load to the anode would offer highest gain.

AG6K said:
Quote
  Correct Tom but since the VHF parasitic suppressor is not in parallel with the anode the engineer must do a series to parallel transformation to find out what the VHF RL is.


Sorry Rich, but the suppressor is in SERIES with the anode path. The easiest way to find the suppressors effect at VHF is to look at series impedance of the path for a normal series connected suppressor.

AG6K: 
Quote
This procedure is covered in "Calculating Power Dissipation in Parasitic-Suppressor Resistors", March, 1989 QST, page 7, 'Finding Impedance by Solving for Admittance'. .® The American Radio Relay League, Inc.

Since the suppressor is in series with the anode path, one only has to look at the series equivalent impedance and calculate heat with current. There is no magic to nichrome. It de-Q's the system at HF much more than a regular suppressor, while not offering anything that cannot be done some other way at VHF or UHF. It's 1920's breadboard suppression technology.

AG6K wrote:
Quote
In "Care and Feeding ..." Eimac recommends operating  Th-W filament tubes at the lowest potential (+1%) that provides no decrease of peak emission - but Not at a potential below the Eimac's recommended minimum filament V.  The minimum recommended  for the 3-400Z and the 3-500Z filament is 4.75V.   What do you find wrong with 4.80V Tom?


If you can't understand what is wrong with your suggestion of something so simple as setting a voltage, how can you ever hope to understand more complex things like suppressors and how they work?? 

If the minimum operating voltage is 4.75 volts and you are the single source telling Hams to set voltage at 4.8 volts, you are causing virtually 100% of people following your advice to set voltage below published voltage levels. This is because you give no instructions to test emission, find the lowest (and/or  highest) mains voltage, what meter to use, or even where or how to measure.

Look at the mess it caused everyone because QST and an author used your 4.8 volts as a target reference! The nichrome is not much different than the filament voltage, and neither is moving a balun to the input of a floating tuner network.

73 Tom
Logged
N4NYY
Member

Posts: 5224




Ignore
« Reply #22 on: October 01, 2011, 08:02:04 AM »

OK, I am trying to follow this thread, but the one dude is putting everything in quotes, including his replies. That is not helping.
Logged
W3LK
Member

Posts: 5639




Ignore
« Reply #23 on: October 01, 2011, 10:22:32 AM »

OK, I am trying to follow this thread, but the one dude is putting everything in quotes, including his replies. That is not helping.

Because the poster doesn't know how to use the quote function correctly. Sad
Logged

A smoking section in a restaurant makes as much sense as a peeing section in a swimming pool.
N4NYY
Member

Posts: 5224




Ignore
« Reply #24 on: October 01, 2011, 11:39:18 AM »

Quote
Because the poster doesn't know how to use the quote function correctly. Sad

Yep. Maybe he can figure it out.
Logged
KA5N
Member

Posts: 4380




Ignore
« Reply #25 on: October 01, 2011, 12:24:12 PM »

Any thread longer than two yards long will take you to places you don't need to be.
Time to tune out.
Allen
Logged
N2EY
Member

Posts: 5083




Ignore
« Reply #26 on: October 01, 2011, 12:51:25 PM »

OK, I am trying to follow this thread

It's really pretty simple.

Tube type HF amplifiers used by hams often have oscillations at frequencies other than the desired ones because there are unavoidable resonances in the circuit. Usually these are in the VHF region. A typical HF amplifier circuit looks different to VHF frequencies in such a way that the wiring itself becomes a tuned circuit.

These "parasitic oscillations" or "parasitics" can't usually be completely cured by neutralization, bypassing or shielding. (Those things help, and are good practice, but most amplifiers need more).

The usual cure is to insert one or more "parasitic suppressors" in the VHF oscillator circuit, in such a way that VHF oscillations can't start or be sustained.

The typical parasitic suppressor is a small inductance in parallel with a noninductive resistor. The idea is that the inductor value is small enough that it is essentially a short-circuit at HF, yet large enough at VHF that the resistor is effectively in the circuit for VHF. The resistor introduces enough loss at VHF that parasitic oscillations can't start or continue.

That's the theory. In practice, suppressing parasitics can take a bit of trading-off, particularly in amps covering the higher HF bands like 10 meters. This is because an inductor-resistor combination with values of L and R large enough to stop a parasitic may have some loss at HF too.

One possible option to reduce losses at HF is to add a capacitor in parallel with the inductor so that the combination becomes a tuned circuit at the VHF parasitic frequency. This works, and in fact the Ancient Ones sometimes did it as a TVI reduction measure. But it has drawbacks:

1) More parts

2) Only suppresses one parasitic frequency

3) Requires adjustment intially.

The usual resistor-inductor combination has been used for decades with great success when properly applied. W8JI's website description of cleaning up a TL-922 is IMHO a good example (complete with pictures).

73 de jim, N2EY

   
Logged
N4NYY
Member

Posts: 5224




Ignore
« Reply #27 on: October 01, 2011, 01:45:19 PM »

Thanks, Jim. I only have minimal experience with suppressors as I typically replace the original, as in Heathkits case. I do not use non-original setups, unless there was a service bulletin to correct a flaw. I try to use these threads to try and learn something, but it is difficult to follow, especially when some answers were put in a quotes highlight.
Logged
W8JI
Member

Posts: 9748


WWW

Ignore
« Reply #28 on: October 01, 2011, 01:57:02 PM »

Thanks, Jim. I only have minimal experience with suppressors as I typically replace the original, as in Heathkits case. I do not use non-original setups, unless there was a service bulletin to correct a flaw. I try to use these threads to try and learn something, but it is difficult to follow, especially when some answers were put in a quotes highlight.

I'll try to expand what I have on this including photos, so if something is unclear let me know.

http://www.w8ji.com/vhf_stability.htm

Tubes with the greatest stability problems are older designs with long thin grid leads. Many modern tube designs in good layouts require no suppression at all. There are about a dozen different ways to skin the same cat, but generally the best way is the way that affects the desired frequencies the least and the indesired frequencies the most.

73 Tom
Logged
AG6K
Member

Posts: 1




Ignore
« Reply #29 on: October 01, 2011, 04:09:43 PM »

In "Care and Feeding ..." Eimac recommends operating  Th-W filament tubes at the lowest potential (+1%) that provides no decrease of peak emission - but Not at a potential below the Eimac's recommended minimum filament V.  The minimum recommended  for the 3-400Z and the 3-500Z filament is 4.75V.   What do you find wrong with 4.80V Tom?

I don't know about Tom, but here's my take on it:

If I had an amplifier with a genuine Eimac 3-500Z or 3-400Z, and it could be guaranteed that the filament voltage would never, ever drop below 4.75 volts if set at 4.8 volts, I might consider running the tube that way. Maybe.

 Heath SB-220s run their 3-500Z filaments close to this potential when operated from 234vac.  One of the tubes in my SB-220 is an Amperex that was mfg in 1967 and it still does its job. 

But such conditions are simply not realistic in 99.99% of amateur radio situations. Here's why.

To meet the above conditions, you need:

1) A true-RMS AC voltmeter with accuracy better than 1%. Otherwise, when you set for 4.8 volts you could actually be at 4.75 volts or lower.

2) A setup that can measure the filament voltage at the tube pins while under full-power operating conditions.

3) A line supply and filament transformer that are regulated to better than 1% under all conditions so that the filament voltage doesn't ever drop below 4.75 volts

  And what happens when a 3-500Z's filament potential drops slightly below 4.75v ?  In tubes with good vacuums, nothing.  However if the tube is slightly gassy. the filament's 1.5% Thorium (Th) content  becomes contaminated and emission decreases.  The fix is to raise the filament V temporarily to expel the O2 and N2 atoms that have contaminated the Th.

4) Test equipment to measure the IMD, power out, etc. as recommended by Eimac, to be sure that 4.8 volts really is OK.

IOW, to run at 4.80 volts when the lower limit of the specification is 4.75 you need a filament supply and voltage measuring system where the total combined errors from all factors is less than 1% under all conditions. Otherwise there will be times when the minimum voltage specification isn't met, the tube runs too cold and useful life is actually reduced.

  but the fix is fairly easy. When I had an 8170 rebuilt by Econco I was surprised to find out that the recarburized 7.5V, 75A filament had so much emission that  I could turn down the potential to 6.7V with no decrease in peak output.  Since the Eimac recommended minimum filament potential was 7.15V, I phoned Dave at Econco, explained what Iwas seeing, and he explained that going below the factory recommended V was only a problem if peak output decreased or if the tube had many air atoms inside. 

The plain and simple fact is that 99.99% of hams, including me, just don't have the resources to run that close to the edge. Not even close.

  the edge is not serious. 

So what 99.99% of hams should do is to aim for the design-center voltage - 5.0 volts under load. That way the chances of being out of specification are minimized and tube life is maximized.  

  only in the minds of those who don't know. Operating a Th-W (thorium-tungsten) filament  more than 2% above the V that produces max PEP reduces tube life by the ratio of the filament potentials raised to the 23.4 power.  IOW, each 3% increase in filament potential reduces emissive life by 50%.  Example 10v/10.3v ^23.4 = 0.5, 10.3v /10v = 2. 

Which is the whole point of the filament-voltage discussion.

Of course it is counter-intuitive that running too cold can be as bad as running too hot. But that's how it is.

   "It ain't what you don't know that gets you into trouble. It's what you know for sure that just ain't so."
 — Mark Twain

>Transmitting tubes with thoriated-tungsten filaments are not the same things as light bulbs.

  Correct.  Light bulbs have 100% W filaments while electrom tubes have 98.5% W filaments.

CHEERS JIM

73 de Jim, N2EY
Logged
Pages: Prev 1 [2] 3 Next   Go Up
  Print  
 
Jump to:  

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