The AM Forum

OK guys, who can answer this?
Friend of mine called me last night rebuilding a old amp.
The Parasitic chokes are burned up and wanted to know where to find
"Carbon" 47 ohm resistors.
Our local parts store only stocks flame-proof resistors now...
Will the metal flame-proofs work or to "construct"
a parasitic choke do you need to use the carbons ones? & why?
73 steve w8tow

Allied.  www.alliedelec.com

Quick and easy.  You hafta pay a handling charge on orders <$25, but they still have carbon resistors, high voltage capacitors, etc.

BTW I was looknig the other night in their catalog and found some FB high-power noninductive resistors.  They look like a transistor and are intended to be mounted on a heatsink, but  they claim <80 nH inductance.  Just what the doctor ordered for a rhoimic termination, grid swamping network, or a dummy load.

73 John

Watch out for situations where you might accidentally over dissipate some of those packages. They can go off like M80's. I had one on a power amplifier screen supply go kaboom. Pictures at 10.

Yes, 2 watters, will do...
RF parts only has 1 watters
tks es 73 steve
8tow

A little discussion about these parasitic chokes, please.  I have heard three schools of thought about them.  1.  I understand Art Collins had an understanding about them and used nichrome wire for the coil around globar resistors of about 100 ohms each.  2. The current thing Steve has mentioned, 3 to 7 turns of buss wire around a carbon comp. resistor of about 47-51 ohms.  And

3. Are they really necessary at all?  I have seen several shortwave, high power stations that do not use them at all.  

So are they really necessary, and if so, what configuration?

Are the parasitic suppressors necessary to supress parasitics that result from poor construction practices?

You bet 100%
But sometimes "poor construction practices" includes tubes, switches, etc., that can't be easy redesigned.

Ok if they exist, what kind of construction?

I've used flame proof resistors at RF with no problems for years.

Length of tube electrodes is more than zero, distances between them are less than infinity.

I found an old gentleman in Florida that stocks or has new old stock of resistors, I believe he specializes in them only  1/8 watt to 1000 watt.  Most any thing you want.  Unfortune has it, there is a minimun order

http://minresistors.com

Art Collins put a ham transmitter on the market in the mid 1930's.  I forget the model number, but it used a single triode (211 or 203A) in the final and ran about 100 watts on AM.  It used the "Collins coupler" tank circuit, an early form of pi-network while nearly all other ham rigs, both commecrial and homebrew, were using inductive coupling in the output.  

What was unusual about this transmitter was how smoothly it operated.  Most ham homebrew rigs were squirrelly to say the least.  The secret:  Collins had figured out that the spastic operation of hf class-c finals was due to vhf parasitics (known as uhf back then).  His rig included the now well-known parasitic choke consisting of a noninductive resistor with a coil of wire wound over it.  With parasitics tamed, the tx tuned smooth as velvet.

I have used 5- to 10-watt carbon resistors for my parasitic suppressors.  Found them in junkboxes at hamfests.  I think they were standard items of manufacture in the WW2 era.  The Gates BC1-T uses a bundle of 1-watters in parallel.  Glo-bars are excellent if you can find them.  I would be leery of "flameproof"  resistors.  It has been my experience that they instantly break out in flame when severely overloaded.  Maybe they don't combust, but the arc that results when they self-destruct sometimes produces a spectacular flame.  Flameproof resistors go excellently with solid state electronics, since both self-destruct immediately once a threshold of instantaneous overload is exceeded.

In one of my  rigs I use a set of Ohmmite parasitic chokes manufactured circa 1939.  They look like standard "brown devil" type 10-20w wirewound resistors with a coil of wire around them, typical parasitic suppressor style.  I'm not sure, but I suspect these are "noninductive" wirewound resistors, with two layers of wire connected in parallel, each wound in the opposite direction as the other.  They serve their function perfectly.  I have a set of spares, but am not interested in parting with them.

It's Ok, it's 10% tolerance.

Where else besides in the parasitic supressor in the plate lead is non-inductive important?  I'm bringing a linear back to life which had carbon resistors as the plate current meter shunt in the B- cathode lead -- 3 1W 4.7 ohm carbon comp resistors in parallel but on the power supply side of an RF choke for two 3CX800A7's grounded grid.

I think the carbon comp resistors can be replaced by a wirewound 1.5 ohm because (a) they're on the DC side of the RF choke and (b) the inductance of the wirewound 1.5 ohm resistor is small anyway, about 200 nH.  

Is this wishful thinking on my part?

Jon, K6JEK

200 nH combined with parasitic capacitance = FB resonant circuit at VHF.

3CX800s are expensive enough as it is without risking them for lack of a 50 cent resistor.

Why not just order them from Allied?

73 John

So this weekend I found a 47 ohm 2 watter and decided to conduct a bit of an exeperiment.  In one of my HB rigs, (1940 PP Taylor T40's....807 driver and 6V6 osc) I replaced an existing 47 ohm paracitic choke I previously constructed using a flame-proof resistor. This resistor was located on the plate of the 807.

One of the grids of the T40s has a Ohmite Z-1 to it.

This rig has always behaved oddly. Tune-up was tricky, esp on 40m.
When modulated, the waveform never looked as good as I would like.


AFTER replacing the paracitic choke on the 807 plate this weekend with a CARBON 47 ohm resistor, it tunes much better AND the ragged edges on the modulation envelope are now gone!!!
If you can get by using the flame-proof resistors, more power to you,
but with old buzzard circuits, find a carbon resistor for the chokes form...
COnsider this:
If the flame-proof resistor has a foil surounding the composite innards, how will the coil you wrap around it behave that close to the foil?
The "Q" of the R-L circuit will change, as will the resonant freq....
I am off to find carbon resistors, got 4 more to change out!!!
73 steve
8tow

Yes.... and keep it authentic. There's nothing like the nostalgic smell of a burned carbon comp parasitic !!!

What about using Ferrite Beads????  I have an old SBE 33 slopbucket rig and for parasitics it uses ferrite beads on the plates leads going to the 27GB5 output tubes.

Anybody got any input on this idea?????

Interesting, Mike.  Are the beads along the whole length of the lead or are there just a few near the cap?

Beads and wire wound resistors will mess up the tank Q.
The resistor blowing up means one of two problems.
Bad layout or excessive L in the suppressor.
Sometimes a high VSWR will also take them out.
Flame proof film resistors can't handle over load as well
as carbon comp but the resistance is more stable.

A strapper amoung us puts his suppressor in the grid where it belongs.
This requires a superior layout though.

My SBE 33 has only one bead near to the plate cap on each tube.

I wonder if the SBE34 might be the same way??

Usually 1 bead will have a series Z of about 50 ohms. Compare that to the plate Z of 2000 ohms. As you go higher in frequency the bead will have more effect on the tank coil. The bead material will be lossy at VHF
to kill the Q at self oscillation frequency. A resistor across an inductor is doing the same thing. The bead will heat up just like the resistor.
The material can be damaged if it gets too hot. I would avoid a bead
in a high power rig where there is a lot of ambient heat.
layout is still the key to a stable rig. The tube plate to plate tune cap
needs to be as short as possible and the cap to ground is also a big deal.

Also, cathode bypass and plate filter caps should be connected as close as possible to each other and to the cathode of the output lamp, also the input cap of the antenna tuner should be connected to them by the shortest way. It is the output current's path. Inductances of conductors, be it a part of a chassys, or just a wire, must be always considered (it is a common mistake!). Also, RF currents are significant in the PA, so if a "bias filter" cap is grounded between antenna tuning breadslicer and a RF plate filter cap there will be a parazitic feedback.

Just draw a full equivalent diagram that includes all wires with their inductances and resistances, and capacitances between elements, to see better how to optimize the layout.

It's a world gone mad.

I was thrilled that my local surplus store had a supply of 1 W 4.7 ohm carbon composition resistors.   Now I can replace the three paralleled 1/2 W carbon resistors that blew with these 1 W carbon.  Imagine my surprise when I measured the inductance of the carbon resistors and discovered it is higher than the metal film 1 W guys.  Both, of course, are much lower than the wirewound I almost put in.

Do I put in the metal film because lower inductance is better or do I put in the carbon composition because everyone knows you should use carbon composition in situations like this -- the cathode lead of a pair of 3CX800A7's, grounded grid.

I think I better have a drink and think it over.

Metal films are always better.  Carbon comps were used in the old buzzard days because that's all they had (within a reasonable cost target).  They're more stable, less inductance, less noise (not that it matters in a parasitic supressor), and if they overheat reasonably they're not permanently damaged as easily as carbon comps.

73 John

I have a different experience, John. Yes, metal films are more stable and less nolsy, but when overloaded they usually produce a smoke and a tiny spark running accross the body... Being a student I used them when did not have a cigarette lighter. I covered one such resistor by a small piece of a cotton from a first aid kit, and using plyers put it into the 220V socket. One Puff! and I'm smocking my cigarette... ;)

I wrapped some elestric toaster wire around a 100 Ohm carbon resistor and stuck it across my mic input lead. Now, no one will answer my CQ.

I doubt very much that metal film resistors "L" measured at 30 mhz and above is lower than a carbon resistor.
I swept several this weekend,
At 1 khz, they measure about the same "L", but at 1 mhz, the metal film (same R values) goes up x 10, and at 50 mhz, the L of a metal film is outta sight,
so.... I think if ya wanna get rid of VHF/UHF paracitics, you would not use a
metal film resistor as a form, but, that is just me....
73 to all

I'll measure some this weekend - I have an old buzzard tube-type Rohde&Schwartz network analyzer.  What values did you measure, Steve?

73 John

I have checked the failure modes of most resistor types. Only wire wound
resistors burn open. All others I have tested go low in resistance before
they burn open. This can cause cascading failure modes. A wire wound resistor always increases in resistance until it opens providing it doesn't flash over with high voltage.

So what if it ain't a carbon resistor???  I have some so called noninductive resistors around here that appear to be  two resistive wire windings that counter each other, they are called noninductive.

 Could these be used as part of a parasitic choke????

I have some of these used as dummie loads, and they do have SWR reflected of about 1.2 to 1.

JOhn, I checked the following values (both flamers and carbons)
39 ohm
47 ohm
56 ohm
63 ohm
I made a typical P choke for an 807, 6 turns of #18 on
the resistor.
oh, these were measured on a HP LCR analyzer
73 steve

6 turns is a lot of L and the 807 is a real VHF generator.

My 4CX3000A has 1 turn of 1/4 inch strap 5/8 inch I.D. with 5 220 ohms 2 watt carbons in parallel. 6 turns in a linear may have problems on 10 meters. I think the SB220 had 2 turns on a 47 ohm 2 watt.   fc

Steve, I'm not sure that flameproofs are metal films, or at least the same type of metal films I'm talking about.  I'm talking about RN55 1 percenters as well as 1/4w film resistors as purchased from DigiKey.

Flameproofs are generally used as current limiting resistors in the primary circuit of siwtch mode power supplies, TV sets, etc., and are not intended for any sort of RF service.  The reason they're "flameproof" is so that they don't present a fire hazard if they overcurrent (although someone posted the exact opposite experience!).

I have a HP digital L-R-C bridge as well.  I'll run a bunch thru and see what I find.

Y'know, I'd think a wirewound resistor would make a swell plate RF choke - it's a great lossy inductor, which would go a long way towards quelling unwanted resonances.  Anyone tried this?

73 John

John,
I had an old MARS article of a 4CX1000A linear. The builder blew the
B&W 800 plate choke and rewound it with wire removed from a power
resistor. He claimed it worked great. The B&W 800 has a resonance
around 25 MHz. The wire resistance should kill the q of the choke.
The trick is finding resistor wire low enough resistance that is long enough.
I know a number of guys use resistance wire for parasitic chokes.

Flame proof resistors are only supposed to contain the flame. My testing did not include these I was doing RLR, RNC, RWR, and RN60 style mil parts.
I found film resistors dropped a lot in value as they got red. We were sensing EED squibs so needed to limit current under all failure contitions.
fc

John, these resistors were "NTE" off the shelf "flameproofs"
2 watters..
I took a hammer to one and inside is a foil around the carbon...
perhaps if a person wanted to, they could crape away the foil,
but I think its easier to buy a old carbon resisitor or perhaps use some
pencil lead!
steve

I have some of those, too, Steve.  I'll do some experimenting with various resistors this weekend - 's the way to find out fer sure!

73 John

I used 3 in parallel 50 ohm wirewound inside 4 turns of #12 copper 3/4 in dia on my 4x1... works FB on 80, 40 and 20.......does not get hot.....Before I put this on top the tube, I had some jumpy tuning and some arcing of the plate tune cap air variable...I tried the wirewounds cause I didn't have any carbons....

I think a wire wound may have a problem with the end termination where the lug is located acting like a shorted turn at each end. RWR resistors have a cap at the end that may not act like a shorted turn. Take a 10 watt or 25 watt wire wound with a shorted turn at each end I suspect this could be a problem. Say you had a 175 watter at a fairly low resistance there
may be enough inductance to act like a plate choke.
When I first built my 4-1 rig I had 4 turns also. It didn't work well on 15 and 10 meters. I eventually got it down to two turns 1/2 inch I.D. and worked well on ten meters. It took a number of layout changes to get it right. It is not easy with a tall tube.   fc

Yes ...I'm still experimenting with the 4x1...When I come across some carbons, I'll try them...I am using a B and W tank coil and a breadslicer @ 4KV and so far it won't go to 10 meters.......

Gentlemen, why don't calculate how a wire is acting, how end short turns are acting, what's inductance, what's resistance we need, and so on, having in mind what do we need to supress parasitics oscillations but don't affect too much output on the high end of the band, instead of just reading tea leaves? :lol:

There's the engineer in you showing, Tolly!

The problem is that we don't know a lot about the actual construction of these resistors.  Molded wirewounds are almost impossible to see how many turns are inside, and whether there is a shorted turn or not.

However, there are certainly computer programs that could model a parasitic supressor - however, they're based on a more or less ideal resistor = and how the resistor affects the performance of the suppressor is the whole thread here.

BTW did you find a manual for your R-392?  If not, there's one here:

http://bama.edebris.com/manuals/military-kg7bz/r392/

73 John

The real problem is the mechanical layout not so much the tube.
I agree with Tolly but we all live out of our junk boxes. VERY low resistance RWR style resistors may work very well.
I actually used them on the modulator plates of my PDM Viking 2. I used
6.8 ohm 3 watt RWR resistors at the plates of the 6DQ5s
They have been there for 20 years now. I never measured the inductance
since the PWM is running at 75 khz. Small pulses still make it throught the circuit.  fc

Thanks John, downloading now!

I hope after disassembling I'll find a way to resolder that filaments inside of the crystal osc chassis...

Tolly,
You might be able to get a new module from Fair Radio Cheap.
Good luck!

Thank Frank,
I have a nice audio chip from TV, it is Hi-Fi amp and was powered exactly from 28V source! However, I'll need to protect it simehow, because a squelch relay will defenitely damage it shorting output.

Tolly that is great. Just mount it to the side wall for plenty of heat sink
then use the base of a dead tube as a plug to connect it to the audio output tube socket. Mil receivers like the Racals and Cubics use a series resistor at the output of the audio IC to protect the output. I think Cubic uses 33 ohms if I remember. This way the IC can handle a dead short
on the output. This would be in the line level path. I have a RA6830 ISB
board with resistors at the aduio amplifier output for each channel that
drive little board mounted 600 ohm line level transformers. I suspect
the R392 does the same thing.Racal and Cubic both
have the connection to the speaker between the IC and resistor. This is not protected. The R392 audio output is 600 ohms. I suspect a 28 volt
audio IC will make a lot of power to drive the output transformer with a
series resistor. You should have no problem driving a speaker if it is matched to 600 ohms. The R392 audio output was under 1 watt if I remember. You might want to AC couple the output of the Amplifier
with around 470 to 2000 UF to get good low end response. fc

Thank you Frank again! Corrections: it was my mistake, the squelch relay don't short output trans, it shorts one more trans that is input trans, I forgot about it.
I am going to mount a heatsink on top of a tube base. And one more redesign: instead of plugging input of the amp into the 6BJ5'th socket I'll crocodile it to the volume regulator through the shielded wire.

Also, I figured out how to convert filament feeding without resoldering wires on tube sockets. What bothered me, were couple tubes buried inside of the crystal oscillator subassembly. Fortunately, a resistor that drops 12V for that tubes is conveniently located underneath the RF chassys and may be easily shotred.

So, 12BA6 and 12AT6 may be easily installed!

AH very good converting to 12 volt tubes. You could use 3 terminal regulators to regulate the 28 volts down to 12V or put pairs in series
to operate off 28 VDC. I thought the 2 tubes in the crystal oscillator
were in series. My 392 was so clean inside I hated to tear into it.
I sold it when I got an R390A.  I have friends who installed relay controlled mechanical filters to help bandwidth skirts.  fc

Yes, filaments of 2 tubes in crystal oscillator are in series, and in series with 80 Ohm resistor. If to short the resistor the whole set may be easily converted to be powered from 12 V source so 12V version of tubes may be used. However, tubes from AF desk must be in such case removed and replaced by a hi-fi IC.  One more store is a crapstal calibrator, where transistors may be inserted in the tube sockets instead of 12AU7.

Tolly, 12AU7s can be wired for 6.3 volts. There are 3 heater pins someone
correct me it has been a lot of years.  This allows you to put both heaters in parallel. good luck!

However I know Frank that the 9'th pin is the middle betwen 4 and 5 pins, but I won't resolder  anything in the radio. It is still original Collins made peace of history.

I understand now.
I once considered building a power supply inside the case. I thought about removing the power connector and install a plate over the hole. Then a power cord could come through the plate. Maybe a little switch would also fit on the plate or the spare fuse holder could be removed and a switch
could go there.
The radio took about 28 volts at 3.5 amps with most of the power used by the audio output. I have heard the radio works better with a little higher plate voltage.

Especially in terms of intermodulations. ;)
It is exactly what I am going to do! :D

I always thought the best thing to do would be rip out all the tubes
and use transistors running class A but AGC would be a problem.
1 module at a time starting from the speaker to the RF input is the path I would have taken. It becomes a big science project.  

BTW I notice you use the term crapstal. This is a good sign you will be using "strap" soon.  welcome!

I prefer to use tubes in RF and 1'st IF.

However, before I saw a schematics, I considered to rewire filaments, because I did not know that B+ and filaments are powered from different pins of sockets. In such case they could be rewired to any voltage. But when I discovered that filaments and plates are powered separately I decided lo leave all wires on place, may be to convert from 26 to 12 volts, since 12 volt versions of the same tubes are much cheaper and widely available.